mirror of
https://github.com/tencentmusic/supersonic.git
synced 2025-12-14 22:25:19 +00:00
headless integrates knowledge (#722)
This commit is contained in:
daikon
@@ -1,334 +0,0 @@
|
||||
package com.hankcs.hanlp.collection.trie.bintrie;
|
||||
|
||||
import com.hankcs.hanlp.corpus.io.ByteArray;
|
||||
import com.tencent.supersonic.chat.core.knowledge.LoadRemoveService;
|
||||
import com.tencent.supersonic.common.util.ContextUtils;
|
||||
import java.io.DataOutputStream;
|
||||
import java.io.IOException;
|
||||
import java.io.ObjectInput;
|
||||
import java.io.ObjectOutput;
|
||||
import java.util.AbstractMap;
|
||||
import java.util.ArrayDeque;
|
||||
import java.util.Arrays;
|
||||
import java.util.List;
|
||||
import java.util.Map;
|
||||
import java.util.Objects;
|
||||
import java.util.Queue;
|
||||
import java.util.Set;
|
||||
import org.slf4j.Logger;
|
||||
import org.slf4j.LoggerFactory;
|
||||
|
||||
|
||||
public abstract class BaseNode<V> implements Comparable<BaseNode> {
|
||||
|
||||
/**
|
||||
* 状态数组，方便读取的时候用
|
||||
*/
|
||||
static final Status[] ARRAY_STATUS = Status.values();
|
||||
|
||||
private static final Logger logger = LoggerFactory.getLogger(BaseNode.class);
|
||||
/**
|
||||
* 子节点
|
||||
*/
|
||||
protected BaseNode[] child;
|
||||
/**
|
||||
* 节点状态
|
||||
*/
|
||||
protected Status status;
|
||||
/**
|
||||
* 节点代表的字符
|
||||
*/
|
||||
protected char c;
|
||||
/**
|
||||
* 节点代表的值
|
||||
*/
|
||||
protected V value;
|
||||
|
||||
protected String prefix = null;
|
||||
|
||||
public BaseNode<V> transition(String path, int begin) {
|
||||
BaseNode<V> cur = this;
|
||||
for (int i = begin; i < path.length(); ++i) {
|
||||
cur = cur.getChild(path.charAt(i));
|
||||
if (cur == null || cur.status == Status.UNDEFINED_0) {
|
||||
return null;
|
||||
}
|
||||
}
|
||||
return cur;
|
||||
}
|
||||
|
||||
public BaseNode<V> transition(char[] path, int begin) {
|
||||
BaseNode<V> cur = this;
|
||||
for (int i = begin; i < path.length; ++i) {
|
||||
cur = cur.getChild(path[i]);
|
||||
if (cur == null || cur.status == Status.UNDEFINED_0) {
|
||||
return null;
|
||||
}
|
||||
}
|
||||
return cur;
|
||||
}
|
||||
|
||||
/**
|
||||
* 转移状态
|
||||
*
|
||||
* @param path
|
||||
* @return
|
||||
*/
|
||||
public BaseNode<V> transition(char path) {
|
||||
BaseNode<V> cur = this;
|
||||
cur = cur.getChild(path);
|
||||
if (cur == null || cur.status == Status.UNDEFINED_0) {
|
||||
return null;
|
||||
}
|
||||
return cur;
|
||||
}
|
||||
|
||||
/**
|
||||
* 添加子节点
|
||||
*
|
||||
* @return true-新增了节点 false-修改了现有节点
|
||||
*/
|
||||
protected abstract boolean addChild(BaseNode node);
|
||||
|
||||
/**
|
||||
* 是否含有子节点
|
||||
*
|
||||
* @param c 子节点的char
|
||||
* @return 是否含有
|
||||
*/
|
||||
protected boolean hasChild(char c) {
|
||||
return getChild(c) != null;
|
||||
}
|
||||
|
||||
protected char getChar() {
|
||||
return c;
|
||||
}
|
||||
|
||||
/**
|
||||
* 获取子节点
|
||||
*
|
||||
* @param c 子节点的char
|
||||
* @return 子节点
|
||||
*/
|
||||
public abstract BaseNode getChild(char c);
|
||||
|
||||
/**
|
||||
* 获取节点对应的值
|
||||
*
|
||||
* @return 值
|
||||
*/
|
||||
public final V getValue() {
|
||||
return value;
|
||||
}
|
||||
|
||||
/**
|
||||
* 设置节点对应的值
|
||||
*
|
||||
* @param value 值
|
||||
*/
|
||||
public final void setValue(V value) {
|
||||
this.value = value;
|
||||
}
|
||||
|
||||
@Override
|
||||
public int compareTo(BaseNode other) {
|
||||
return compareTo(other.getChar());
|
||||
}
|
||||
|
||||
/**
|
||||
* 重载，与字符的比较
|
||||
*
|
||||
* @param other
|
||||
* @return
|
||||
*/
|
||||
public int compareTo(char other) {
|
||||
if (this.c > other) {
|
||||
return 1;
|
||||
}
|
||||
if (this.c < other) {
|
||||
return -1;
|
||||
}
|
||||
return 0;
|
||||
}
|
||||
|
||||
/**
|
||||
* 获取节点的成词状态
|
||||
*
|
||||
* @return
|
||||
*/
|
||||
public Status getStatus() {
|
||||
return status;
|
||||
}
|
||||
|
||||
protected void walk(StringBuilder sb, Set<Map.Entry<String, V>> entrySet) {
|
||||
sb.append(c);
|
||||
if (status == Status.WORD_MIDDLE_2 || status == Status.WORD_END_3) {
|
||||
entrySet.add(new TrieEntry(sb.toString(), value));
|
||||
}
|
||||
if (child == null) {
|
||||
return;
|
||||
}
|
||||
for (BaseNode node : child) {
|
||||
if (node == null) {
|
||||
continue;
|
||||
}
|
||||
node.walk(new StringBuilder(sb.toString()), entrySet);
|
||||
}
|
||||
}
|
||||
|
||||
protected void walkToSave(DataOutputStream out) throws IOException {
|
||||
out.writeChar(c);
|
||||
out.writeInt(status.ordinal());
|
||||
int childSize = 0;
|
||||
if (child != null) {
|
||||
childSize = child.length;
|
||||
}
|
||||
out.writeInt(childSize);
|
||||
if (child == null) {
|
||||
return;
|
||||
}
|
||||
for (BaseNode node : child) {
|
||||
node.walkToSave(out);
|
||||
}
|
||||
}
|
||||
|
||||
protected void walkToSave(ObjectOutput out) throws IOException {
|
||||
out.writeChar(c);
|
||||
out.writeInt(status.ordinal());
|
||||
if (status == Status.WORD_END_3 || status == Status.WORD_MIDDLE_2) {
|
||||
out.writeObject(value);
|
||||
}
|
||||
int childSize = 0;
|
||||
if (child != null) {
|
||||
childSize = child.length;
|
||||
}
|
||||
out.writeInt(childSize);
|
||||
if (child == null) {
|
||||
return;
|
||||
}
|
||||
for (BaseNode node : child) {
|
||||
node.walkToSave(out);
|
||||
}
|
||||
}
|
||||
|
||||
protected void walkToLoad(ByteArray byteArray, _ValueArray<V> valueArray) {
|
||||
c = byteArray.nextChar();
|
||||
status = ARRAY_STATUS[byteArray.nextInt()];
|
||||
if (status == Status.WORD_END_3 || status == Status.WORD_MIDDLE_2) {
|
||||
value = valueArray.nextValue();
|
||||
}
|
||||
int childSize = byteArray.nextInt();
|
||||
child = new BaseNode[childSize];
|
||||
for (int i = 0; i < childSize; ++i) {
|
||||
child[i] = new Node<V>();
|
||||
child[i].walkToLoad(byteArray, valueArray);
|
||||
}
|
||||
}
|
||||
|
||||
protected void walkToLoad(ObjectInput byteArray) throws IOException, ClassNotFoundException {
|
||||
c = byteArray.readChar();
|
||||
status = ARRAY_STATUS[byteArray.readInt()];
|
||||
if (status == Status.WORD_END_3 || status == Status.WORD_MIDDLE_2) {
|
||||
value = (V) byteArray.readObject();
|
||||
}
|
||||
int childSize = byteArray.readInt();
|
||||
child = new BaseNode[childSize];
|
||||
for (int i = 0; i < childSize; ++i) {
|
||||
child[i] = new Node<V>();
|
||||
child[i].walkToLoad(byteArray);
|
||||
}
|
||||
}
|
||||
|
||||
public enum Status {
|
||||
/**
|
||||
* 未指定，用于删除词条
|
||||
*/
|
||||
UNDEFINED_0,
|
||||
/**
|
||||
* 不是词语的结尾
|
||||
*/
|
||||
NOT_WORD_1,
|
||||
/**
|
||||
* 是个词语的结尾，并且还可以继续
|
||||
*/
|
||||
WORD_MIDDLE_2,
|
||||
/**
|
||||
* 是个词语的结尾，并且没有继续
|
||||
*/
|
||||
WORD_END_3,
|
||||
}
|
||||
|
||||
public class TrieEntry extends AbstractMap.SimpleEntry<String, V> implements Comparable<TrieEntry> {
|
||||
|
||||
public TrieEntry(String key, V value) {
|
||||
super(key, value);
|
||||
}
|
||||
|
||||
@Override
|
||||
public int compareTo(TrieEntry o) {
|
||||
return getKey().compareTo(String.valueOf(o.getKey()));
|
||||
}
|
||||
}
|
||||
|
||||
@Override
|
||||
public String toString() {
|
||||
return "BaseNode{"
|
||||
+ "child="
|
||||
+ Arrays.toString(child)
|
||||
+ ", status="
|
||||
+ status
|
||||
+ ", c="
|
||||
+ c
|
||||
+ ", value="
|
||||
+ value
|
||||
+ ", prefix='"
|
||||
+ prefix
|
||||
+ '\''
|
||||
+ '}';
|
||||
}
|
||||
|
||||
public void walkNode(Set<Map.Entry<String, V>> entrySet, Integer agentId, Set<Long> detectModelIds) {
|
||||
if (status == Status.WORD_MIDDLE_2 || status == Status.WORD_END_3) {
|
||||
LoadRemoveService loadRemoveService = ContextUtils.getBean(LoadRemoveService.class);
|
||||
logger.debug("agentId:{},detectModelIds:{},before:{}", agentId, detectModelIds, value.toString());
|
||||
List natures = loadRemoveService.removeNatures((List) value, agentId, detectModelIds);
|
||||
String name = this.prefix != null ? this.prefix + c : "" + c;
|
||||
logger.debug("name:{},after:{},natures:{}", name, (List) value, natures);
|
||||
entrySet.add(new TrieEntry(name, (V) natures));
|
||||
}
|
||||
}
|
||||
|
||||
/***
|
||||
* walk limit
|
||||
* @param sb
|
||||
* @param entrySet
|
||||
* @param limit
|
||||
*/
|
||||
public void walkLimit(StringBuilder sb, Set<Map.Entry<String, V>> entrySet, int limit, Integer agentId,
|
||||
Set<Long> detectModelIds) {
|
||||
Queue<BaseNode> queue = new ArrayDeque<>();
|
||||
this.prefix = sb.toString();
|
||||
queue.add(this);
|
||||
while (!queue.isEmpty()) {
|
||||
if (entrySet.size() >= limit) {
|
||||
break;
|
||||
}
|
||||
BaseNode root = queue.poll();
|
||||
if (root == null) {
|
||||
continue;
|
||||
}
|
||||
root.walkNode(entrySet, agentId, detectModelIds);
|
||||
if (root.child == null) {
|
||||
continue;
|
||||
}
|
||||
String prefix = root.prefix + root.c;
|
||||
for (BaseNode node : root.child) {
|
||||
if (Objects.nonNull(node)) {
|
||||
node.prefix = prefix;
|
||||
queue.add(node);
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
}
|
||||
@@ -1,393 +0,0 @@
|
||||
package com.hankcs.hanlp.dictionary;
|
||||
|
||||
|
||||
import static com.hankcs.hanlp.utility.Predefine.logger;
|
||||
|
||||
import com.hankcs.hanlp.HanLP;
|
||||
import com.hankcs.hanlp.collection.trie.DoubleArrayTrie;
|
||||
import com.hankcs.hanlp.corpus.io.ByteArray;
|
||||
import com.hankcs.hanlp.corpus.io.IOUtil;
|
||||
import com.hankcs.hanlp.corpus.tag.Nature;
|
||||
import com.hankcs.hanlp.utility.Predefine;
|
||||
import com.hankcs.hanlp.utility.TextUtility;
|
||||
|
||||
import java.io.BufferedOutputStream;
|
||||
import java.io.BufferedReader;
|
||||
import java.io.DataOutputStream;
|
||||
import java.io.FileNotFoundException;
|
||||
import java.io.IOException;
|
||||
import java.io.InputStreamReader;
|
||||
import java.io.Serializable;
|
||||
import java.util.Collection;
|
||||
import java.util.TreeMap;
|
||||
|
||||
/**
|
||||
* 使用DoubleArrayTrie实现的核心词典
|
||||
*/
|
||||
public class CoreDictionary {
|
||||
|
||||
public static DoubleArrayTrie<Attribute> trie = new DoubleArrayTrie<Attribute>();
|
||||
|
||||
public static final String PATH = HanLP.Config.CoreDictionaryPath;
|
||||
|
||||
// 自动加载词典
|
||||
static {
|
||||
long start = System.currentTimeMillis();
|
||||
if (!load(PATH)) {
|
||||
throw new IllegalArgumentException("核心词典" + PATH + "加载失败");
|
||||
} else {
|
||||
logger.info(PATH + "加载成功，" + trie.size() + "个词条，耗时" + (System.currentTimeMillis() - start) + "ms");
|
||||
}
|
||||
}
|
||||
|
||||
// 一些特殊的WORD_ID
|
||||
public static final int NR_WORD_ID = getWordID(Predefine.TAG_PEOPLE);
|
||||
public static final int NS_WORD_ID = getWordID(Predefine.TAG_PLACE);
|
||||
public static final int NT_WORD_ID = getWordID(Predefine.TAG_GROUP);
|
||||
public static final int T_WORD_ID = getWordID(Predefine.TAG_TIME);
|
||||
public static final int X_WORD_ID = getWordID(Predefine.TAG_CLUSTER);
|
||||
public static final int M_WORD_ID = getWordID(Predefine.TAG_NUMBER);
|
||||
public static final int NX_WORD_ID = getWordID(Predefine.TAG_PROPER);
|
||||
|
||||
private static boolean load(String path) {
|
||||
logger.info("核心词典开始加载:" + path);
|
||||
if (loadDat(path)) {
|
||||
return true;
|
||||
}
|
||||
TreeMap<String, Attribute> map = new TreeMap<String, Attribute>();
|
||||
BufferedReader br = null;
|
||||
try {
|
||||
br = new BufferedReader(new InputStreamReader(IOUtil.newInputStream(path), "UTF-8"));
|
||||
String line;
|
||||
int totalFrequency = 0;
|
||||
long start = System.currentTimeMillis();
|
||||
while ((line = br.readLine()) != null) {
|
||||
String[] param = line.split("\\s");
|
||||
int natureCount = (param.length - 1) / 2;
|
||||
Attribute attribute = new Attribute(natureCount);
|
||||
for (int i = 0; i < natureCount; ++i) {
|
||||
attribute.nature[i] = Nature.create(param[1 + 2 * i]);
|
||||
attribute.frequency[i] = Integer.parseInt(param[2 + 2 * i]);
|
||||
attribute.totalFrequency += attribute.frequency[i];
|
||||
}
|
||||
map.put(param[0], attribute);
|
||||
totalFrequency += attribute.totalFrequency;
|
||||
}
|
||||
logger.info(
|
||||
"核心词典读入词条" + map.size() + " 全部频次" + totalFrequency + "，耗时" + (System.currentTimeMillis() - start)
|
||||
+ "ms");
|
||||
br.close();
|
||||
trie.build(map);
|
||||
logger.info("核心词典加载成功:" + trie.size() + "个词条，下面将写入缓存……");
|
||||
try {
|
||||
DataOutputStream out = new DataOutputStream(
|
||||
new BufferedOutputStream(IOUtil.newOutputStream(path + Predefine.BIN_EXT)));
|
||||
Collection<Attribute> attributeList = map.values();
|
||||
out.writeInt(attributeList.size());
|
||||
for (Attribute attribute : attributeList) {
|
||||
out.writeInt(attribute.totalFrequency);
|
||||
out.writeInt(attribute.nature.length);
|
||||
for (int i = 0; i < attribute.nature.length; ++i) {
|
||||
out.writeInt(attribute.nature[i].ordinal());
|
||||
out.writeInt(attribute.frequency[i]);
|
||||
}
|
||||
}
|
||||
trie.save(out);
|
||||
out.writeInt(totalFrequency);
|
||||
Predefine.setTotalFrequency(totalFrequency);
|
||||
out.close();
|
||||
} catch (Exception e) {
|
||||
logger.warning("保存失败" + e);
|
||||
return false;
|
||||
}
|
||||
} catch (FileNotFoundException e) {
|
||||
logger.warning("核心词典" + path + "不存在！" + e);
|
||||
return false;
|
||||
} catch (IOException e) {
|
||||
logger.warning("核心词典" + path + "读取错误！" + e);
|
||||
return false;
|
||||
}
|
||||
|
||||
return true;
|
||||
}
|
||||
|
||||
/**
|
||||
* 从磁盘加载双数组
|
||||
*
|
||||
* @param path
|
||||
* @return
|
||||
*/
|
||||
static boolean loadDat(String path) {
|
||||
try {
|
||||
ByteArray byteArray = ByteArray.createByteArray(path + Predefine.BIN_EXT);
|
||||
if (byteArray == null) {
|
||||
return false;
|
||||
}
|
||||
int size = byteArray.nextInt();
|
||||
Attribute[] attributes = new Attribute[size];
|
||||
final Nature[] natureIndexArray = Nature.values();
|
||||
for (int i = 0; i < size; ++i) {
|
||||
// 第一个是全部频次，第二个是词性个数
|
||||
int currentTotalFrequency = byteArray.nextInt();
|
||||
int length = byteArray.nextInt();
|
||||
attributes[i] = new Attribute(length);
|
||||
attributes[i].totalFrequency = currentTotalFrequency;
|
||||
for (int j = 0; j < length; ++j) {
|
||||
attributes[i].nature[j] = natureIndexArray[byteArray.nextInt()];
|
||||
attributes[i].frequency[j] = byteArray.nextInt();
|
||||
}
|
||||
}
|
||||
if (!trie.load(byteArray, attributes)) {
|
||||
return false;
|
||||
}
|
||||
int totalFrequency = 0;
|
||||
if (byteArray.hasMore()) {
|
||||
totalFrequency = byteArray.nextInt();
|
||||
} else {
|
||||
for (Attribute attribute : attributes) {
|
||||
totalFrequency += attribute.totalFrequency;
|
||||
}
|
||||
}
|
||||
Predefine.setTotalFrequency(totalFrequency);
|
||||
} catch (Exception e) {
|
||||
logger.warning("读取失败，问题发生在" + e);
|
||||
return false;
|
||||
}
|
||||
return true;
|
||||
}
|
||||
|
||||
/**
|
||||
* 获取条目
|
||||
*
|
||||
* @param key
|
||||
* @return
|
||||
*/
|
||||
public static Attribute get(String key) {
|
||||
return trie.get(key);
|
||||
}
|
||||
|
||||
/**
|
||||
* 获取条目
|
||||
*
|
||||
* @param wordID
|
||||
* @return
|
||||
*/
|
||||
public static Attribute get(int wordID) {
|
||||
return trie.get(wordID);
|
||||
}
|
||||
|
||||
/**
|
||||
* 获取词频
|
||||
*
|
||||
* @param term
|
||||
* @return
|
||||
*/
|
||||
public static int getTermFrequency(String term) {
|
||||
Attribute attribute = get(term);
|
||||
if (attribute == null) {
|
||||
return 0;
|
||||
}
|
||||
return attribute.totalFrequency;
|
||||
}
|
||||
|
||||
/**
|
||||
* 是否包含词语
|
||||
*
|
||||
* @param key
|
||||
* @return
|
||||
*/
|
||||
public static boolean contains(String key) {
|
||||
return trie.get(key) != null;
|
||||
}
|
||||
|
||||
/**
|
||||
* 核心词典中的词属性
|
||||
*/
|
||||
public static class Attribute implements Serializable {
|
||||
|
||||
/**
|
||||
* 词性列表
|
||||
*/
|
||||
public Nature[] nature;
|
||||
/**
|
||||
* 词性对应的词频
|
||||
*/
|
||||
public int[] frequency;
|
||||
|
||||
public int totalFrequency;
|
||||
public String original = null;
|
||||
|
||||
|
||||
public Attribute(int size) {
|
||||
nature = new Nature[size];
|
||||
frequency = new int[size];
|
||||
}
|
||||
|
||||
public Attribute(Nature[] nature, int[] frequency) {
|
||||
this.nature = nature;
|
||||
this.frequency = frequency;
|
||||
}
|
||||
|
||||
public Attribute(Nature nature, int frequency) {
|
||||
this(1);
|
||||
this.nature[0] = nature;
|
||||
this.frequency[0] = frequency;
|
||||
totalFrequency = frequency;
|
||||
}
|
||||
|
||||
public Attribute(Nature[] nature, int[] frequency, int totalFrequency) {
|
||||
this.nature = nature;
|
||||
this.frequency = frequency;
|
||||
this.totalFrequency = totalFrequency;
|
||||
}
|
||||
|
||||
/**
|
||||
* 使用单个词性，默认词频1000构造
|
||||
*
|
||||
* @param nature
|
||||
*/
|
||||
public Attribute(Nature nature) {
|
||||
this(nature, 1000);
|
||||
}
|
||||
|
||||
public static Attribute create(String natureWithFrequency) {
|
||||
try {
|
||||
String[] param = natureWithFrequency.split(" ");
|
||||
if (param.length % 2 != 0) {
|
||||
return new Attribute(Nature.create(natureWithFrequency.trim()), 1); // 儿童锁
|
||||
}
|
||||
int natureCount = param.length / 2;
|
||||
Attribute attribute = new Attribute(natureCount);
|
||||
for (int i = 0; i < natureCount; ++i) {
|
||||
attribute.nature[i] = Nature.create(param[2 * i]);
|
||||
attribute.frequency[i] = Integer.parseInt(param[1 + 2 * i]);
|
||||
attribute.totalFrequency += attribute.frequency[i];
|
||||
}
|
||||
return attribute;
|
||||
} catch (Exception e) {
|
||||
logger.warning("使用字符串" + natureWithFrequency + "创建词条属性失败！" + TextUtility.exceptionToString(e));
|
||||
return null;
|
||||
}
|
||||
}
|
||||
|
||||
/**
|
||||
* 从字节流中加载
|
||||
*
|
||||
* @param byteArray
|
||||
* @param natureIndexArray
|
||||
* @return
|
||||
*/
|
||||
public static Attribute create(ByteArray byteArray, Nature[] natureIndexArray) {
|
||||
int currentTotalFrequency = byteArray.nextInt();
|
||||
int length = byteArray.nextInt();
|
||||
Attribute attribute = new Attribute(length);
|
||||
attribute.totalFrequency = currentTotalFrequency;
|
||||
for (int j = 0; j < length; ++j) {
|
||||
attribute.nature[j] = natureIndexArray[byteArray.nextInt()];
|
||||
attribute.frequency[j] = byteArray.nextInt();
|
||||
}
|
||||
|
||||
return attribute;
|
||||
}
|
||||
|
||||
/**
|
||||
* 获取词性的词频
|
||||
*
|
||||
* @param nature 字符串词性
|
||||
* @return 词频
|
||||
* @deprecated 推荐使用Nature参数！
|
||||
*/
|
||||
public int getNatureFrequency(String nature) {
|
||||
try {
|
||||
Nature pos = Nature.create(nature);
|
||||
return getNatureFrequency(pos);
|
||||
} catch (IllegalArgumentException e) {
|
||||
return 0;
|
||||
}
|
||||
}
|
||||
|
||||
/**
|
||||
* 获取词性的词频
|
||||
*
|
||||
* @param nature 词性
|
||||
* @return 词频
|
||||
*/
|
||||
public int getNatureFrequency(final Nature nature) {
|
||||
int i = 0;
|
||||
for (Nature pos : this.nature) {
|
||||
if (nature == pos) {
|
||||
return frequency[i];
|
||||
}
|
||||
++i;
|
||||
}
|
||||
return 0;
|
||||
}
|
||||
|
||||
/**
|
||||
* 是否有某个词性
|
||||
*
|
||||
* @param nature
|
||||
* @return
|
||||
*/
|
||||
public boolean hasNature(Nature nature) {
|
||||
return getNatureFrequency(nature) > 0;
|
||||
}
|
||||
|
||||
/**
|
||||
* 是否有以某个前缀开头的词性
|
||||
*
|
||||
* @param prefix 词性前缀，比如u会查询是否有ude, uzhe等等
|
||||
* @return
|
||||
*/
|
||||
public boolean hasNatureStartsWith(String prefix) {
|
||||
for (Nature n : nature) {
|
||||
if (n.startsWith(prefix)) {
|
||||
return true;
|
||||
}
|
||||
}
|
||||
return false;
|
||||
}
|
||||
|
||||
@Override
|
||||
public String toString() {
|
||||
final StringBuilder sb = new StringBuilder();
|
||||
for (int i = 0; i < nature.length; ++i) {
|
||||
sb.append(nature[i]).append(' ').append(frequency[i]).append(' ');
|
||||
}
|
||||
return sb.toString();
|
||||
}
|
||||
|
||||
public void save(DataOutputStream out) throws IOException {
|
||||
out.writeInt(totalFrequency);
|
||||
out.writeInt(nature.length);
|
||||
for (int i = 0; i < nature.length; ++i) {
|
||||
out.writeInt(nature[i].ordinal());
|
||||
out.writeInt(frequency[i]);
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
/**
|
||||
* 获取词语的ID
|
||||
*
|
||||
* @param a 词语
|
||||
* @return ID, 如果不存在, 则返回-1
|
||||
*/
|
||||
public static int getWordID(String a) {
|
||||
return CoreDictionary.trie.exactMatchSearch(a);
|
||||
}
|
||||
|
||||
/**
|
||||
* 热更新核心词典<br>
|
||||
* 集群环境（或其他IOAdapter）需要自行删除缓存文件
|
||||
*
|
||||
* @return 是否成功
|
||||
*/
|
||||
public static boolean reload() {
|
||||
String path = CoreDictionary.PATH;
|
||||
IOUtil.deleteFile(path + Predefine.BIN_EXT);
|
||||
|
||||
return load(path);
|
||||
}
|
||||
}
|
||||
|
||||
@@ -1,341 +0,0 @@
|
||||
package com.hankcs.hanlp.seg;
|
||||
|
||||
|
||||
import com.hankcs.hanlp.algorithm.Viterbi;
|
||||
import com.hankcs.hanlp.collection.AhoCorasick.AhoCorasickDoubleArrayTrie;
|
||||
import com.hankcs.hanlp.collection.trie.DoubleArrayTrie;
|
||||
import com.hankcs.hanlp.corpus.tag.Nature;
|
||||
import com.hankcs.hanlp.dictionary.CoreDictionary;
|
||||
import com.hankcs.hanlp.dictionary.CoreDictionaryTransformMatrixDictionary;
|
||||
import com.hankcs.hanlp.dictionary.other.CharType;
|
||||
import com.hankcs.hanlp.seg.NShort.Path.AtomNode;
|
||||
import com.hankcs.hanlp.seg.common.Graph;
|
||||
import com.hankcs.hanlp.seg.common.Term;
|
||||
import com.hankcs.hanlp.seg.common.Vertex;
|
||||
import com.hankcs.hanlp.seg.common.WordNet;
|
||||
import com.hankcs.hanlp.utility.TextUtility;
|
||||
import java.util.ArrayList;
|
||||
import java.util.Iterator;
|
||||
import java.util.LinkedList;
|
||||
import java.util.List;
|
||||
import java.util.ListIterator;
|
||||
|
||||
|
||||
public abstract class WordBasedSegment extends Segment {
|
||||
|
||||
public WordBasedSegment() {
|
||||
}
|
||||
|
||||
protected static void generateWord(List<Vertex> linkedArray, WordNet wordNetOptimum) {
|
||||
fixResultByRule(linkedArray);
|
||||
wordNetOptimum.addAll(linkedArray);
|
||||
}
|
||||
|
||||
protected static void fixResultByRule(List<Vertex> linkedArray) {
|
||||
mergeContinueNumIntoOne(linkedArray);
|
||||
changeDelimiterPOS(linkedArray);
|
||||
splitMiddleSlashFromDigitalWords(linkedArray);
|
||||
checkDateElements(linkedArray);
|
||||
}
|
||||
|
||||
static void changeDelimiterPOS(List<Vertex> linkedArray) {
|
||||
Iterator var1 = linkedArray.iterator();
|
||||
|
||||
while (true) {
|
||||
Vertex vertex;
|
||||
do {
|
||||
if (!var1.hasNext()) {
|
||||
return;
|
||||
}
|
||||
|
||||
vertex = (Vertex) var1.next();
|
||||
} while (!vertex.realWord.equals("－－") && !vertex.realWord.equals("—") && !vertex.realWord.equals("-"));
|
||||
|
||||
vertex.confirmNature(Nature.w);
|
||||
}
|
||||
}
|
||||
|
||||
private static void splitMiddleSlashFromDigitalWords(List<Vertex> linkedArray) {
|
||||
if (linkedArray.size() >= 2) {
|
||||
ListIterator<Vertex> listIterator = linkedArray.listIterator();
|
||||
Vertex next = (Vertex) listIterator.next();
|
||||
|
||||
for (Vertex current = next; listIterator.hasNext(); current = next) {
|
||||
next = (Vertex) listIterator.next();
|
||||
Nature currentNature = current.getNature();
|
||||
if (currentNature == Nature.nx && (next.hasNature(Nature.q) || next.hasNature(Nature.n))) {
|
||||
String[] param = current.realWord.split("-", 1);
|
||||
if (param.length == 2 && TextUtility.isAllNum(param[0]) && TextUtility.isAllNum(param[1])) {
|
||||
current = current.copy();
|
||||
current.realWord = param[0];
|
||||
current.confirmNature(Nature.m);
|
||||
listIterator.previous();
|
||||
listIterator.previous();
|
||||
listIterator.set(current);
|
||||
listIterator.next();
|
||||
listIterator.add(Vertex.newPunctuationInstance("-"));
|
||||
listIterator.add(Vertex.newNumberInstance(param[1]));
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
}
|
||||
}
|
||||
|
||||
private static void checkDateElements(List<Vertex> linkedArray) {
|
||||
if (linkedArray.size() >= 2) {
|
||||
ListIterator<Vertex> listIterator = linkedArray.listIterator();
|
||||
Vertex next = (Vertex) listIterator.next();
|
||||
|
||||
for (Vertex current = next; listIterator.hasNext(); current = next) {
|
||||
next = (Vertex) listIterator.next();
|
||||
if (TextUtility.isAllNum(current.realWord) || TextUtility.isAllChineseNum(current.realWord)) {
|
||||
String nextWord = next.realWord;
|
||||
if (nextWord.length() == 1 && "月日时分秒".contains(nextWord)
|
||||
|| nextWord.length() == 2 && nextWord.equals("月份")) {
|
||||
mergeDate(listIterator, next, current);
|
||||
} else if (nextWord.equals("年")) {
|
||||
if (TextUtility.isYearTime(current.realWord)) {
|
||||
mergeDate(listIterator, next, current);
|
||||
} else {
|
||||
current.confirmNature(Nature.m);
|
||||
}
|
||||
} else if (current.realWord.endsWith("点")) {
|
||||
current.confirmNature(Nature.t, true);
|
||||
} else {
|
||||
char[] tmpCharArray = current.realWord.toCharArray();
|
||||
String lastChar = String.valueOf(tmpCharArray[tmpCharArray.length - 1]);
|
||||
if (!"∶·．／./".contains(lastChar)) {
|
||||
current.confirmNature(Nature.m, true);
|
||||
} else if (current.realWord.length() > 1) {
|
||||
char last = current.realWord.charAt(current.realWord.length() - 1);
|
||||
current = Vertex.newNumberInstance(
|
||||
current.realWord.substring(0, current.realWord.length() - 1));
|
||||
listIterator.previous();
|
||||
listIterator.previous();
|
||||
listIterator.set(current);
|
||||
listIterator.next();
|
||||
listIterator.add(Vertex.newPunctuationInstance(String.valueOf(last)));
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
}
|
||||
}
|
||||
|
||||
private static void mergeDate(ListIterator<Vertex> listIterator, Vertex next, Vertex current) {
|
||||
current = Vertex.newTimeInstance(current.realWord + next.realWord);
|
||||
listIterator.previous();
|
||||
listIterator.previous();
|
||||
listIterator.set(current);
|
||||
listIterator.next();
|
||||
listIterator.next();
|
||||
listIterator.remove();
|
||||
}
|
||||
|
||||
protected static List<Term> convert(List<Vertex> vertexList) {
|
||||
return convert(vertexList, false);
|
||||
}
|
||||
|
||||
protected static Graph generateBiGraph(WordNet wordNet) {
|
||||
return wordNet.toGraph();
|
||||
}
|
||||
|
||||
/**
|
||||
* @deprecated
|
||||
*/
|
||||
private static List<AtomNode> atomSegment(String sSentence, int start, int end) {
|
||||
if (end < start) {
|
||||
throw new RuntimeException("start=" + start + " < end=" + end);
|
||||
} else {
|
||||
List<AtomNode> atomSegment = new ArrayList();
|
||||
int pCur = 0;
|
||||
StringBuilder sb = new StringBuilder();
|
||||
char[] charArray = sSentence.substring(start, end).toCharArray();
|
||||
int[] charTypeArray = new int[charArray.length];
|
||||
|
||||
for (int i = 0; i < charArray.length; ++i) {
|
||||
char c = charArray[i];
|
||||
charTypeArray[i] = CharType.get(c);
|
||||
if (c == '.' && i < charArray.length - 1 && CharType.get(charArray[i + 1]) == 9) {
|
||||
charTypeArray[i] = 9;
|
||||
} else if (c == '.' && i < charArray.length - 1 && charArray[i + 1] >= '0' && charArray[i + 1] <= '9') {
|
||||
charTypeArray[i] = 5;
|
||||
} else if (charTypeArray[i] == 8) {
|
||||
charTypeArray[i] = 5;
|
||||
}
|
||||
}
|
||||
|
||||
while (true) {
|
||||
while (true) {
|
||||
while (pCur < charArray.length) {
|
||||
int nCurType = charTypeArray[pCur];
|
||||
if (nCurType != 7 && nCurType != 10 && nCurType != 6 && nCurType != 17) {
|
||||
if (pCur < charArray.length - 1 && (nCurType == 5 || nCurType == 9)) {
|
||||
sb.delete(0, sb.length());
|
||||
sb.append(charArray[pCur]);
|
||||
boolean reachEnd = true;
|
||||
|
||||
while (pCur < charArray.length - 1) {
|
||||
++pCur;
|
||||
int nNextType = charTypeArray[pCur];
|
||||
if (nNextType != nCurType) {
|
||||
reachEnd = false;
|
||||
break;
|
||||
}
|
||||
|
||||
sb.append(charArray[pCur]);
|
||||
}
|
||||
|
||||
atomSegment.add(new AtomNode(sb.toString(), nCurType));
|
||||
if (reachEnd) {
|
||||
++pCur;
|
||||
}
|
||||
} else {
|
||||
atomSegment.add(new AtomNode(charArray[pCur], nCurType));
|
||||
++pCur;
|
||||
}
|
||||
} else {
|
||||
String single = String.valueOf(charArray[pCur]);
|
||||
if (single.length() != 0) {
|
||||
atomSegment.add(new AtomNode(single, nCurType));
|
||||
}
|
||||
|
||||
++pCur;
|
||||
}
|
||||
}
|
||||
|
||||
return atomSegment;
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
private static void mergeContinueNumIntoOne(List<Vertex> linkedArray) {
|
||||
if (linkedArray.size() >= 2) {
|
||||
ListIterator<Vertex> listIterator = linkedArray.listIterator();
|
||||
Vertex next = (Vertex) listIterator.next();
|
||||
Vertex current = next;
|
||||
|
||||
while (true) {
|
||||
while (listIterator.hasNext()) {
|
||||
next = (Vertex) listIterator.next();
|
||||
if (!TextUtility.isAllNum(current.realWord) && !TextUtility.isAllChineseNum(current.realWord)
|
||||
|| !TextUtility.isAllNum(next.realWord) && !TextUtility.isAllChineseNum(next.realWord)) {
|
||||
current = next;
|
||||
} else {
|
||||
current = Vertex.newNumberInstance(current.realWord + next.realWord);
|
||||
listIterator.previous();
|
||||
listIterator.previous();
|
||||
listIterator.set(current);
|
||||
listIterator.next();
|
||||
listIterator.next();
|
||||
listIterator.remove();
|
||||
}
|
||||
}
|
||||
|
||||
return;
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
protected void generateWordNet(final WordNet wordNetStorage) {
|
||||
final char[] charArray = wordNetStorage.charArray;
|
||||
DoubleArrayTrie.Searcher searcher = CoreDictionary.trie.getSearcher(charArray, 0);
|
||||
|
||||
while (searcher.next()) {
|
||||
wordNetStorage.add(searcher.begin + 1, new Vertex(new String(charArray, searcher.begin, searcher.length),
|
||||
(CoreDictionary.Attribute) searcher.value, searcher.index));
|
||||
}
|
||||
|
||||
if (this.config.forceCustomDictionary) {
|
||||
this.customDictionary.parseText(charArray, new AhoCorasickDoubleArrayTrie.IHit<CoreDictionary.Attribute>() {
|
||||
public void hit(int begin, int end, CoreDictionary.Attribute value) {
|
||||
wordNetStorage.add(begin + 1, new Vertex(new String(charArray, begin, end - begin), value));
|
||||
}
|
||||
});
|
||||
}
|
||||
|
||||
LinkedList<Vertex>[] vertexes = wordNetStorage.getVertexes();
|
||||
int i = 1;
|
||||
|
||||
while (true) {
|
||||
while (i < vertexes.length) {
|
||||
if (vertexes[i].isEmpty()) {
|
||||
int j;
|
||||
for (j = i + 1;
|
||||
j < vertexes.length - 1 && (vertexes[j].isEmpty() || CharType.get(charArray[j - 1]) == 11);
|
||||
++j) {
|
||||
}
|
||||
|
||||
wordNetStorage.add(i, quickAtomSegment(charArray, i - 1, j - 1));
|
||||
i = j;
|
||||
} else {
|
||||
i += ((Vertex) vertexes[i].getLast()).realWord.length();
|
||||
}
|
||||
}
|
||||
|
||||
return;
|
||||
}
|
||||
}
|
||||
|
||||
protected List<Term> decorateResultForIndexMode(List<Vertex> vertexList, WordNet wordNetAll) {
|
||||
List<Term> termList = new LinkedList();
|
||||
int line = 1;
|
||||
ListIterator<Vertex> listIterator = vertexList.listIterator();
|
||||
listIterator.next();
|
||||
int length = vertexList.size() - 2;
|
||||
|
||||
for (int i = 0; i < length; ++i) {
|
||||
Vertex vertex = (Vertex) listIterator.next();
|
||||
Term termMain = convert(vertex);
|
||||
//termList.add(termMain);
|
||||
addTerms(termList, vertex, line - 1);
|
||||
termMain.offset = line - 1;
|
||||
if (vertex.realWord.length() > 2) {
|
||||
label43:
|
||||
for (int currentLine = line; currentLine < line + vertex.realWord.length(); ++currentLine) {
|
||||
Iterator iterator = wordNetAll.descendingIterator(currentLine);
|
||||
|
||||
while (true) {
|
||||
Vertex smallVertex;
|
||||
do {
|
||||
if (!iterator.hasNext()) {
|
||||
continue label43;
|
||||
}
|
||||
smallVertex = (Vertex) iterator.next();
|
||||
} while ((termMain.nature != Nature.mq || !smallVertex.hasNature(Nature.q))
|
||||
&& smallVertex.realWord.length() < this.config.indexMode);
|
||||
|
||||
if (smallVertex != vertex
|
||||
&& currentLine + smallVertex.realWord.length() <= line + vertex.realWord.length()) {
|
||||
listIterator.add(smallVertex);
|
||||
//Term termSub = convert(smallVertex);
|
||||
//termSub.offset = currentLine - 1;
|
||||
//termList.add(termSub);
|
||||
addTerms(termList, smallVertex, currentLine - 1);
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
line += vertex.realWord.length();
|
||||
}
|
||||
|
||||
return termList;
|
||||
}
|
||||
|
||||
protected static void speechTagging(List<Vertex> vertexList) {
|
||||
Viterbi.compute(vertexList, CoreDictionaryTransformMatrixDictionary.transformMatrixDictionary);
|
||||
}
|
||||
|
||||
protected void addTerms(List<Term> terms, Vertex vertex, int offset) {
|
||||
for (int i = 0; i < vertex.attribute.nature.length; i++) {
|
||||
Term term = new Term(vertex.realWord, vertex.attribute.nature[i]);
|
||||
term.setFrequency(vertex.attribute.frequency[i]);
|
||||
term.offset = offset;
|
||||
terms.add(term);
|
||||
}
|
||||
}
|
||||
}
|
||||
@@ -1,69 +0,0 @@
|
||||
package com.hankcs.hanlp.seg.common;
|
||||
|
||||
import com.hankcs.hanlp.corpus.tag.Nature;
|
||||
import com.hankcs.hanlp.dictionary.CoreDictionary;
|
||||
import com.hankcs.hanlp.dictionary.CustomDictionary;
|
||||
import com.tencent.supersonic.chat.core.utils.HanlpHelper;
|
||||
import lombok.Data;
|
||||
import lombok.ToString;
|
||||
|
||||
@Data
|
||||
@ToString
|
||||
public class Term {
|
||||
|
||||
public String word;
|
||||
|
||||
public Nature nature;
|
||||
public int offset;
|
||||
public int frequency = 0;
|
||||
|
||||
public Term(String word, Nature nature) {
|
||||
this.word = word;
|
||||
this.nature = nature;
|
||||
}
|
||||
|
||||
public Term(String word, Nature nature, int offset) {
|
||||
this.word = word;
|
||||
this.nature = nature;
|
||||
this.offset = offset;
|
||||
}
|
||||
|
||||
public Term(String word, Nature nature, int offset, int frequency) {
|
||||
this.word = word;
|
||||
this.nature = nature;
|
||||
this.offset = offset;
|
||||
this.frequency = frequency;
|
||||
}
|
||||
|
||||
public int length() {
|
||||
return this.word.length();
|
||||
}
|
||||
|
||||
public int getFrequency() {
|
||||
if (frequency > 0) {
|
||||
return frequency;
|
||||
}
|
||||
String wordOri = word.toLowerCase();
|
||||
CoreDictionary.Attribute attribute = HanlpHelper.getDynamicCustomDictionary().get(wordOri);
|
||||
if (attribute == null) {
|
||||
attribute = CoreDictionary.get(wordOri);
|
||||
if (attribute == null) {
|
||||
attribute = CustomDictionary.get(wordOri);
|
||||
}
|
||||
}
|
||||
if (attribute != null && nature != null && attribute.hasNature(nature)) {
|
||||
return attribute.getNatureFrequency(nature);
|
||||
}
|
||||
return attribute == null ? 0 : attribute.totalFrequency;
|
||||
}
|
||||
|
||||
public boolean equals(Object obj) {
|
||||
if (obj instanceof Term) {
|
||||
Term term = (Term) obj;
|
||||
if (this.nature == term.nature && this.word.equals(term.word)) {
|
||||
return true;
|
||||
}
|
||||
}
|
||||
return super.equals(obj);
|
||||
}
|
||||
}
|
||||
@@ -1,7 +1,9 @@
|
||||
package com.tencent.supersonic.chat.core.config;
|
||||
|
||||
import com.tencent.supersonic.chat.core.utils.HanlpHelper;
|
||||
import com.tencent.supersonic.headless.core.knowledge.helper.HanlpHelper;
|
||||
|
||||
import java.io.FileNotFoundException;
|
||||
|
||||
import lombok.Data;
|
||||
import lombok.extern.slf4j.Slf4j;
|
||||
import org.springframework.beans.factory.annotation.Value;
|
||||
@@ -11,7 +13,7 @@ import org.springframework.context.annotation.Configuration;
|
||||
@Data
|
||||
@Configuration
|
||||
@Slf4j
|
||||
public class LocalFileConfig {
|
||||
public class ChatLocalFileConfig {
|
||||
|
||||
|
||||
@Value("${dict.directory.latest:/data/dictionary/custom}")
|
||||
@@ -1,6 +1,6 @@
|
||||
package com.tencent.supersonic.chat.core.corrector;
|
||||
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElement;
|
||||
import com.tencent.supersonic.headless.api.pojo.SchemaElement;
|
||||
import com.tencent.supersonic.chat.api.pojo.SemanticParseInfo;
|
||||
import com.tencent.supersonic.chat.api.pojo.SemanticSchema;
|
||||
import com.tencent.supersonic.chat.core.pojo.QueryContext;
|
||||
|
||||
@@ -1,7 +1,7 @@
|
||||
package com.tencent.supersonic.chat.core.corrector;
|
||||
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElement;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaValueMap;
|
||||
import com.tencent.supersonic.headless.api.pojo.SchemaElement;
|
||||
import com.tencent.supersonic.headless.api.pojo.SchemaValueMap;
|
||||
import com.tencent.supersonic.chat.api.pojo.SemanticParseInfo;
|
||||
import com.tencent.supersonic.chat.api.pojo.SemanticSchema;
|
||||
import com.tencent.supersonic.chat.api.pojo.request.QueryFilters;
|
||||
|
||||
@@ -1,30 +0,0 @@
|
||||
package com.tencent.supersonic.chat.core.knowledge;
|
||||
|
||||
import com.google.common.base.Objects;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElement;
|
||||
import lombok.Data;
|
||||
import lombok.ToString;
|
||||
|
||||
@Data
|
||||
@ToString
|
||||
public class DatabaseMapResult extends MapResult {
|
||||
|
||||
private SchemaElement schemaElement;
|
||||
|
||||
@Override
|
||||
public boolean equals(Object o) {
|
||||
if (this == o) {
|
||||
return true;
|
||||
}
|
||||
if (o == null || getClass() != o.getClass()) {
|
||||
return false;
|
||||
}
|
||||
DatabaseMapResult that = (DatabaseMapResult) o;
|
||||
return Objects.equal(name, that.name) && Objects.equal(schemaElement, that.schemaElement);
|
||||
}
|
||||
|
||||
@Override
|
||||
public int hashCode() {
|
||||
return Objects.hashCode(name, schemaElement);
|
||||
}
|
||||
}
|
||||
@@ -1,13 +0,0 @@
|
||||
package com.tencent.supersonic.chat.core.knowledge;
|
||||
|
||||
import java.util.List;
|
||||
import lombok.Data;
|
||||
|
||||
|
||||
@Data
|
||||
public class DictConfig {
|
||||
|
||||
private Long modelId;
|
||||
|
||||
private List<DimValueInfo> dimValueInfoList;
|
||||
}
|
||||
@@ -1,31 +0,0 @@
|
||||
|
||||
package com.tencent.supersonic.chat.core.knowledge;
|
||||
|
||||
public enum DictUpdateMode {
|
||||
|
||||
OFFLINE_FULL("OFFLINE_FULL"),
|
||||
OFFLINE_MODEL("OFFLINE_MODEL"),
|
||||
REALTIME_ADD("REALTIME_ADD"),
|
||||
REALTIME_DELETE("REALTIME_DELETE"),
|
||||
NOT_SUPPORT("NOT_SUPPORT");
|
||||
|
||||
private String value;
|
||||
|
||||
DictUpdateMode(String value) {
|
||||
this.value = value;
|
||||
}
|
||||
|
||||
public static DictUpdateMode of(String value) {
|
||||
for (DictUpdateMode item : DictUpdateMode.values()) {
|
||||
if (item.value.equalsIgnoreCase(value)) {
|
||||
return item;
|
||||
}
|
||||
}
|
||||
return DictUpdateMode.NOT_SUPPORT;
|
||||
}
|
||||
|
||||
public String getValue() {
|
||||
return value;
|
||||
}
|
||||
|
||||
}
|
||||
@@ -1,34 +0,0 @@
|
||||
package com.tencent.supersonic.chat.core.knowledge;
|
||||
|
||||
import java.util.Objects;
|
||||
import lombok.Data;
|
||||
import lombok.ToString;
|
||||
|
||||
/***
|
||||
* word nature
|
||||
*/
|
||||
@Data
|
||||
@ToString
|
||||
public class DictWord {
|
||||
|
||||
private String word;
|
||||
private String nature;
|
||||
private String natureWithFrequency;
|
||||
|
||||
@Override
|
||||
public boolean equals(Object o) {
|
||||
if (this == o) {
|
||||
return true;
|
||||
}
|
||||
if (o == null || getClass() != o.getClass()) {
|
||||
return false;
|
||||
}
|
||||
DictWord that = (DictWord) o;
|
||||
return Objects.equals(word, that.word) && Objects.equals(natureWithFrequency, that.natureWithFrequency);
|
||||
}
|
||||
|
||||
@Override
|
||||
public int hashCode() {
|
||||
return Objects.hash(word, natureWithFrequency);
|
||||
}
|
||||
}
|
||||
@@ -1,38 +0,0 @@
|
||||
package com.tencent.supersonic.chat.core.knowledge.dictionary;
|
||||
|
||||
import com.hankcs.hanlp.corpus.tag.Nature;
|
||||
import com.hankcs.hanlp.dictionary.CoreDictionary;
|
||||
import java.util.Collections;
|
||||
import java.util.Comparator;
|
||||
import java.util.HashMap;
|
||||
import java.util.LinkedList;
|
||||
import java.util.List;
|
||||
import java.util.Map;
|
||||
import java.util.stream.Collectors;
|
||||
import java.util.stream.IntStream;
|
||||
|
||||
/**
|
||||
* Dictionary Attribute Util
|
||||
*/
|
||||
public class DictionaryAttributeUtil {
|
||||
|
||||
public static CoreDictionary.Attribute getAttribute(CoreDictionary.Attribute old, CoreDictionary.Attribute add) {
|
||||
Map<Nature, Integer> map = new HashMap<>();
|
||||
IntStream.range(0, old.nature.length).boxed().forEach(i -> map.put(old.nature[i], old.frequency[i]));
|
||||
IntStream.range(0, add.nature.length).boxed().forEach(i -> map.put(add.nature[i], add.frequency[i]));
|
||||
List<Map.Entry<Nature, Integer>> list = new LinkedList<Map.Entry<Nature, Integer>>(map.entrySet());
|
||||
Collections.sort(list, new Comparator<Map.Entry<Nature, Integer>>() {
|
||||
public int compare(Map.Entry<Nature, Integer> o1, Map.Entry<Nature, Integer> o2) {
|
||||
return o2.getValue() - o1.getValue();
|
||||
}
|
||||
});
|
||||
CoreDictionary.Attribute attribute = new CoreDictionary.Attribute(
|
||||
list.stream().map(i -> i.getKey()).collect(Collectors.toList()).toArray(new Nature[0]),
|
||||
list.stream().map(i -> i.getValue()).mapToInt(Integer::intValue).toArray(),
|
||||
list.stream().map(i -> i.getValue()).findFirst().get());
|
||||
if (old.original != null || add.original != null) {
|
||||
attribute.original = add.original != null ? add.original : old.original;
|
||||
}
|
||||
return attribute;
|
||||
}
|
||||
}
|
||||
@@ -1,18 +0,0 @@
|
||||
package com.tencent.supersonic.chat.core.knowledge;
|
||||
|
||||
|
||||
import java.util.HashMap;
|
||||
import java.util.List;
|
||||
import java.util.Map;
|
||||
|
||||
import lombok.Data;
|
||||
|
||||
@Data
|
||||
public class DimValue2DictCommand {
|
||||
|
||||
private DictUpdateMode updateMode;
|
||||
|
||||
private List<Long> modelIds;
|
||||
|
||||
private Map<Long, List<Long>> modelAndDimPair = new HashMap<>();
|
||||
}
|
||||
@@ -1,31 +0,0 @@
|
||||
package com.tencent.supersonic.chat.core.knowledge;
|
||||
|
||||
|
||||
import com.tencent.supersonic.common.pojo.enums.TaskStatusEnum;
|
||||
|
||||
import java.util.Date;
|
||||
import java.util.Set;
|
||||
|
||||
import lombok.Data;
|
||||
|
||||
@Data
|
||||
public class DimValueDictInfo {
|
||||
|
||||
private Long id;
|
||||
|
||||
private String name;
|
||||
|
||||
private String description;
|
||||
|
||||
private String command;
|
||||
|
||||
private TaskStatusEnum status;
|
||||
|
||||
private String createdBy;
|
||||
|
||||
private Date createdAt;
|
||||
|
||||
private Long elapsedMs;
|
||||
|
||||
private Set<Long> dimIds;
|
||||
}
|
||||
@@ -1,26 +0,0 @@
|
||||
package com.tencent.supersonic.chat.core.knowledge;
|
||||
|
||||
|
||||
import com.tencent.supersonic.common.pojo.enums.TypeEnums;
|
||||
import java.util.List;
|
||||
import javax.validation.constraints.NotNull;
|
||||
|
||||
public class DimValueInfo {
|
||||
|
||||
/**
|
||||
* metricId、DimensionId、domainId
|
||||
*/
|
||||
private Long itemId;
|
||||
|
||||
/**
|
||||
* type: IntentionTypeEnum
|
||||
* temporarily only supports dimension-related information
|
||||
*/
|
||||
@NotNull
|
||||
private TypeEnums type = TypeEnums.DIMENSION;
|
||||
|
||||
private List<String> blackList;
|
||||
private List<String> whiteList;
|
||||
private List<String> ruleList;
|
||||
private Boolean isDictInfo;
|
||||
}
|
||||
@@ -1,34 +0,0 @@
|
||||
package com.tencent.supersonic.chat.core.knowledge;
|
||||
|
||||
import com.google.common.base.Objects;
|
||||
import java.util.Map;
|
||||
import lombok.Data;
|
||||
import lombok.ToString;
|
||||
|
||||
@Data
|
||||
@ToString
|
||||
public class EmbeddingResult extends MapResult {
|
||||
|
||||
private String id;
|
||||
|
||||
private double distance;
|
||||
|
||||
private Map<String, String> metadata;
|
||||
|
||||
@Override
|
||||
public boolean equals(Object o) {
|
||||
if (this == o) {
|
||||
return true;
|
||||
}
|
||||
if (o == null || getClass() != o.getClass()) {
|
||||
return false;
|
||||
}
|
||||
EmbeddingResult that = (EmbeddingResult) o;
|
||||
return Objects.equal(id, that.id);
|
||||
}
|
||||
|
||||
@Override
|
||||
public int hashCode() {
|
||||
return Objects.hashCode(id);
|
||||
}
|
||||
}
|
||||
@@ -1,56 +0,0 @@
|
||||
package com.tencent.supersonic.chat.core.knowledge;
|
||||
|
||||
import java.util.List;
|
||||
|
||||
public interface FileHandler {
|
||||
|
||||
/**
|
||||
* backup files to a specific directory
|
||||
* config: dict.directory.backup
|
||||
*
|
||||
* @param fileName
|
||||
*/
|
||||
void backupFile(String fileName);
|
||||
|
||||
/**
|
||||
* create a directory
|
||||
*
|
||||
* @param path
|
||||
*/
|
||||
void createDir(String path);
|
||||
|
||||
Boolean existPath(String path);
|
||||
|
||||
/**
|
||||
* write data to a specific file,
|
||||
* config dir: dict.directory.latest
|
||||
*
|
||||
* @param data
|
||||
* @param fileName
|
||||
* @param append
|
||||
*/
|
||||
void writeFile(List<String> data, String fileName, Boolean append);
|
||||
|
||||
/**
|
||||
* get the knowledge file root directory
|
||||
*
|
||||
* @return
|
||||
*/
|
||||
String getDictRootPath();
|
||||
|
||||
/**
|
||||
* delete dictionary file
|
||||
* automatic backup
|
||||
*
|
||||
* @param fileName
|
||||
* @return
|
||||
*/
|
||||
Boolean deleteDictFile(String fileName);
|
||||
|
||||
/**
|
||||
* delete files directly without backup
|
||||
*
|
||||
* @param fileName
|
||||
*/
|
||||
void deleteFile(String fileName);
|
||||
}
|
||||
@@ -1,32 +0,0 @@
|
||||
package com.tencent.supersonic.chat.core.knowledge;
|
||||
|
||||
import com.hankcs.hanlp.corpus.io.IIOAdapter;
|
||||
import java.io.IOException;
|
||||
import java.io.InputStream;
|
||||
import java.io.OutputStream;
|
||||
import java.net.URI;
|
||||
|
||||
import lombok.extern.slf4j.Slf4j;
|
||||
import org.apache.hadoop.conf.Configuration;
|
||||
import org.apache.hadoop.fs.FileSystem;
|
||||
import org.apache.hadoop.fs.Path;
|
||||
|
||||
@Slf4j
|
||||
public class HadoopFileIOAdapter implements IIOAdapter {
|
||||
|
||||
@Override
|
||||
public InputStream open(String path) throws IOException {
|
||||
log.info("open:{}", path);
|
||||
Configuration conf = new Configuration();
|
||||
FileSystem fs = FileSystem.get(URI.create(path), conf);
|
||||
return fs.open(new Path(path));
|
||||
}
|
||||
|
||||
@Override
|
||||
public OutputStream create(String path) throws IOException {
|
||||
log.info("create:{}", path);
|
||||
Configuration conf = new Configuration();
|
||||
FileSystem fs = FileSystem.get(URI.create(path), conf);
|
||||
return fs.create(new Path(path));
|
||||
}
|
||||
}
|
||||
@@ -1,44 +0,0 @@
|
||||
package com.tencent.supersonic.chat.core.knowledge;
|
||||
|
||||
import com.google.common.base.Objects;
|
||||
import java.util.List;
|
||||
import lombok.Data;
|
||||
import lombok.ToString;
|
||||
|
||||
@Data
|
||||
@ToString
|
||||
public class HanlpMapResult extends MapResult {
|
||||
|
||||
private List<String> natures;
|
||||
private int offset = 0;
|
||||
|
||||
private double similarity;
|
||||
|
||||
public HanlpMapResult(String name, List<String> natures, String detectWord) {
|
||||
this.name = name;
|
||||
this.natures = natures;
|
||||
this.detectWord = detectWord;
|
||||
}
|
||||
|
||||
@Override
|
||||
public boolean equals(Object o) {
|
||||
if (this == o) {
|
||||
return true;
|
||||
}
|
||||
if (o == null || getClass() != o.getClass()) {
|
||||
return false;
|
||||
}
|
||||
HanlpMapResult hanlpMapResult = (HanlpMapResult) o;
|
||||
return Objects.equal(name, hanlpMapResult.name) && Objects.equal(natures, hanlpMapResult.natures);
|
||||
}
|
||||
|
||||
@Override
|
||||
public int hashCode() {
|
||||
return Objects.hashCode(name, natures);
|
||||
}
|
||||
|
||||
public void setOffset(int offset) {
|
||||
this.offset = offset;
|
||||
}
|
||||
|
||||
}
|
||||
@@ -1,55 +0,0 @@
|
||||
package com.tencent.supersonic.chat.core.knowledge;
|
||||
|
||||
import com.tencent.supersonic.chat.core.utils.NatureHelper;
|
||||
import lombok.Data;
|
||||
import org.apache.commons.lang3.StringUtils;
|
||||
import org.springframework.beans.factory.annotation.Value;
|
||||
import org.springframework.stereotype.Service;
|
||||
import org.springframework.util.CollectionUtils;
|
||||
|
||||
import java.util.ArrayList;
|
||||
import java.util.List;
|
||||
import java.util.Objects;
|
||||
import java.util.Set;
|
||||
|
||||
@Data
|
||||
@Service
|
||||
public class LoadRemoveService {
|
||||
|
||||
@Value("${mapper.remove.agentId:}")
|
||||
private Integer mapperRemoveAgentId;
|
||||
|
||||
@Value("${mapper.remove.nature.prefix:}")
|
||||
private String mapperRemoveNaturePrefix;
|
||||
|
||||
public List removeNatures(List value, Integer agentId, Set<Long> detectModelIds) {
|
||||
if (CollectionUtils.isEmpty(value)) {
|
||||
return value;
|
||||
}
|
||||
List<String> resultList = new ArrayList<>(value);
|
||||
if (!CollectionUtils.isEmpty(detectModelIds)) {
|
||||
resultList.removeIf(nature -> {
|
||||
if (Objects.isNull(nature)) {
|
||||
return false;
|
||||
}
|
||||
Long modelId = NatureHelper.getViewId(nature);
|
||||
if (Objects.nonNull(modelId)) {
|
||||
return !detectModelIds.contains(modelId);
|
||||
}
|
||||
return false;
|
||||
});
|
||||
}
|
||||
if (Objects.nonNull(mapperRemoveAgentId)
|
||||
&& mapperRemoveAgentId.equals(agentId)
|
||||
&& StringUtils.isNotBlank(mapperRemoveNaturePrefix)) {
|
||||
resultList.removeIf(nature -> {
|
||||
if (Objects.isNull(nature)) {
|
||||
return false;
|
||||
}
|
||||
return nature.startsWith(mapperRemoveNaturePrefix);
|
||||
});
|
||||
}
|
||||
return resultList;
|
||||
}
|
||||
|
||||
}
|
||||
@@ -1,127 +0,0 @@
|
||||
package com.tencent.supersonic.chat.core.knowledge;
|
||||
|
||||
|
||||
import com.tencent.supersonic.chat.core.config.LocalFileConfig;
|
||||
import java.io.BufferedWriter;
|
||||
import java.io.IOException;
|
||||
import java.nio.charset.StandardCharsets;
|
||||
import java.nio.file.Files;
|
||||
import java.nio.file.Path;
|
||||
import java.nio.file.Paths;
|
||||
import java.nio.file.StandardCopyOption;
|
||||
import java.nio.file.StandardOpenOption;
|
||||
import java.util.List;
|
||||
import lombok.extern.slf4j.Slf4j;
|
||||
import org.springframework.stereotype.Component;
|
||||
import org.springframework.util.CollectionUtils;
|
||||
|
||||
@Slf4j
|
||||
@Component
|
||||
public class LocalFileHandler implements FileHandler {
|
||||
|
||||
private final LocalFileConfig localFileConfig;
|
||||
|
||||
public LocalFileHandler(LocalFileConfig localFileConfig) {
|
||||
this.localFileConfig = localFileConfig;
|
||||
}
|
||||
|
||||
@Override
|
||||
public void backupFile(String fileName) {
|
||||
String dictDirectoryBackup = localFileConfig.getDictDirectoryBackup();
|
||||
if (!existPath(dictDirectoryBackup)) {
|
||||
createDir(dictDirectoryBackup);
|
||||
}
|
||||
|
||||
String source = localFileConfig.getDictDirectoryLatest() + "/" + fileName;
|
||||
String target = dictDirectoryBackup + "/" + fileName;
|
||||
Path sourcePath = Paths.get(source);
|
||||
Path targetPath = Paths.get(target);
|
||||
try {
|
||||
Files.copy(sourcePath, targetPath, StandardCopyOption.REPLACE_EXISTING);
|
||||
log.info("backupFile successfully! path:{}", targetPath.toAbsolutePath());
|
||||
} catch (IOException e) {
|
||||
log.info("Failed to copy file: " + e.getMessage());
|
||||
}
|
||||
|
||||
}
|
||||
|
||||
@Override
|
||||
public void createDir(String directoryPath) {
|
||||
Path path = Paths.get(directoryPath);
|
||||
try {
|
||||
Files.createDirectories(path);
|
||||
log.info("Directory created successfully!");
|
||||
} catch (IOException e) {
|
||||
log.info("Failed to create directory: " + e.getMessage());
|
||||
}
|
||||
}
|
||||
|
||||
@Override
|
||||
public void deleteFile(String filePath) {
|
||||
Path path = Paths.get(filePath);
|
||||
try {
|
||||
Files.delete(path);
|
||||
log.info("File:{} deleted successfully!", getAbsolutePath(filePath));
|
||||
} catch (IOException e) {
|
||||
log.warn("Failed to delete file:{}, e:", getAbsolutePath(filePath), e);
|
||||
}
|
||||
}
|
||||
|
||||
@Override
|
||||
public Boolean existPath(String pathStr) {
|
||||
Path path = Paths.get(pathStr);
|
||||
if (Files.exists(path)) {
|
||||
log.info("path:{} exists!", getAbsolutePath(pathStr));
|
||||
return true;
|
||||
} else {
|
||||
log.info("path:{} not exists!", getAbsolutePath(pathStr));
|
||||
}
|
||||
return false;
|
||||
}
|
||||
|
||||
@Override
|
||||
public void writeFile(List<String> lines, String fileName, Boolean append) {
|
||||
String dictDirectoryLatest = localFileConfig.getDictDirectoryLatest();
|
||||
if (!existPath(dictDirectoryLatest)) {
|
||||
createDir(dictDirectoryLatest);
|
||||
}
|
||||
String filePath = dictDirectoryLatest + "/" + fileName;
|
||||
if (existPath(filePath)) {
|
||||
backupFile(fileName);
|
||||
}
|
||||
try (BufferedWriter writer = getWriter(filePath, append)) {
|
||||
if (!CollectionUtils.isEmpty(lines)) {
|
||||
for (String line : lines) {
|
||||
writer.write(line);
|
||||
writer.newLine();
|
||||
}
|
||||
}
|
||||
log.info("File:{} written successfully!", getAbsolutePath(filePath));
|
||||
} catch (IOException e) {
|
||||
log.info("Failed to write file:{}, e:", getAbsolutePath(filePath), e);
|
||||
}
|
||||
}
|
||||
|
||||
public String getAbsolutePath(String path) {
|
||||
return Paths.get(path).toAbsolutePath().toString();
|
||||
}
|
||||
|
||||
@Override
|
||||
public String getDictRootPath() {
|
||||
return Paths.get(localFileConfig.getDictDirectoryLatest()).toAbsolutePath().toString();
|
||||
}
|
||||
|
||||
@Override
|
||||
public Boolean deleteDictFile(String fileName) {
|
||||
backupFile(fileName);
|
||||
deleteFile(localFileConfig.getDictDirectoryLatest() + "/" + fileName);
|
||||
return true;
|
||||
}
|
||||
|
||||
private BufferedWriter getWriter(String filePath, Boolean append) throws IOException {
|
||||
if (append) {
|
||||
return Files.newBufferedWriter(Paths.get(filePath), StandardCharsets.UTF_8, StandardOpenOption.APPEND);
|
||||
}
|
||||
return Files.newBufferedWriter(Paths.get(filePath), StandardCharsets.UTF_8);
|
||||
}
|
||||
}
|
||||
@@ -1,13 +0,0 @@
|
||||
package com.tencent.supersonic.chat.core.knowledge;
|
||||
|
||||
import java.io.Serializable;
|
||||
import lombok.Data;
|
||||
import lombok.ToString;
|
||||
|
||||
@Data
|
||||
@ToString
|
||||
public class MapResult implements Serializable {
|
||||
|
||||
protected String name;
|
||||
protected String detectWord;
|
||||
}
|
||||
@@ -1,396 +0,0 @@
|
||||
package com.tencent.supersonic.chat.core.knowledge;
|
||||
|
||||
import static com.hankcs.hanlp.utility.Predefine.logger;
|
||||
|
||||
import com.hankcs.hanlp.HanLP;
|
||||
import com.hankcs.hanlp.collection.trie.DoubleArrayTrie;
|
||||
import com.hankcs.hanlp.collection.trie.bintrie.BinTrie;
|
||||
import com.hankcs.hanlp.corpus.io.ByteArray;
|
||||
import com.hankcs.hanlp.corpus.io.IOUtil;
|
||||
import com.hankcs.hanlp.corpus.tag.Nature;
|
||||
import com.hankcs.hanlp.dictionary.CoreDictionary;
|
||||
import com.hankcs.hanlp.dictionary.DynamicCustomDictionary;
|
||||
import com.hankcs.hanlp.dictionary.other.CharTable;
|
||||
import com.hankcs.hanlp.seg.common.Term;
|
||||
import com.hankcs.hanlp.utility.LexiconUtility;
|
||||
import com.hankcs.hanlp.utility.Predefine;
|
||||
import com.hankcs.hanlp.utility.TextUtility;
|
||||
import com.tencent.supersonic.chat.core.knowledge.dictionary.DictionaryAttributeUtil;
|
||||
import com.tencent.supersonic.chat.core.utils.HanlpHelper;
|
||||
import java.io.BufferedOutputStream;
|
||||
import java.io.BufferedReader;
|
||||
import java.io.DataOutputStream;
|
||||
import java.io.File;
|
||||
import java.io.FileNotFoundException;
|
||||
import java.io.IOException;
|
||||
import java.io.InputStreamReader;
|
||||
import java.util.Arrays;
|
||||
import java.util.Comparator;
|
||||
import java.util.LinkedHashSet;
|
||||
import java.util.LinkedList;
|
||||
import java.util.List;
|
||||
import java.util.Map;
|
||||
import java.util.Objects;
|
||||
import java.util.PriorityQueue;
|
||||
import java.util.TreeMap;
|
||||
import java.util.concurrent.ConcurrentHashMap;
|
||||
|
||||
public class MultiCustomDictionary extends DynamicCustomDictionary {
|
||||
|
||||
public static int MAX_SIZE = 10;
|
||||
public static Boolean removeDuplicates = true;
|
||||
public static ConcurrentHashMap<String, PriorityQueue<Term>> NATURE_TO_VALUES = new ConcurrentHashMap<>();
|
||||
private static boolean addToSuggesterTrie = true;
|
||||
|
||||
public MultiCustomDictionary() {
|
||||
this(HanLP.Config.CustomDictionaryPath);
|
||||
}
|
||||
|
||||
public MultiCustomDictionary(String... path) {
|
||||
super(path);
|
||||
}
|
||||
|
||||
/***
|
||||
* load dictionary
|
||||
* @param path
|
||||
* @param defaultNature
|
||||
* @param map
|
||||
* @param customNatureCollector
|
||||
* @param addToSuggeterTrie
|
||||
* @return
|
||||
*/
|
||||
public static boolean load(String path, Nature defaultNature, TreeMap<String, CoreDictionary.Attribute> map,
|
||||
LinkedHashSet<Nature> customNatureCollector, boolean addToSuggeterTrie) {
|
||||
try {
|
||||
String splitter = "\\s";
|
||||
if (path.endsWith(".csv")) {
|
||||
splitter = ",";
|
||||
}
|
||||
|
||||
BufferedReader br = new BufferedReader(new InputStreamReader(IOUtil.newInputStream(path), "UTF-8"));
|
||||
boolean firstLine = true;
|
||||
|
||||
while (true) {
|
||||
String[] param;
|
||||
do {
|
||||
String line;
|
||||
if ((line = br.readLine()) == null) {
|
||||
br.close();
|
||||
return true;
|
||||
}
|
||||
|
||||
if (firstLine) {
|
||||
line = IOUtil.removeUTF8BOM(line);
|
||||
firstLine = false;
|
||||
}
|
||||
|
||||
param = line.split(splitter);
|
||||
} while (param[0].length() == 0);
|
||||
|
||||
if (HanLP.Config.Normalization) {
|
||||
param[0] = CharTable.convert(param[0]);
|
||||
}
|
||||
|
||||
int natureCount = (param.length - 1) / 2;
|
||||
CoreDictionary.Attribute attribute;
|
||||
boolean isLetters = isLetters(param[0]);
|
||||
String original = null;
|
||||
String word = getWordBySpace(param[0]);
|
||||
if (isLetters) {
|
||||
original = word;
|
||||
word = word.toLowerCase();
|
||||
}
|
||||
if (natureCount == 0) {
|
||||
attribute = new CoreDictionary.Attribute(defaultNature);
|
||||
} else {
|
||||
attribute = new CoreDictionary.Attribute(natureCount);
|
||||
|
||||
for (int i = 0; i < natureCount; ++i) {
|
||||
attribute.nature[i] = LexiconUtility.convertStringToNature(param[1 + 2 * i],
|
||||
customNatureCollector);
|
||||
attribute.frequency[i] = Integer.parseInt(param[2 + 2 * i]);
|
||||
attribute.totalFrequency += attribute.frequency[i];
|
||||
}
|
||||
}
|
||||
attribute.original = original;
|
||||
|
||||
if (removeDuplicates && map.containsKey(word)) {
|
||||
attribute = DictionaryAttributeUtil.getAttribute(map.get(word), attribute);
|
||||
}
|
||||
map.put(word, attribute);
|
||||
if (addToSuggeterTrie) {
|
||||
SearchService.put(word, attribute);
|
||||
}
|
||||
for (int i = 0; i < attribute.nature.length; i++) {
|
||||
Nature nature = attribute.nature[i];
|
||||
PriorityQueue<Term> priorityQueue = NATURE_TO_VALUES.get(nature.toString());
|
||||
if (Objects.isNull(priorityQueue)) {
|
||||
priorityQueue = new PriorityQueue<>(MAX_SIZE,
|
||||
Comparator.comparingInt(Term::getFrequency).reversed());
|
||||
NATURE_TO_VALUES.put(nature.toString(), priorityQueue);
|
||||
}
|
||||
Term term = new Term(word, nature);
|
||||
term.setFrequency(attribute.frequency[i]);
|
||||
if (!priorityQueue.contains(term) && priorityQueue.size() < MAX_SIZE) {
|
||||
priorityQueue.add(term);
|
||||
}
|
||||
}
|
||||
}
|
||||
} catch (Exception var12) {
|
||||
logger.severe("自定义词典" + path + "读取错误！" + var12);
|
||||
return false;
|
||||
}
|
||||
}
|
||||
|
||||
public boolean load(String... path) {
|
||||
this.path = path;
|
||||
long start = System.currentTimeMillis();
|
||||
if (!this.loadMainDictionary(path[0])) {
|
||||
Predefine.logger.warning("自定义词典" + Arrays.toString(path) + "加载失败");
|
||||
return false;
|
||||
} else {
|
||||
Predefine.logger.info(
|
||||
"自定义词典加载成功:" + this.dat.size() + "个词条，耗时" + (System.currentTimeMillis() - start) + "ms");
|
||||
this.path = path;
|
||||
return true;
|
||||
}
|
||||
}
|
||||
|
||||
/***
|
||||
* load main dictionary
|
||||
* @param mainPath
|
||||
* @param path
|
||||
* @param dat
|
||||
* @param isCache
|
||||
* @param addToSuggestTrie
|
||||
* @return
|
||||
*/
|
||||
public static boolean loadMainDictionary(String mainPath, String[] path,
|
||||
DoubleArrayTrie<CoreDictionary.Attribute> dat, boolean isCache, boolean addToSuggestTrie) {
|
||||
Predefine.logger.info("自定义词典开始加载:" + mainPath);
|
||||
if (loadDat(mainPath, dat)) {
|
||||
return true;
|
||||
} else {
|
||||
TreeMap<String, CoreDictionary.Attribute> map = new TreeMap();
|
||||
LinkedHashSet customNatureCollector = new LinkedHashSet();
|
||||
|
||||
try {
|
||||
for (String p : path) {
|
||||
Nature defaultNature = Nature.n;
|
||||
File file = new File(p);
|
||||
String fileName = file.getName();
|
||||
int cut = fileName.lastIndexOf(32);
|
||||
if (cut > 0) {
|
||||
String nature = fileName.substring(cut + 1);
|
||||
p = file.getParent() + File.separator + fileName.substring(0, cut);
|
||||
|
||||
try {
|
||||
defaultNature = LexiconUtility.convertStringToNature(nature, customNatureCollector);
|
||||
} catch (Exception var16) {
|
||||
Predefine.logger.severe("配置文件【" + p + "】写错了！" + var16);
|
||||
continue;
|
||||
}
|
||||
}
|
||||
|
||||
Predefine.logger.info("以默认词性[" + defaultNature + "]加载自定义词典" + p + "中……");
|
||||
boolean success = load(p, defaultNature, map, customNatureCollector, addToSuggestTrie);
|
||||
if (!success) {
|
||||
Predefine.logger.warning("失败：" + p);
|
||||
}
|
||||
}
|
||||
|
||||
if (map.size() == 0) {
|
||||
Predefine.logger.warning("没有加载到任何词条");
|
||||
map.put("未##它", null);
|
||||
}
|
||||
|
||||
logger.info("正在构建DoubleArrayTrie……");
|
||||
dat.build(map);
|
||||
if (addToSuggestTrie) {
|
||||
// SearchService.save();
|
||||
}
|
||||
if (isCache) {
|
||||
// 缓存成dat文件，下次加载会快很多
|
||||
logger.info("正在缓存词典为dat文件……");
|
||||
// 缓存值文件
|
||||
List<CoreDictionary.Attribute> attributeList = new LinkedList<CoreDictionary.Attribute>();
|
||||
for (Map.Entry<String, CoreDictionary.Attribute> entry : map.entrySet()) {
|
||||
attributeList.add(entry.getValue());
|
||||
}
|
||||
|
||||
DataOutputStream out = new DataOutputStream(
|
||||
new BufferedOutputStream(IOUtil.newOutputStream(mainPath + ".bin")));
|
||||
if (customNatureCollector.isEmpty()) {
|
||||
for (int i = Nature.begin.ordinal() + 1; i < Nature.values().length; ++i) {
|
||||
Nature nature = Nature.values()[i];
|
||||
if (Objects.nonNull(nature)) {
|
||||
customNatureCollector.add(nature);
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
IOUtil.writeCustomNature(out, customNatureCollector);
|
||||
out.writeInt(attributeList.size());
|
||||
|
||||
for (CoreDictionary.Attribute attribute : attributeList) {
|
||||
attribute.save(out);
|
||||
}
|
||||
|
||||
dat.save(out);
|
||||
out.close();
|
||||
}
|
||||
} catch (FileNotFoundException var17) {
|
||||
logger.severe("自定义词典" + mainPath + "不存在！" + var17);
|
||||
return false;
|
||||
} catch (IOException var18) {
|
||||
logger.severe("自定义词典" + mainPath + "读取错误！" + var18);
|
||||
return false;
|
||||
} catch (Exception var19) {
|
||||
logger.warning("自定义词典" + mainPath + "缓存失败！\n" + TextUtility.exceptionToString(var19));
|
||||
}
|
||||
|
||||
return true;
|
||||
}
|
||||
}
|
||||
|
||||
public boolean loadMainDictionary(String mainPath) {
|
||||
return loadMainDictionary(mainPath, this.path, this.dat, true, addToSuggesterTrie);
|
||||
}
|
||||
|
||||
public static boolean loadDat(String path, DoubleArrayTrie<CoreDictionary.Attribute> dat) {
|
||||
return loadDat(path, HanLP.Config.CustomDictionaryPath, dat);
|
||||
}
|
||||
|
||||
public static boolean loadDat(String path, String[] customDicPath, DoubleArrayTrie<CoreDictionary.Attribute> dat) {
|
||||
try {
|
||||
if (HanLP.Config.CustomDictionaryAutoRefreshCache && isDicNeedUpdate(path, customDicPath)) {
|
||||
return false;
|
||||
} else {
|
||||
ByteArray byteArray = ByteArray.createByteArray(path + ".bin");
|
||||
if (byteArray == null) {
|
||||
return false;
|
||||
} else {
|
||||
int size = byteArray.nextInt();
|
||||
if (size < 0) {
|
||||
while (true) {
|
||||
++size;
|
||||
if (size > 0) {
|
||||
size = byteArray.nextInt();
|
||||
break;
|
||||
}
|
||||
|
||||
Nature.create(byteArray.nextString());
|
||||
}
|
||||
}
|
||||
|
||||
CoreDictionary.Attribute[] attributes = new CoreDictionary.Attribute[size];
|
||||
Nature[] natureIndexArray = Nature.values();
|
||||
|
||||
for (int i = 0; i < size; ++i) {
|
||||
int currentTotalFrequency = byteArray.nextInt();
|
||||
int length = byteArray.nextInt();
|
||||
attributes[i] = new CoreDictionary.Attribute(length);
|
||||
attributes[i].totalFrequency = currentTotalFrequency;
|
||||
|
||||
for (int j = 0; j < length; ++j) {
|
||||
attributes[i].nature[j] = natureIndexArray[byteArray.nextInt()];
|
||||
attributes[i].frequency[j] = byteArray.nextInt();
|
||||
}
|
||||
}
|
||||
|
||||
if (!dat.load(byteArray, attributes)) {
|
||||
return false;
|
||||
} else {
|
||||
return true;
|
||||
}
|
||||
}
|
||||
}
|
||||
} catch (Exception var11) {
|
||||
logger.warning("读取失败，问题发生在" + TextUtility.exceptionToString(var11));
|
||||
return false;
|
||||
}
|
||||
}
|
||||
|
||||
public static boolean isLetters(String str) {
|
||||
char[] chars = str.toCharArray();
|
||||
if (chars.length <= 1) {
|
||||
return false;
|
||||
}
|
||||
for (int i = 0; i < chars.length; i++) {
|
||||
if ((chars[i] >= 'A' && chars[i] <= 'Z')) {
|
||||
return true;
|
||||
}
|
||||
}
|
||||
return false;
|
||||
}
|
||||
|
||||
public static boolean isLowerLetter(String str) {
|
||||
char[] chars = str.toCharArray();
|
||||
for (int i = 0; i < chars.length; i++) {
|
||||
if ((chars[i] >= 'a' && chars[i] <= 'z')) {
|
||||
return true;
|
||||
}
|
||||
}
|
||||
return false;
|
||||
}
|
||||
|
||||
public static String getWordBySpace(String word) {
|
||||
if (word.contains(HanlpHelper.SPACE_SPILT)) {
|
||||
return word.replace(HanlpHelper.SPACE_SPILT, " ");
|
||||
}
|
||||
return word;
|
||||
}
|
||||
|
||||
public boolean reload() {
|
||||
if (this.path != null && this.path.length != 0) {
|
||||
IOUtil.deleteFile(this.path[0] + ".bin");
|
||||
Boolean loadCacheOk = this.loadDat(this.path[0], this.path, this.dat);
|
||||
if (!loadCacheOk) {
|
||||
return this.loadMainDictionary(this.path[0], this.path, this.dat, true, addToSuggesterTrie);
|
||||
}
|
||||
}
|
||||
return false;
|
||||
|
||||
}
|
||||
|
||||
public synchronized boolean insert(String word, String natureWithFrequency) {
|
||||
if (word == null) {
|
||||
return false;
|
||||
} else {
|
||||
if (HanLP.Config.Normalization) {
|
||||
word = CharTable.convert(word);
|
||||
}
|
||||
CoreDictionary.Attribute att = natureWithFrequency == null ? new CoreDictionary.Attribute(Nature.nz, 1)
|
||||
: CoreDictionary.Attribute.create(natureWithFrequency);
|
||||
boolean isLetters = isLetters(word);
|
||||
word = getWordBySpace(word);
|
||||
String original = null;
|
||||
if (isLetters) {
|
||||
original = word;
|
||||
word = word.toLowerCase();
|
||||
}
|
||||
if (att == null) {
|
||||
return false;
|
||||
} else if (this.dat.containsKey(word)) {
|
||||
att.original = original;
|
||||
att = DictionaryAttributeUtil.getAttribute(this.dat.get(word), att);
|
||||
this.dat.set(word, att);
|
||||
// return true;
|
||||
} else {
|
||||
if (this.trie == null) {
|
||||
this.trie = new BinTrie();
|
||||
}
|
||||
att.original = original;
|
||||
if (this.trie.containsKey(word)) {
|
||||
att = DictionaryAttributeUtil.getAttribute(this.trie.get(word), att);
|
||||
}
|
||||
this.trie.put(word, att);
|
||||
// return true;
|
||||
}
|
||||
if (addToSuggesterTrie) {
|
||||
SearchService.put(word, att);
|
||||
}
|
||||
return true;
|
||||
}
|
||||
}
|
||||
}
|
||||
@@ -1,170 +0,0 @@
|
||||
package com.tencent.supersonic.chat.core.knowledge;
|
||||
|
||||
import com.hankcs.hanlp.collection.trie.bintrie.BaseNode;
|
||||
import com.hankcs.hanlp.collection.trie.bintrie.BinTrie;
|
||||
import com.hankcs.hanlp.corpus.tag.Nature;
|
||||
import com.hankcs.hanlp.dictionary.CoreDictionary;
|
||||
import com.hankcs.hanlp.seg.common.Term;
|
||||
import com.tencent.supersonic.chat.api.pojo.request.DimensionValueReq;
|
||||
import com.tencent.supersonic.chat.core.knowledge.dictionary.DictionaryAttributeUtil;
|
||||
import com.tencent.supersonic.common.pojo.enums.DictWordType;
|
||||
import java.util.ArrayList;
|
||||
import java.util.Arrays;
|
||||
import java.util.List;
|
||||
import java.util.Map;
|
||||
import java.util.Objects;
|
||||
import java.util.PriorityQueue;
|
||||
import java.util.Set;
|
||||
import java.util.TreeMap;
|
||||
import java.util.TreeSet;
|
||||
import java.util.stream.Collectors;
|
||||
import lombok.extern.slf4j.Slf4j;
|
||||
import org.apache.commons.lang3.StringUtils;
|
||||
import org.springframework.util.CollectionUtils;
|
||||
|
||||
@Slf4j
|
||||
public class SearchService {
|
||||
|
||||
public static final int SEARCH_SIZE = 200;
|
||||
private static BinTrie<List<String>> trie;
|
||||
private static BinTrie<List<String>> suffixTrie;
|
||||
|
||||
static {
|
||||
trie = new BinTrie<>();
|
||||
suffixTrie = new BinTrie<>();
|
||||
}
|
||||
|
||||
/***
|
||||
* prefix Search
|
||||
* @param key
|
||||
* @return
|
||||
*/
|
||||
public static List<HanlpMapResult> prefixSearch(String key, int limit, Integer agentId, Set<Long> detectModelIds) {
|
||||
return prefixSearch(key, limit, agentId, trie, detectModelIds);
|
||||
}
|
||||
|
||||
public static List<HanlpMapResult> prefixSearch(String key, int limit, Integer agentId,
|
||||
BinTrie<List<String>> binTrie, Set<Long> detectModelIds) {
|
||||
Set<Map.Entry<String, List<String>>> result = prefixSearchLimit(key, limit, binTrie, agentId, detectModelIds);
|
||||
return result.stream().map(
|
||||
entry -> {
|
||||
String name = entry.getKey().replace("#", " ");
|
||||
return new HanlpMapResult(name, entry.getValue(), key);
|
||||
}
|
||||
).sorted((a, b) -> -(b.getName().length() - a.getName().length()))
|
||||
.limit(SEARCH_SIZE)
|
||||
.collect(Collectors.toList());
|
||||
}
|
||||
|
||||
/***
|
||||
* suffix Search
|
||||
* @param key
|
||||
* @return
|
||||
*/
|
||||
public static List<HanlpMapResult> suffixSearch(String key, int limit, Integer agentId, Set<Long> detectModelIds) {
|
||||
String reverseDetectSegment = StringUtils.reverse(key);
|
||||
return suffixSearch(reverseDetectSegment, limit, agentId, suffixTrie, detectModelIds);
|
||||
}
|
||||
|
||||
public static List<HanlpMapResult> suffixSearch(String key, int limit, Integer agentId,
|
||||
BinTrie<List<String>> binTrie, Set<Long> detectModelIds) {
|
||||
Set<Map.Entry<String, List<String>>> result = prefixSearchLimit(key, limit, binTrie, agentId, detectModelIds);
|
||||
return result.stream().map(
|
||||
entry -> {
|
||||
String name = entry.getKey().replace("#", " ");
|
||||
List<String> natures = entry.getValue().stream()
|
||||
.map(nature -> nature.replaceAll(DictWordType.SUFFIX.getType(), ""))
|
||||
.collect(Collectors.toList());
|
||||
name = StringUtils.reverse(name);
|
||||
return new HanlpMapResult(name, natures, key);
|
||||
}
|
||||
).sorted((a, b) -> -(b.getName().length() - a.getName().length()))
|
||||
.limit(SEARCH_SIZE)
|
||||
.collect(Collectors.toList());
|
||||
}
|
||||
|
||||
private static Set<Map.Entry<String, List<String>>> prefixSearchLimit(String key, int limit,
|
||||
BinTrie<List<String>> binTrie, Integer agentId, Set<Long> detectModelIds) {
|
||||
key = key.toLowerCase();
|
||||
Set<Map.Entry<String, List<String>>> entrySet = new TreeSet<Map.Entry<String, List<String>>>();
|
||||
|
||||
StringBuilder sb = new StringBuilder();
|
||||
if (StringUtils.isNotBlank(key)) {
|
||||
sb = new StringBuilder(key.substring(0, key.length() - 1));
|
||||
}
|
||||
BaseNode branch = binTrie;
|
||||
char[] chars = key.toCharArray();
|
||||
for (char aChar : chars) {
|
||||
if (branch == null) {
|
||||
return entrySet;
|
||||
}
|
||||
branch = branch.getChild(aChar);
|
||||
}
|
||||
|
||||
if (branch == null) {
|
||||
return entrySet;
|
||||
}
|
||||
branch.walkLimit(sb, entrySet, limit, agentId, detectModelIds);
|
||||
return entrySet;
|
||||
}
|
||||
|
||||
public static void clear() {
|
||||
log.info("clear all trie");
|
||||
trie = new BinTrie<>();
|
||||
suffixTrie = new BinTrie<>();
|
||||
}
|
||||
|
||||
public static void put(String key, CoreDictionary.Attribute attribute) {
|
||||
trie.put(key, getValue(attribute.nature));
|
||||
}
|
||||
|
||||
public static void loadSuffix(List<DictWord> suffixes) {
|
||||
if (CollectionUtils.isEmpty(suffixes)) {
|
||||
return;
|
||||
}
|
||||
TreeMap<String, CoreDictionary.Attribute> map = new TreeMap();
|
||||
for (DictWord suffix : suffixes) {
|
||||
CoreDictionary.Attribute attributeNew = suffix.getNatureWithFrequency() == null
|
||||
? new CoreDictionary.Attribute(Nature.nz, 1)
|
||||
: CoreDictionary.Attribute.create(suffix.getNatureWithFrequency());
|
||||
if (map.containsKey(suffix.getWord())) {
|
||||
attributeNew = DictionaryAttributeUtil.getAttribute(map.get(suffix.getWord()), attributeNew);
|
||||
}
|
||||
map.put(suffix.getWord(), attributeNew);
|
||||
}
|
||||
for (Map.Entry<String, CoreDictionary.Attribute> stringAttributeEntry : map.entrySet()) {
|
||||
putSuffix(stringAttributeEntry.getKey(), stringAttributeEntry.getValue());
|
||||
}
|
||||
}
|
||||
|
||||
public static void putSuffix(String key, CoreDictionary.Attribute attribute) {
|
||||
Nature[] nature = attribute.nature;
|
||||
suffixTrie.put(key, getValue(nature));
|
||||
}
|
||||
|
||||
private static List<String> getValue(Nature[] nature) {
|
||||
return Arrays.stream(nature).map(entry -> entry.toString()).collect(Collectors.toList());
|
||||
}
|
||||
|
||||
public static void remove(DictWord dictWord, Nature[] natures) {
|
||||
trie.remove(dictWord.getWord());
|
||||
if (Objects.nonNull(natures) && natures.length > 0) {
|
||||
trie.put(dictWord.getWord(), getValue(natures));
|
||||
}
|
||||
if (dictWord.getNature().contains(DictWordType.METRIC.getType()) || dictWord.getNature()
|
||||
.contains(DictWordType.DIMENSION.getType())) {
|
||||
suffixTrie.remove(dictWord.getWord());
|
||||
}
|
||||
}
|
||||
|
||||
public static List<String> getDimensionValue(DimensionValueReq dimensionValueReq) {
|
||||
String nature = DictWordType.NATURE_SPILT + dimensionValueReq.getModelId() + DictWordType.NATURE_SPILT
|
||||
+ dimensionValueReq.getElementID();
|
||||
PriorityQueue<Term> terms = MultiCustomDictionary.NATURE_TO_VALUES.get(nature);
|
||||
if (org.apache.commons.collections.CollectionUtils.isEmpty(terms)) {
|
||||
return new ArrayList<>();
|
||||
}
|
||||
return terms.stream().map(term -> term.getWord()).collect(Collectors.toList());
|
||||
}
|
||||
}
|
||||
|
||||
@@ -1,22 +0,0 @@
|
||||
package com.tencent.supersonic.chat.core.knowledge;
|
||||
|
||||
import lombok.Builder;
|
||||
import lombok.Data;
|
||||
import lombok.ToString;
|
||||
|
||||
import java.io.Serializable;
|
||||
|
||||
@Data
|
||||
@ToString
|
||||
@Builder
|
||||
public class ViewInfoStat implements Serializable {
|
||||
|
||||
private long viewCount;
|
||||
|
||||
private long metricViewCount;
|
||||
|
||||
private long dimensionViewCount;
|
||||
|
||||
private long dimensionValueViewCount;
|
||||
|
||||
}
|
||||
@@ -1,38 +0,0 @@
|
||||
package com.tencent.supersonic.chat.core.knowledge.builder;
|
||||
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElement;
|
||||
import com.tencent.supersonic.chat.core.knowledge.DictWord;
|
||||
import java.util.ArrayList;
|
||||
import java.util.List;
|
||||
import lombok.extern.slf4j.Slf4j;
|
||||
|
||||
/**
|
||||
* base word nature
|
||||
*/
|
||||
@Slf4j
|
||||
public abstract class BaseWordBuilder {
|
||||
|
||||
public static final Long DEFAULT_FREQUENCY = 100000L;
|
||||
|
||||
public List<DictWord> getDictWords(List<SchemaElement> schemaElements) {
|
||||
List<DictWord> dictWords = new ArrayList<>();
|
||||
try {
|
||||
dictWords = getDictWordsWithException(schemaElements);
|
||||
} catch (Exception e) {
|
||||
log.error("getWordNatureList error,", e);
|
||||
}
|
||||
return dictWords;
|
||||
}
|
||||
|
||||
protected List<DictWord> getDictWordsWithException(List<SchemaElement> schemaElements) {
|
||||
|
||||
List<DictWord> dictWords = new ArrayList<>();
|
||||
|
||||
for (SchemaElement schemaElement : schemaElements) {
|
||||
dictWords.addAll(doGet(schemaElement.getName(), schemaElement));
|
||||
}
|
||||
return dictWords;
|
||||
}
|
||||
|
||||
protected abstract List<DictWord> doGet(String word, SchemaElement schemaElement);
|
||||
}
|
||||
@@ -1,64 +0,0 @@
|
||||
package com.tencent.supersonic.chat.core.knowledge.builder;
|
||||
|
||||
import com.google.common.collect.Lists;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElement;
|
||||
import com.tencent.supersonic.chat.core.knowledge.DictWord;
|
||||
import com.tencent.supersonic.common.pojo.enums.DictWordType;
|
||||
import org.apache.commons.lang3.StringUtils;
|
||||
import org.springframework.beans.factory.annotation.Value;
|
||||
import org.springframework.stereotype.Service;
|
||||
import org.springframework.util.CollectionUtils;
|
||||
|
||||
import java.util.ArrayList;
|
||||
import java.util.List;
|
||||
|
||||
/**
|
||||
* dimension word nature
|
||||
*/
|
||||
@Service
|
||||
public class DimensionWordBuilder extends BaseWordBuilder {
|
||||
|
||||
@Value("${nlp.dimension.use.suffix:true}")
|
||||
private boolean nlpDimensionUseSuffix = true;
|
||||
|
||||
@Override
|
||||
public List<DictWord> doGet(String word, SchemaElement schemaElement) {
|
||||
List<DictWord> result = Lists.newArrayList();
|
||||
result.add(getOnwWordNature(word, schemaElement, false));
|
||||
result.addAll(getOnwWordNatureAlias(schemaElement, false));
|
||||
if (nlpDimensionUseSuffix) {
|
||||
String reverseWord = StringUtils.reverse(word);
|
||||
if (StringUtils.isNotEmpty(word) && !word.equalsIgnoreCase(reverseWord)) {
|
||||
result.add(getOnwWordNature(reverseWord, schemaElement, true));
|
||||
}
|
||||
}
|
||||
return result;
|
||||
}
|
||||
|
||||
private DictWord getOnwWordNature(String word, SchemaElement schemaElement, boolean isSuffix) {
|
||||
DictWord dictWord = new DictWord();
|
||||
dictWord.setWord(word);
|
||||
Long viewId = schemaElement.getView();
|
||||
String nature = DictWordType.NATURE_SPILT + viewId + DictWordType.NATURE_SPILT + schemaElement.getId()
|
||||
+ DictWordType.DIMENSION.getType();
|
||||
if (isSuffix) {
|
||||
nature = DictWordType.NATURE_SPILT + viewId + DictWordType.NATURE_SPILT + schemaElement.getId()
|
||||
+ DictWordType.SUFFIX.getType() + DictWordType.DIMENSION.getType();
|
||||
}
|
||||
dictWord.setNatureWithFrequency(String.format("%s " + DEFAULT_FREQUENCY, nature));
|
||||
return dictWord;
|
||||
}
|
||||
|
||||
private List<DictWord> getOnwWordNatureAlias(SchemaElement schemaElement, boolean isSuffix) {
|
||||
List<DictWord> dictWords = new ArrayList<>();
|
||||
if (CollectionUtils.isEmpty(schemaElement.getAlias())) {
|
||||
return dictWords;
|
||||
}
|
||||
|
||||
for (String alias : schemaElement.getAlias()) {
|
||||
dictWords.add(getOnwWordNature(alias, schemaElement, false));
|
||||
}
|
||||
return dictWords;
|
||||
}
|
||||
|
||||
}
|
||||
@@ -1,45 +0,0 @@
|
||||
package com.tencent.supersonic.chat.core.knowledge.builder;
|
||||
|
||||
|
||||
import com.google.common.collect.Lists;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElement;
|
||||
import com.tencent.supersonic.chat.core.knowledge.DictWord;
|
||||
import com.tencent.supersonic.common.pojo.enums.DictWordType;
|
||||
import lombok.extern.slf4j.Slf4j;
|
||||
import org.springframework.stereotype.Service;
|
||||
import org.springframework.util.CollectionUtils;
|
||||
|
||||
import java.util.List;
|
||||
import java.util.Objects;
|
||||
|
||||
/**
|
||||
* dimension value wordNature
|
||||
*/
|
||||
@Service
|
||||
@Slf4j
|
||||
public class EntityWordBuilder extends BaseWordBuilder {
|
||||
|
||||
@Override
|
||||
public List<DictWord> doGet(String word, SchemaElement schemaElement) {
|
||||
List<DictWord> result = Lists.newArrayList();
|
||||
|
||||
if (Objects.isNull(schemaElement)) {
|
||||
return result;
|
||||
}
|
||||
|
||||
Long view = schemaElement.getView();
|
||||
String nature = DictWordType.NATURE_SPILT + view + DictWordType.NATURE_SPILT + schemaElement.getId()
|
||||
+ DictWordType.ENTITY.getType();
|
||||
|
||||
if (!CollectionUtils.isEmpty(schemaElement.getAlias())) {
|
||||
schemaElement.getAlias().stream().forEach(alias -> {
|
||||
DictWord dictWordAlias = new DictWord();
|
||||
dictWordAlias.setWord(alias);
|
||||
dictWordAlias.setNatureWithFrequency(String.format("%s " + DEFAULT_FREQUENCY * 2, nature));
|
||||
result.add(dictWordAlias);
|
||||
});
|
||||
}
|
||||
return result;
|
||||
}
|
||||
|
||||
}
|
||||
@@ -1,64 +0,0 @@
|
||||
package com.tencent.supersonic.chat.core.knowledge.builder;
|
||||
|
||||
import com.google.common.collect.Lists;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElement;
|
||||
import com.tencent.supersonic.chat.core.knowledge.DictWord;
|
||||
import com.tencent.supersonic.common.pojo.enums.DictWordType;
|
||||
import org.apache.commons.lang3.StringUtils;
|
||||
import org.springframework.beans.factory.annotation.Value;
|
||||
import org.springframework.stereotype.Service;
|
||||
import org.springframework.util.CollectionUtils;
|
||||
|
||||
import java.util.ArrayList;
|
||||
import java.util.List;
|
||||
|
||||
/**
|
||||
* Metric DictWord
|
||||
*/
|
||||
@Service
|
||||
public class MetricWordBuilder extends BaseWordBuilder {
|
||||
|
||||
@Value("${nlp.metric.use.suffix:true}")
|
||||
private boolean nlpMetricUseSuffix = true;
|
||||
|
||||
@Override
|
||||
public List<DictWord> doGet(String word, SchemaElement schemaElement) {
|
||||
List<DictWord> result = Lists.newArrayList();
|
||||
result.add(getOnwWordNature(word, schemaElement, false));
|
||||
result.addAll(getOnwWordNatureAlias(schemaElement, false));
|
||||
if (nlpMetricUseSuffix) {
|
||||
String reverseWord = StringUtils.reverse(word);
|
||||
if (!word.equalsIgnoreCase(reverseWord)) {
|
||||
result.add(getOnwWordNature(reverseWord, schemaElement, true));
|
||||
}
|
||||
}
|
||||
return result;
|
||||
}
|
||||
|
||||
private DictWord getOnwWordNature(String word, SchemaElement schemaElement, boolean isSuffix) {
|
||||
DictWord dictWord = new DictWord();
|
||||
dictWord.setWord(word);
|
||||
Long viewId = schemaElement.getView();
|
||||
String nature = DictWordType.NATURE_SPILT + viewId + DictWordType.NATURE_SPILT + schemaElement.getId()
|
||||
+ DictWordType.METRIC.getType();
|
||||
if (isSuffix) {
|
||||
nature = DictWordType.NATURE_SPILT + viewId + DictWordType.NATURE_SPILT + schemaElement.getId()
|
||||
+ DictWordType.SUFFIX.getType() + DictWordType.METRIC.getType();
|
||||
}
|
||||
dictWord.setNatureWithFrequency(String.format("%s " + DEFAULT_FREQUENCY, nature));
|
||||
return dictWord;
|
||||
}
|
||||
|
||||
private List<DictWord> getOnwWordNatureAlias(SchemaElement schemaElement, boolean isSuffix) {
|
||||
List<DictWord> dictWords = new ArrayList<>();
|
||||
if (CollectionUtils.isEmpty(schemaElement.getAlias())) {
|
||||
return dictWords;
|
||||
}
|
||||
|
||||
for (String alias : schemaElement.getAlias()) {
|
||||
dictWords.add(getOnwWordNature(alias, schemaElement, false));
|
||||
}
|
||||
return dictWords;
|
||||
}
|
||||
|
||||
}
|
||||
@@ -1,44 +0,0 @@
|
||||
package com.tencent.supersonic.chat.core.knowledge.builder;
|
||||
|
||||
import com.google.common.collect.Lists;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElement;
|
||||
import com.tencent.supersonic.chat.core.knowledge.DictWord;
|
||||
import com.tencent.supersonic.common.pojo.enums.DictWordType;
|
||||
import lombok.extern.slf4j.Slf4j;
|
||||
import org.apache.commons.collections.CollectionUtils;
|
||||
import org.springframework.stereotype.Service;
|
||||
|
||||
import java.util.List;
|
||||
|
||||
/**
|
||||
* model word nature
|
||||
*/
|
||||
@Service
|
||||
@Slf4j
|
||||
public class ModelWordBuilder extends BaseWordBuilder {
|
||||
|
||||
@Override
|
||||
public List<DictWord> doGet(String word, SchemaElement schemaElement) {
|
||||
List<DictWord> result = Lists.newArrayList();
|
||||
//modelName
|
||||
DictWord dictWord = buildDictWord(word, schemaElement.getView());
|
||||
result.add(dictWord);
|
||||
//alias
|
||||
List<String> aliasList = schemaElement.getAlias();
|
||||
if (CollectionUtils.isNotEmpty(aliasList)) {
|
||||
for (String alias : aliasList) {
|
||||
result.add(buildDictWord(alias, schemaElement.getView()));
|
||||
}
|
||||
}
|
||||
return result;
|
||||
}
|
||||
|
||||
private DictWord buildDictWord(String word, Long modelId) {
|
||||
DictWord dictWord = new DictWord();
|
||||
dictWord.setWord(word);
|
||||
String nature = DictWordType.NATURE_SPILT + modelId;
|
||||
dictWord.setNatureWithFrequency(String.format("%s " + DEFAULT_FREQUENCY, nature));
|
||||
return dictWord;
|
||||
}
|
||||
|
||||
}
|
||||
@@ -1,41 +0,0 @@
|
||||
package com.tencent.supersonic.chat.core.knowledge.builder;
|
||||
|
||||
|
||||
import com.google.common.collect.Lists;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElement;
|
||||
import com.tencent.supersonic.chat.core.knowledge.DictWord;
|
||||
import com.tencent.supersonic.common.pojo.enums.DictWordType;
|
||||
import lombok.extern.slf4j.Slf4j;
|
||||
import org.springframework.stereotype.Service;
|
||||
import org.springframework.util.CollectionUtils;
|
||||
|
||||
import java.util.List;
|
||||
import java.util.Objects;
|
||||
|
||||
/**
|
||||
* dimension value wordNature
|
||||
*/
|
||||
@Service
|
||||
@Slf4j
|
||||
public class ValueWordBuilder extends BaseWordBuilder {
|
||||
|
||||
@Override
|
||||
public List<DictWord> doGet(String word, SchemaElement schemaElement) {
|
||||
|
||||
List<DictWord> result = Lists.newArrayList();
|
||||
if (Objects.nonNull(schemaElement) && !CollectionUtils.isEmpty(schemaElement.getAlias())) {
|
||||
|
||||
schemaElement.getAlias().stream().forEach(value -> {
|
||||
DictWord dictWord = new DictWord();
|
||||
Long viewId = schemaElement.getView();
|
||||
String nature = DictWordType.NATURE_SPILT + viewId + DictWordType.NATURE_SPILT + schemaElement.getId();
|
||||
dictWord.setNatureWithFrequency(String.format("%s " + DEFAULT_FREQUENCY, nature));
|
||||
dictWord.setWord(value);
|
||||
result.add(dictWord);
|
||||
});
|
||||
}
|
||||
log.debug("ValueWordBuilder, result:{}", result);
|
||||
return result;
|
||||
}
|
||||
|
||||
}
|
||||
@@ -1,26 +0,0 @@
|
||||
package com.tencent.supersonic.chat.core.knowledge.builder;
|
||||
|
||||
|
||||
import com.tencent.supersonic.common.pojo.enums.DictWordType;
|
||||
import java.util.Map;
|
||||
import java.util.concurrent.ConcurrentHashMap;
|
||||
|
||||
/**
|
||||
* DictWord Strategy Factory
|
||||
*/
|
||||
public class WordBuilderFactory {
|
||||
|
||||
private static Map<DictWordType, BaseWordBuilder> wordNatures = new ConcurrentHashMap<>();
|
||||
|
||||
static {
|
||||
wordNatures.put(DictWordType.DIMENSION, new DimensionWordBuilder());
|
||||
wordNatures.put(DictWordType.METRIC, new MetricWordBuilder());
|
||||
wordNatures.put(DictWordType.VIEW, new ModelWordBuilder());
|
||||
wordNatures.put(DictWordType.ENTITY, new EntityWordBuilder());
|
||||
wordNatures.put(DictWordType.VALUE, new ValueWordBuilder());
|
||||
}
|
||||
|
||||
public static BaseWordBuilder get(DictWordType strategyType) {
|
||||
return wordNatures.get(strategyType);
|
||||
}
|
||||
}
|
||||
@@ -1,8 +1,8 @@
|
||||
package com.tencent.supersonic.chat.core.mapper;
|
||||
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElement;
|
||||
import com.tencent.supersonic.headless.api.pojo.SchemaElement;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElementMatch;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElementType;
|
||||
import com.tencent.supersonic.headless.api.pojo.SchemaElementType;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaMapInfo;
|
||||
import com.tencent.supersonic.chat.api.pojo.SemanticSchema;
|
||||
import com.tencent.supersonic.chat.api.pojo.ViewSchema;
|
||||
|
||||
@@ -1,8 +1,8 @@
|
||||
package com.tencent.supersonic.chat.core.mapper;
|
||||
|
||||
import com.hankcs.hanlp.seg.common.Term;
|
||||
import com.tencent.supersonic.chat.core.pojo.QueryContext;
|
||||
import com.tencent.supersonic.chat.core.utils.NatureHelper;
|
||||
import com.tencent.supersonic.headless.api.pojo.response.S2Term;
|
||||
import com.tencent.supersonic.headless.core.knowledge.helper.NatureHelper;
|
||||
import lombok.extern.slf4j.Slf4j;
|
||||
import org.apache.commons.collections4.CollectionUtils;
|
||||
import org.apache.commons.lang3.StringUtils;
|
||||
@@ -28,7 +28,8 @@ public abstract class BaseMatchStrategy<T> implements MatchStrategy<T> {
|
||||
private MapperHelper mapperHelper;
|
||||
|
||||
@Override
|
||||
public Map<MatchText, List<T>> match(QueryContext queryContext, List<Term> terms, Set<Long> detectViewIds) {
|
||||
public Map<MatchText, List<T>> match(QueryContext queryContext, List<S2Term> terms,
|
||||
Set<Long> detectViewIds) {
|
||||
String text = queryContext.getQueryText();
|
||||
if (Objects.isNull(terms) || StringUtils.isEmpty(text)) {
|
||||
return null;
|
||||
@@ -43,7 +44,7 @@ public abstract class BaseMatchStrategy<T> implements MatchStrategy<T> {
|
||||
return result;
|
||||
}
|
||||
|
||||
public List<T> detect(QueryContext queryContext, List<Term> terms, Set<Long> detectModelIds) {
|
||||
public List<T> detect(QueryContext queryContext, List<S2Term> terms, Set<Long> detectModelIds) {
|
||||
Map<Integer, Integer> regOffsetToLength = getRegOffsetToLength(terms);
|
||||
String text = queryContext.getQueryText();
|
||||
Set<T> results = new HashSet<>();
|
||||
@@ -72,9 +73,10 @@ public abstract class BaseMatchStrategy<T> implements MatchStrategy<T> {
|
||||
return;
|
||||
}
|
||||
|
||||
public Map<Integer, Integer> getRegOffsetToLength(List<Term> terms) {
|
||||
return terms.stream().sorted(Comparator.comparing(Term::length))
|
||||
.collect(Collectors.toMap(Term::getOffset, term -> term.word.length(), (value1, value2) -> value2));
|
||||
public Map<Integer, Integer> getRegOffsetToLength(List<S2Term> terms) {
|
||||
return terms.stream().sorted(Comparator.comparing(S2Term::length))
|
||||
.collect(Collectors.toMap(S2Term::getOffset, term -> term.word.length(),
|
||||
(value1, value2) -> value2));
|
||||
}
|
||||
|
||||
public void selectResultInOneRound(Set<T> existResults, List<T> oneRoundResults) {
|
||||
@@ -102,7 +104,7 @@ public abstract class BaseMatchStrategy<T> implements MatchStrategy<T> {
|
||||
}
|
||||
}
|
||||
|
||||
public List<T> getMatches(QueryContext queryContext, List<Term> terms) {
|
||||
public List<T> getMatches(QueryContext queryContext, List<S2Term> terms) {
|
||||
Set<Long> viewIds = mapperHelper.getViewIds(queryContext.getViewId(), queryContext.getAgent());
|
||||
terms = filterByViewId(terms, viewIds);
|
||||
Map<MatchText, List<T>> matchResult = match(queryContext, terms, viewIds);
|
||||
@@ -120,7 +122,7 @@ public abstract class BaseMatchStrategy<T> implements MatchStrategy<T> {
|
||||
return matches;
|
||||
}
|
||||
|
||||
public List<Term> filterByViewId(List<Term> terms, Set<Long> viewIds) {
|
||||
public List<S2Term> filterByViewId(List<S2Term> terms, Set<Long> viewIds) {
|
||||
logTerms(terms);
|
||||
if (CollectionUtils.isNotEmpty(viewIds)) {
|
||||
terms = terms.stream().filter(term -> {
|
||||
@@ -136,11 +138,11 @@ public abstract class BaseMatchStrategy<T> implements MatchStrategy<T> {
|
||||
return terms;
|
||||
}
|
||||
|
||||
public void logTerms(List<Term> terms) {
|
||||
public void logTerms(List<S2Term> terms) {
|
||||
if (CollectionUtils.isEmpty(terms)) {
|
||||
return;
|
||||
}
|
||||
for (Term term : terms) {
|
||||
for (S2Term term : terms) {
|
||||
log.debug("word:{},nature:{},frequency:{}", term.word, term.nature.toString(), term.getFrequency());
|
||||
}
|
||||
}
|
||||
|
||||
@@ -1,10 +1,10 @@
|
||||
package com.tencent.supersonic.chat.core.mapper;
|
||||
|
||||
import com.hankcs.hanlp.seg.common.Term;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElement;
|
||||
import com.tencent.supersonic.headless.api.pojo.SchemaElement;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElementMatch;
|
||||
import com.tencent.supersonic.chat.core.config.OptimizationConfig;
|
||||
import com.tencent.supersonic.chat.core.knowledge.DatabaseMapResult;
|
||||
import com.tencent.supersonic.headless.api.pojo.response.S2Term;
|
||||
import com.tencent.supersonic.headless.core.knowledge.DatabaseMapResult;
|
||||
import com.tencent.supersonic.chat.core.pojo.QueryContext;
|
||||
import com.tencent.supersonic.common.pojo.Constants;
|
||||
import lombok.extern.slf4j.Slf4j;
|
||||
@@ -36,7 +36,7 @@ public class DatabaseMatchStrategy extends BaseMatchStrategy<DatabaseMapResult>
|
||||
private List<SchemaElement> allElements;
|
||||
|
||||
@Override
|
||||
public Map<MatchText, List<DatabaseMapResult>> match(QueryContext queryContext, List<Term> terms,
|
||||
public Map<MatchText, List<DatabaseMapResult>> match(QueryContext queryContext, List<S2Term> terms,
|
||||
Set<Long> detectModelIds) {
|
||||
this.allElements = getSchemaElements(queryContext);
|
||||
return super.match(queryContext, terms, detectModelIds);
|
||||
|
||||
@@ -1,13 +1,13 @@
|
||||
package com.tencent.supersonic.chat.core.mapper;
|
||||
|
||||
import com.alibaba.fastjson.JSONObject;
|
||||
import com.hankcs.hanlp.seg.common.Term;
|
||||
import com.tencent.supersonic.chat.core.pojo.QueryContext;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElement;
|
||||
import com.tencent.supersonic.headless.api.pojo.SchemaElement;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElementMatch;
|
||||
import com.tencent.supersonic.chat.core.knowledge.EmbeddingResult;
|
||||
import com.tencent.supersonic.chat.core.knowledge.builder.BaseWordBuilder;
|
||||
import com.tencent.supersonic.chat.core.utils.HanlpHelper;
|
||||
import com.tencent.supersonic.headless.api.pojo.response.S2Term;
|
||||
import com.tencent.supersonic.headless.core.knowledge.EmbeddingResult;
|
||||
import com.tencent.supersonic.headless.core.knowledge.builder.BaseWordBuilder;
|
||||
import com.tencent.supersonic.headless.core.knowledge.helper.HanlpHelper;
|
||||
import com.tencent.supersonic.common.util.ContextUtils;
|
||||
import com.tencent.supersonic.common.util.embedding.Retrieval;
|
||||
import java.util.List;
|
||||
@@ -26,7 +26,7 @@ public class EmbeddingMapper extends BaseMapper {
|
||||
public void doMap(QueryContext queryContext) {
|
||||
//1. query from embedding by queryText
|
||||
String queryText = queryContext.getQueryText();
|
||||
List<Term> terms = HanlpHelper.getTerms(queryText);
|
||||
List<S2Term> terms = HanlpHelper.getTerms(queryText);
|
||||
|
||||
EmbeddingMatchStrategy matchStrategy = ContextUtils.getBean(EmbeddingMatchStrategy.class);
|
||||
List<EmbeddingResult> matchResults = matchStrategy.getMatches(queryContext, terms);
|
||||
|
||||
@@ -2,7 +2,7 @@ package com.tencent.supersonic.chat.core.mapper;
|
||||
|
||||
import com.google.common.collect.Lists;
|
||||
import com.tencent.supersonic.chat.core.config.OptimizationConfig;
|
||||
import com.tencent.supersonic.chat.core.knowledge.EmbeddingResult;
|
||||
import com.tencent.supersonic.headless.core.knowledge.EmbeddingResult;
|
||||
import com.tencent.supersonic.chat.core.pojo.QueryContext;
|
||||
import com.tencent.supersonic.common.config.EmbeddingConfig;
|
||||
import com.tencent.supersonic.common.pojo.Constants;
|
||||
|
||||
@@ -1,9 +1,9 @@
|
||||
package com.tencent.supersonic.chat.core.mapper;
|
||||
|
||||
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElement;
|
||||
import com.tencent.supersonic.headless.api.pojo.SchemaElement;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElementMatch;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElementType;
|
||||
import com.tencent.supersonic.headless.api.pojo.SchemaElementType;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaMapInfo;
|
||||
import com.tencent.supersonic.chat.api.pojo.SemanticSchema;
|
||||
import com.tencent.supersonic.chat.api.pojo.ViewSchema;
|
||||
|
||||
@@ -1,9 +1,9 @@
|
||||
package com.tencent.supersonic.chat.core.mapper;
|
||||
|
||||
import com.hankcs.hanlp.seg.common.Term;
|
||||
import com.tencent.supersonic.chat.core.config.OptimizationConfig;
|
||||
import com.tencent.supersonic.chat.core.knowledge.HanlpMapResult;
|
||||
import com.tencent.supersonic.chat.core.knowledge.SearchService;
|
||||
import com.tencent.supersonic.headless.api.pojo.response.S2Term;
|
||||
import com.tencent.supersonic.headless.core.knowledge.HanlpMapResult;
|
||||
import com.tencent.supersonic.headless.core.knowledge.SearchService;
|
||||
import com.tencent.supersonic.chat.core.pojo.QueryContext;
|
||||
import com.tencent.supersonic.common.pojo.Constants;
|
||||
import java.util.HashMap;
|
||||
@@ -35,7 +35,7 @@ public class HanlpDictMatchStrategy extends BaseMatchStrategy<HanlpMapResult> {
|
||||
private OptimizationConfig optimizationConfig;
|
||||
|
||||
@Override
|
||||
public Map<MatchText, List<HanlpMapResult>> match(QueryContext queryContext, List<Term> terms,
|
||||
public Map<MatchText, List<HanlpMapResult>> match(QueryContext queryContext, List<S2Term> terms,
|
||||
Set<Long> detectModelIds) {
|
||||
String text = queryContext.getQueryText();
|
||||
if (Objects.isNull(terms) || StringUtils.isEmpty(text)) {
|
||||
@@ -60,16 +60,15 @@ public class HanlpDictMatchStrategy extends BaseMatchStrategy<HanlpMapResult> {
|
||||
public void detectByStep(QueryContext queryContext, Set<HanlpMapResult> existResults, Set<Long> detectModelIds,
|
||||
Integer startIndex, Integer index, int offset) {
|
||||
String text = queryContext.getQueryText();
|
||||
Integer agentId = queryContext.getAgentId();
|
||||
String detectSegment = text.substring(startIndex, index);
|
||||
|
||||
// step1. pre search
|
||||
Integer oneDetectionMaxSize = optimizationConfig.getOneDetectionMaxSize();
|
||||
LinkedHashSet<HanlpMapResult> hanlpMapResults = SearchService.prefixSearch(detectSegment, oneDetectionMaxSize,
|
||||
agentId, detectModelIds).stream().collect(Collectors.toCollection(LinkedHashSet::new));
|
||||
detectModelIds).stream().collect(Collectors.toCollection(LinkedHashSet::new));
|
||||
// step2. suffix search
|
||||
LinkedHashSet<HanlpMapResult> suffixHanlpMapResults = SearchService.suffixSearch(detectSegment,
|
||||
oneDetectionMaxSize, agentId, detectModelIds).stream()
|
||||
oneDetectionMaxSize, detectModelIds).stream()
|
||||
.collect(Collectors.toCollection(LinkedHashSet::new));
|
||||
|
||||
hanlpMapResults.addAll(suffixHanlpMapResults);
|
||||
|
||||
@@ -1,15 +1,16 @@
|
||||
package com.tencent.supersonic.chat.core.mapper;
|
||||
|
||||
import com.hankcs.hanlp.seg.common.Term;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElement;
|
||||
import com.tencent.supersonic.headless.api.pojo.SchemaElement;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElementMatch;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElementType;
|
||||
import com.tencent.supersonic.headless.api.pojo.SchemaElementType;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaMapInfo;
|
||||
import com.tencent.supersonic.chat.core.knowledge.DatabaseMapResult;
|
||||
import com.tencent.supersonic.chat.core.knowledge.HanlpMapResult;
|
||||
import com.tencent.supersonic.headless.api.pojo.response.S2Term;
|
||||
import com.tencent.supersonic.headless.core.knowledge.DatabaseMapResult;
|
||||
import com.tencent.supersonic.headless.core.knowledge.HanlpMapResult;
|
||||
import com.tencent.supersonic.chat.core.pojo.QueryContext;
|
||||
import com.tencent.supersonic.chat.core.utils.HanlpHelper;
|
||||
import com.tencent.supersonic.chat.core.utils.NatureHelper;
|
||||
import com.tencent.supersonic.headless.server.service.KnowledgeService;
|
||||
import com.tencent.supersonic.headless.core.knowledge.helper.HanlpHelper;
|
||||
import com.tencent.supersonic.headless.core.knowledge.helper.NatureHelper;
|
||||
import com.tencent.supersonic.common.util.ContextUtils;
|
||||
import lombok.extern.slf4j.Slf4j;
|
||||
import org.springframework.util.CollectionUtils;
|
||||
@@ -32,7 +33,8 @@ public class KeywordMapper extends BaseMapper {
|
||||
public void doMap(QueryContext queryContext) {
|
||||
String queryText = queryContext.getQueryText();
|
||||
//1.hanlpDict Match
|
||||
List<Term> terms = HanlpHelper.getTerms(queryText);
|
||||
KnowledgeService knowledgeService = ContextUtils.getBean(KnowledgeService.class);
|
||||
List<S2Term> terms = knowledgeService.getTerms(queryText);
|
||||
HanlpDictMatchStrategy hanlpMatchStrategy = ContextUtils.getBean(HanlpDictMatchStrategy.class);
|
||||
|
||||
List<HanlpMapResult> hanlpMapResults = hanlpMatchStrategy.getMatches(queryContext, terms);
|
||||
@@ -46,7 +48,7 @@ public class KeywordMapper extends BaseMapper {
|
||||
}
|
||||
|
||||
private void convertHanlpMapResultToMapInfo(List<HanlpMapResult> mapResults, QueryContext queryContext,
|
||||
List<Term> terms) {
|
||||
List<S2Term> terms) {
|
||||
if (CollectionUtils.isEmpty(mapResults)) {
|
||||
return;
|
||||
}
|
||||
|
||||
@@ -1,10 +1,10 @@
|
||||
package com.tencent.supersonic.chat.core.mapper;
|
||||
|
||||
import com.hankcs.hanlp.algorithm.EditDistance;
|
||||
import com.hankcs.hanlp.seg.common.Term;
|
||||
import com.tencent.supersonic.chat.core.agent.Agent;
|
||||
import com.tencent.supersonic.chat.core.config.OptimizationConfig;
|
||||
import com.tencent.supersonic.chat.core.utils.NatureHelper;
|
||||
import com.tencent.supersonic.headless.api.pojo.response.S2Term;
|
||||
import com.tencent.supersonic.headless.core.knowledge.helper.NatureHelper;
|
||||
import lombok.Data;
|
||||
import lombok.extern.slf4j.Slf4j;
|
||||
import org.springframework.beans.factory.annotation.Autowired;
|
||||
@@ -36,8 +36,8 @@ public class MapperHelper {
|
||||
return index;
|
||||
}
|
||||
|
||||
public Integer getStepOffset(List<Term> termList, Integer index) {
|
||||
List<Integer> offsetList = termList.stream().sorted(Comparator.comparing(Term::getOffset))
|
||||
public Integer getStepOffset(List<S2Term> termList, Integer index) {
|
||||
List<Integer> offsetList = termList.stream().sorted(Comparator.comparing(S2Term::getOffset))
|
||||
.map(term -> term.getOffset()).collect(Collectors.toList());
|
||||
|
||||
for (int j = 0; j < termList.size() - 1; j++) {
|
||||
|
||||
@@ -1,7 +1,8 @@
|
||||
package com.tencent.supersonic.chat.core.mapper;
|
||||
|
||||
import com.hankcs.hanlp.seg.common.Term;
|
||||
import com.tencent.supersonic.chat.core.pojo.QueryContext;
|
||||
import com.tencent.supersonic.headless.api.pojo.response.S2Term;
|
||||
|
||||
import java.util.List;
|
||||
import java.util.Map;
|
||||
import java.util.Set;
|
||||
@@ -12,6 +13,6 @@ import java.util.Set;
|
||||
*/
|
||||
public interface MatchStrategy<T> {
|
||||
|
||||
Map<MatchText, List<T>> match(QueryContext queryContext, List<Term> terms, Set<Long> detectModelId);
|
||||
Map<MatchText, List<T>> match(QueryContext queryContext, List<S2Term> terms, Set<Long> detectModelId);
|
||||
|
||||
}
|
||||
@@ -1,6 +1,6 @@
|
||||
package com.tencent.supersonic.chat.core.mapper;
|
||||
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElementType;
|
||||
import com.tencent.supersonic.headless.api.pojo.SchemaElementType;
|
||||
import java.io.Serializable;
|
||||
import lombok.Data;
|
||||
import lombok.ToString;
|
||||
|
||||
@@ -1,13 +1,13 @@
|
||||
package com.tencent.supersonic.chat.core.mapper;
|
||||
|
||||
import com.google.common.collect.Lists;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElement;
|
||||
import com.tencent.supersonic.headless.api.pojo.SchemaElement;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElementMatch;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElementType;
|
||||
import com.tencent.supersonic.headless.api.pojo.SchemaElementType;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaMapInfo;
|
||||
import com.tencent.supersonic.chat.api.pojo.request.QueryFilter;
|
||||
import com.tencent.supersonic.chat.api.pojo.request.QueryFilters;
|
||||
import com.tencent.supersonic.chat.core.knowledge.builder.BaseWordBuilder;
|
||||
import com.tencent.supersonic.headless.core.knowledge.builder.BaseWordBuilder;
|
||||
import com.tencent.supersonic.chat.core.pojo.QueryContext;
|
||||
import com.tencent.supersonic.common.pojo.Constants;
|
||||
import lombok.extern.slf4j.Slf4j;
|
||||
|
||||
@@ -1,9 +1,9 @@
|
||||
package com.tencent.supersonic.chat.core.mapper;
|
||||
|
||||
import com.google.common.collect.Lists;
|
||||
import com.hankcs.hanlp.seg.common.Term;
|
||||
import com.tencent.supersonic.chat.core.knowledge.HanlpMapResult;
|
||||
import com.tencent.supersonic.chat.core.knowledge.SearchService;
|
||||
import com.tencent.supersonic.headless.api.pojo.response.S2Term;
|
||||
import com.tencent.supersonic.headless.core.knowledge.HanlpMapResult;
|
||||
import com.tencent.supersonic.headless.core.knowledge.SearchService;
|
||||
import com.tencent.supersonic.chat.core.pojo.QueryContext;
|
||||
import com.tencent.supersonic.common.pojo.enums.DictWordType;
|
||||
import java.util.List;
|
||||
@@ -26,7 +26,7 @@ public class SearchMatchStrategy extends BaseMatchStrategy<HanlpMapResult> {
|
||||
private static final int SEARCH_SIZE = 3;
|
||||
|
||||
@Override
|
||||
public Map<MatchText, List<HanlpMapResult>> match(QueryContext queryContext, List<Term> originals,
|
||||
public Map<MatchText, List<HanlpMapResult>> match(QueryContext queryContext, List<S2Term> originals,
|
||||
Set<Long> detectModelIds) {
|
||||
String text = queryContext.getQueryText();
|
||||
Map<Integer, Integer> regOffsetToLength = getRegOffsetToLength(originals);
|
||||
@@ -52,9 +52,9 @@ public class SearchMatchStrategy extends BaseMatchStrategy<HanlpMapResult> {
|
||||
|
||||
if (StringUtils.isNotEmpty(detectSegment)) {
|
||||
List<HanlpMapResult> hanlpMapResults = SearchService.prefixSearch(detectSegment,
|
||||
SearchService.SEARCH_SIZE, queryContext.getAgentId(), detectModelIds);
|
||||
SearchService.SEARCH_SIZE, detectModelIds);
|
||||
List<HanlpMapResult> suffixHanlpMapResults = SearchService.suffixSearch(
|
||||
detectSegment, SEARCH_SIZE, queryContext.getAgentId(), detectModelIds);
|
||||
detectSegment, SEARCH_SIZE, detectModelIds);
|
||||
hanlpMapResults.addAll(suffixHanlpMapResults);
|
||||
// remove entity name where search
|
||||
hanlpMapResults = hanlpMapResults.stream().filter(entry -> {
|
||||
|
||||
@@ -1,7 +1,7 @@
|
||||
package com.tencent.supersonic.chat.core.parser;
|
||||
|
||||
import com.tencent.supersonic.auth.api.authentication.pojo.User;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElement;
|
||||
import com.tencent.supersonic.headless.api.pojo.SchemaElement;
|
||||
import com.tencent.supersonic.chat.api.pojo.SemanticParseInfo;
|
||||
import com.tencent.supersonic.chat.api.pojo.SemanticSchema;
|
||||
import com.tencent.supersonic.chat.api.pojo.response.SqlInfo;
|
||||
|
||||
@@ -3,7 +3,7 @@ package com.tencent.supersonic.chat.core.parser.plugin;
|
||||
import com.google.common.collect.Lists;
|
||||
import com.google.common.collect.Sets;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElementMatch;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElementType;
|
||||
import com.tencent.supersonic.headless.api.pojo.SchemaElementType;
|
||||
import com.tencent.supersonic.chat.api.pojo.SemanticParseInfo;
|
||||
import com.tencent.supersonic.chat.api.pojo.request.QueryFilter;
|
||||
import com.tencent.supersonic.chat.api.pojo.request.QueryFilters;
|
||||
|
||||
@@ -1,7 +1,7 @@
|
||||
package com.tencent.supersonic.chat.core.parser.sql.llm;
|
||||
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElementMatch;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElementType;
|
||||
import com.tencent.supersonic.headless.api.pojo.SchemaElementType;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaMapInfo;
|
||||
import com.tencent.supersonic.chat.core.pojo.QueryContext;
|
||||
import com.tencent.supersonic.chat.core.query.SemanticQuery;
|
||||
|
||||
@@ -1,16 +1,16 @@
|
||||
package com.tencent.supersonic.chat.core.parser.sql.llm;
|
||||
|
||||
import com.google.common.collect.Lists;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElement;
|
||||
import com.tencent.supersonic.headless.api.pojo.SchemaElement;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElementMatch;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElementType;
|
||||
import com.tencent.supersonic.headless.api.pojo.SchemaElementType;
|
||||
import com.tencent.supersonic.chat.api.pojo.SemanticSchema;
|
||||
import com.tencent.supersonic.chat.core.agent.Agent;
|
||||
import com.tencent.supersonic.chat.core.agent.AgentToolType;
|
||||
import com.tencent.supersonic.chat.core.agent.NL2SQLTool;
|
||||
import com.tencent.supersonic.chat.core.config.LLMParserConfig;
|
||||
import com.tencent.supersonic.chat.core.config.OptimizationConfig;
|
||||
import com.tencent.supersonic.chat.core.knowledge.semantic.SemanticInterpreter;
|
||||
import com.tencent.supersonic.chat.core.query.semantic.SemanticInterpreter;
|
||||
import com.tencent.supersonic.chat.core.parser.SatisfactionChecker;
|
||||
import com.tencent.supersonic.chat.core.pojo.QueryContext;
|
||||
import com.tencent.supersonic.chat.core.query.llm.s2sql.LLMReq;
|
||||
|
||||
@@ -1,7 +1,7 @@
|
||||
package com.tencent.supersonic.chat.core.parser.sql.rule;
|
||||
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElementMatch;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElementType;
|
||||
import com.tencent.supersonic.headless.api.pojo.SchemaElementType;
|
||||
import com.tencent.supersonic.chat.core.parser.SemanticParser;
|
||||
import com.tencent.supersonic.chat.core.pojo.ChatContext;
|
||||
import com.tencent.supersonic.chat.core.pojo.QueryContext;
|
||||
|
||||
@@ -3,9 +3,9 @@ package com.tencent.supersonic.chat.core.plugin;
|
||||
import com.alibaba.fastjson.JSONObject;
|
||||
import com.google.common.collect.Lists;
|
||||
import com.google.common.collect.Sets;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElement;
|
||||
import com.tencent.supersonic.headless.api.pojo.SchemaElement;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElementMatch;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElementType;
|
||||
import com.tencent.supersonic.headless.api.pojo.SchemaElementType;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaMapInfo;
|
||||
import com.tencent.supersonic.chat.core.agent.Agent;
|
||||
import com.tencent.supersonic.chat.core.agent.AgentToolType;
|
||||
|
||||
@@ -6,7 +6,7 @@ import com.tencent.supersonic.chat.api.pojo.SemanticParseInfo;
|
||||
import com.tencent.supersonic.chat.api.pojo.SemanticSchema;
|
||||
import com.tencent.supersonic.chat.api.pojo.response.SqlInfo;
|
||||
import com.tencent.supersonic.chat.core.config.OptimizationConfig;
|
||||
import com.tencent.supersonic.chat.core.knowledge.semantic.SemanticInterpreter;
|
||||
import com.tencent.supersonic.chat.core.query.semantic.SemanticInterpreter;
|
||||
import com.tencent.supersonic.chat.core.utils.ComponentFactory;
|
||||
import com.tencent.supersonic.chat.core.utils.QueryReqBuilder;
|
||||
import com.tencent.supersonic.common.pojo.Aggregator;
|
||||
|
||||
@@ -2,14 +2,14 @@ package com.tencent.supersonic.chat.core.query.llm.analytics;
|
||||
|
||||
import com.google.common.collect.Lists;
|
||||
import com.tencent.supersonic.auth.api.authentication.pojo.User;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElement;
|
||||
import com.tencent.supersonic.headless.api.pojo.SchemaElement;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElementMatch;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElementType;
|
||||
import com.tencent.supersonic.headless.api.pojo.SchemaElementType;
|
||||
import com.tencent.supersonic.chat.api.pojo.SemanticSchema;
|
||||
import com.tencent.supersonic.chat.api.pojo.response.QueryResult;
|
||||
import com.tencent.supersonic.chat.api.pojo.response.QueryState;
|
||||
import com.tencent.supersonic.chat.core.config.OptimizationConfig;
|
||||
import com.tencent.supersonic.chat.core.knowledge.semantic.SemanticInterpreter;
|
||||
import com.tencent.supersonic.chat.core.query.semantic.SemanticInterpreter;
|
||||
import com.tencent.supersonic.chat.core.query.QueryManager;
|
||||
import com.tencent.supersonic.chat.core.query.llm.LLMSemanticQuery;
|
||||
import com.tencent.supersonic.chat.core.utils.ComponentFactory;
|
||||
|
||||
@@ -2,7 +2,7 @@ package com.tencent.supersonic.chat.core.query.llm.s2sql;
|
||||
|
||||
import com.tencent.supersonic.auth.api.authentication.pojo.User;
|
||||
import com.tencent.supersonic.chat.api.pojo.SemanticSchema;
|
||||
import com.tencent.supersonic.chat.core.knowledge.semantic.SemanticInterpreter;
|
||||
import com.tencent.supersonic.chat.core.query.semantic.SemanticInterpreter;
|
||||
import com.tencent.supersonic.chat.api.pojo.response.QueryResult;
|
||||
import com.tencent.supersonic.chat.api.pojo.response.QueryState;
|
||||
import com.tencent.supersonic.chat.api.pojo.response.SqlInfo;
|
||||
|
||||
@@ -3,7 +3,7 @@ package com.tencent.supersonic.chat.core.query.plugin;
|
||||
import com.google.common.collect.Lists;
|
||||
import com.tencent.supersonic.auth.api.authentication.pojo.User;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElementMatch;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElementType;
|
||||
import com.tencent.supersonic.headless.api.pojo.SchemaElementType;
|
||||
import com.tencent.supersonic.chat.api.pojo.SemanticSchema;
|
||||
import com.tencent.supersonic.chat.api.pojo.request.QueryFilter;
|
||||
import com.tencent.supersonic.chat.api.pojo.request.QueryFilters;
|
||||
|
||||
@@ -1,7 +1,7 @@
|
||||
package com.tencent.supersonic.chat.core.query.rule;
|
||||
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElementMatch;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElementType;
|
||||
import com.tencent.supersonic.headless.api.pojo.SchemaElementType;
|
||||
import com.tencent.supersonic.common.pojo.enums.AggregateTypeEnum;
|
||||
import lombok.Data;
|
||||
import lombok.ToString;
|
||||
|
||||
@@ -2,16 +2,16 @@
|
||||
package com.tencent.supersonic.chat.core.query.rule;
|
||||
|
||||
import com.tencent.supersonic.auth.api.authentication.pojo.User;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElement;
|
||||
import com.tencent.supersonic.headless.api.pojo.SchemaElement;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElementMatch;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElementType;
|
||||
import com.tencent.supersonic.headless.api.pojo.SchemaElementType;
|
||||
import com.tencent.supersonic.chat.api.pojo.SemanticParseInfo;
|
||||
import com.tencent.supersonic.chat.api.pojo.SemanticSchema;
|
||||
import com.tencent.supersonic.chat.api.pojo.request.QueryFilter;
|
||||
import com.tencent.supersonic.chat.api.pojo.response.QueryResult;
|
||||
import com.tencent.supersonic.chat.api.pojo.response.QueryState;
|
||||
import com.tencent.supersonic.chat.core.config.OptimizationConfig;
|
||||
import com.tencent.supersonic.chat.core.knowledge.semantic.SemanticInterpreter;
|
||||
import com.tencent.supersonic.chat.core.query.semantic.SemanticInterpreter;
|
||||
import com.tencent.supersonic.chat.core.pojo.ChatContext;
|
||||
import com.tencent.supersonic.chat.core.pojo.QueryContext;
|
||||
import com.tencent.supersonic.chat.core.query.BaseSemanticQuery;
|
||||
|
||||
@@ -1,6 +1,6 @@
|
||||
package com.tencent.supersonic.chat.core.query.rule.metric;
|
||||
|
||||
import static com.tencent.supersonic.chat.api.pojo.SchemaElementType.VALUE;
|
||||
import static com.tencent.supersonic.headless.api.pojo.SchemaElementType.VALUE;
|
||||
import static com.tencent.supersonic.chat.core.query.rule.QueryMatchOption.OptionType.REQUIRED;
|
||||
import static com.tencent.supersonic.chat.core.query.rule.QueryMatchOption.RequireNumberType.AT_LEAST;
|
||||
|
||||
|
||||
@@ -2,8 +2,8 @@ package com.tencent.supersonic.chat.core.query.rule.metric;
|
||||
|
||||
import org.springframework.stereotype.Component;
|
||||
|
||||
import static com.tencent.supersonic.chat.api.pojo.SchemaElementType.DIMENSION;
|
||||
import static com.tencent.supersonic.chat.api.pojo.SchemaElementType.VALUE;
|
||||
import static com.tencent.supersonic.headless.api.pojo.SchemaElementType.DIMENSION;
|
||||
import static com.tencent.supersonic.headless.api.pojo.SchemaElementType.VALUE;
|
||||
import static com.tencent.supersonic.chat.core.query.rule.QueryMatchOption.OptionType.OPTIONAL;
|
||||
import static com.tencent.supersonic.chat.core.query.rule.QueryMatchOption.RequireNumberType.AT_LEAST;
|
||||
import static com.tencent.supersonic.chat.core.query.rule.QueryMatchOption.OptionType.REQUIRED;
|
||||
|
||||
@@ -4,7 +4,7 @@ import static com.tencent.supersonic.chat.core.query.rule.QueryMatchOption.Optio
|
||||
import static com.tencent.supersonic.chat.core.query.rule.QueryMatchOption.RequireNumberType.AT_MOST;
|
||||
|
||||
import com.tencent.supersonic.auth.api.authentication.pojo.User;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElementType;
|
||||
import com.tencent.supersonic.headless.api.pojo.SchemaElementType;
|
||||
import com.tencent.supersonic.chat.api.pojo.response.QueryResult;
|
||||
import org.springframework.stereotype.Component;
|
||||
@Component
|
||||
|
||||
@@ -1,8 +1,8 @@
|
||||
package com.tencent.supersonic.chat.core.query.rule.metric;
|
||||
|
||||
import static com.tencent.supersonic.chat.api.pojo.SchemaElementType.METRIC;
|
||||
import static com.tencent.supersonic.chat.core.query.rule.QueryMatchOption.OptionType.REQUIRED;
|
||||
import static com.tencent.supersonic.chat.core.query.rule.QueryMatchOption.RequireNumberType.AT_LEAST;
|
||||
import static com.tencent.supersonic.headless.api.pojo.SchemaElementType.METRIC;
|
||||
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElementMatch;
|
||||
import com.tencent.supersonic.chat.api.pojo.ViewSchema;
|
||||
@@ -29,7 +29,7 @@ public abstract class MetricSemanticQuery extends RuleSemanticQuery {
|
||||
|
||||
@Override
|
||||
public List<SchemaElementMatch> match(List<SchemaElementMatch> candidateElementMatches,
|
||||
QueryContext queryCtx) {
|
||||
QueryContext queryCtx) {
|
||||
return super.match(candidateElementMatches, queryCtx);
|
||||
}
|
||||
|
||||
|
||||
@@ -1,7 +1,7 @@
|
||||
package com.tencent.supersonic.chat.core.query.rule.metric;
|
||||
|
||||
import static com.tencent.supersonic.chat.api.pojo.SchemaElementType.ENTITY;
|
||||
import static com.tencent.supersonic.chat.api.pojo.SchemaElementType.ID;
|
||||
import static com.tencent.supersonic.headless.api.pojo.SchemaElementType.ENTITY;
|
||||
import static com.tencent.supersonic.headless.api.pojo.SchemaElementType.ID;
|
||||
import static com.tencent.supersonic.chat.core.query.rule.QueryMatchOption.OptionType.REQUIRED;
|
||||
import static com.tencent.supersonic.chat.core.query.rule.QueryMatchOption.RequireNumberType.AT_LEAST;
|
||||
|
||||
|
||||
@@ -1,13 +1,13 @@
|
||||
package com.tencent.supersonic.chat.core.query.rule.metric;
|
||||
|
||||
import static com.tencent.supersonic.chat.api.pojo.SchemaElementType.DIMENSION;
|
||||
import static com.tencent.supersonic.chat.api.pojo.SchemaElementType.VALUE;
|
||||
import static com.tencent.supersonic.headless.api.pojo.SchemaElementType.DIMENSION;
|
||||
import static com.tencent.supersonic.headless.api.pojo.SchemaElementType.VALUE;
|
||||
import static com.tencent.supersonic.chat.core.query.rule.QueryMatchOption.OptionType.OPTIONAL;
|
||||
import static com.tencent.supersonic.chat.core.query.rule.QueryMatchOption.OptionType.REQUIRED;
|
||||
import static com.tencent.supersonic.chat.core.query.rule.QueryMatchOption.RequireNumberType.AT_LEAST;
|
||||
import static com.tencent.supersonic.common.pojo.Constants.DESC_UPPER;
|
||||
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElement;
|
||||
import com.tencent.supersonic.headless.api.pojo.SchemaElement;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElementMatch;
|
||||
import com.tencent.supersonic.chat.core.pojo.ChatContext;
|
||||
import com.tencent.supersonic.chat.core.pojo.QueryContext;
|
||||
|
||||
@@ -1,7 +1,7 @@
|
||||
package com.tencent.supersonic.chat.core.query.rule.tag;
|
||||
|
||||
import static com.tencent.supersonic.chat.api.pojo.SchemaElementType.DIMENSION;
|
||||
import static com.tencent.supersonic.chat.api.pojo.SchemaElementType.ID;
|
||||
import static com.tencent.supersonic.headless.api.pojo.SchemaElementType.DIMENSION;
|
||||
import static com.tencent.supersonic.headless.api.pojo.SchemaElementType.ID;
|
||||
import static com.tencent.supersonic.chat.core.query.rule.QueryMatchOption.RequireNumberType.AT_LEAST;
|
||||
import static com.tencent.supersonic.chat.core.query.rule.QueryMatchOption.OptionType.REQUIRED;
|
||||
|
||||
|
||||
@@ -1,6 +1,6 @@
|
||||
package com.tencent.supersonic.chat.core.query.rule.tag;
|
||||
|
||||
import static com.tencent.supersonic.chat.api.pojo.SchemaElementType.VALUE;
|
||||
import static com.tencent.supersonic.headless.api.pojo.SchemaElementType.VALUE;
|
||||
import static com.tencent.supersonic.chat.core.query.rule.QueryMatchOption.OptionType.REQUIRED;
|
||||
import static com.tencent.supersonic.chat.core.query.rule.QueryMatchOption.RequireNumberType.AT_LEAST;
|
||||
|
||||
|
||||
@@ -1,6 +1,6 @@
|
||||
package com.tencent.supersonic.chat.core.query.rule.tag;
|
||||
|
||||
import static com.tencent.supersonic.chat.api.pojo.SchemaElementType.ID;
|
||||
import static com.tencent.supersonic.headless.api.pojo.SchemaElementType.ID;
|
||||
import static com.tencent.supersonic.chat.core.query.rule.QueryMatchOption.OptionType.REQUIRED;
|
||||
import static com.tencent.supersonic.chat.core.query.rule.QueryMatchOption.RequireNumberType.AT_LEAST;
|
||||
|
||||
|
||||
@@ -1,7 +1,7 @@
|
||||
package com.tencent.supersonic.chat.core.query.rule.tag;
|
||||
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElement;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElementType;
|
||||
import com.tencent.supersonic.headless.api.pojo.SchemaElement;
|
||||
import com.tencent.supersonic.headless.api.pojo.SchemaElementType;
|
||||
import com.tencent.supersonic.chat.api.pojo.SemanticParseInfo;
|
||||
import com.tencent.supersonic.chat.api.pojo.ViewSchema;
|
||||
import com.tencent.supersonic.chat.core.pojo.ChatContext;
|
||||
|
||||
@@ -15,7 +15,7 @@ import java.time.LocalDate;
|
||||
import java.util.List;
|
||||
import java.util.Objects;
|
||||
|
||||
import static com.tencent.supersonic.chat.api.pojo.SchemaElementType.ENTITY;
|
||||
import static com.tencent.supersonic.headless.api.pojo.SchemaElementType.ENTITY;
|
||||
import static com.tencent.supersonic.chat.core.query.rule.QueryMatchOption.OptionType.REQUIRED;
|
||||
import static com.tencent.supersonic.chat.core.query.rule.QueryMatchOption.RequireNumberType.AT_LEAST;
|
||||
|
||||
|
||||
@@ -1,4 +1,4 @@
|
||||
package com.tencent.supersonic.chat.core.knowledge.semantic;
|
||||
package com.tencent.supersonic.chat.core.query.semantic;
|
||||
|
||||
import com.google.common.cache.Cache;
|
||||
import com.google.common.cache.CacheBuilder;
|
||||
@@ -1,4 +1,4 @@
|
||||
package com.tencent.supersonic.chat.core.knowledge.semantic;
|
||||
package com.tencent.supersonic.chat.core.query.semantic;
|
||||
|
||||
import com.github.pagehelper.PageInfo;
|
||||
import com.tencent.supersonic.auth.api.authentication.pojo.User;
|
||||
@@ -1,4 +1,4 @@
|
||||
package com.tencent.supersonic.chat.core.knowledge.semantic;
|
||||
package com.tencent.supersonic.chat.core.query.semantic;
|
||||
|
||||
import com.github.pagehelper.PageInfo;
|
||||
import com.google.gson.Gson;
|
||||
@@ -1,4 +1,4 @@
|
||||
package com.tencent.supersonic.chat.core.knowledge.semantic;
|
||||
package com.tencent.supersonic.chat.core.query.semantic;
|
||||
|
||||
import com.github.pagehelper.PageInfo;
|
||||
import com.tencent.supersonic.auth.api.authentication.pojo.User;
|
||||
@@ -1,10 +1,10 @@
|
||||
package com.tencent.supersonic.chat.core.knowledge.semantic;
|
||||
package com.tencent.supersonic.chat.core.query.semantic;
|
||||
|
||||
import com.google.common.collect.Lists;
|
||||
import com.tencent.supersonic.chat.api.pojo.RelatedSchemaElement;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElement;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElementType;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaValueMap;
|
||||
import com.tencent.supersonic.headless.api.pojo.RelatedSchemaElement;
|
||||
import com.tencent.supersonic.headless.api.pojo.SchemaElement;
|
||||
import com.tencent.supersonic.headless.api.pojo.SchemaElementType;
|
||||
import com.tencent.supersonic.headless.api.pojo.SchemaValueMap;
|
||||
import com.tencent.supersonic.chat.api.pojo.ViewSchema;
|
||||
import com.tencent.supersonic.headless.api.pojo.DimValueMap;
|
||||
import com.tencent.supersonic.headless.api.pojo.RelateDimension;
|
||||
@@ -1,6 +1,6 @@
|
||||
package com.tencent.supersonic.chat.core.utils;
|
||||
|
||||
import com.tencent.supersonic.chat.core.knowledge.semantic.SemanticInterpreter;
|
||||
import com.tencent.supersonic.chat.core.query.semantic.SemanticInterpreter;
|
||||
import com.tencent.supersonic.chat.core.parser.JavaLLMProxy;
|
||||
import com.tencent.supersonic.chat.core.parser.LLMProxy;
|
||||
import com.tencent.supersonic.chat.core.parser.sql.llm.ViewResolver;
|
||||
|
||||
@@ -1,219 +0,0 @@
|
||||
package com.tencent.supersonic.chat.core.utils;
|
||||
|
||||
import com.tencent.supersonic.auth.api.authentication.pojo.User;
|
||||
import com.tencent.supersonic.chat.core.config.DefaultMetric;
|
||||
import com.tencent.supersonic.chat.core.config.Dim4Dict;
|
||||
import com.tencent.supersonic.chat.core.knowledge.semantic.SemanticInterpreter;
|
||||
import com.tencent.supersonic.common.pojo.Aggregator;
|
||||
import com.tencent.supersonic.common.pojo.Constants;
|
||||
import com.tencent.supersonic.common.pojo.DateConf;
|
||||
import com.tencent.supersonic.common.pojo.Filter;
|
||||
import com.tencent.supersonic.common.pojo.Order;
|
||||
import com.tencent.supersonic.common.pojo.QueryColumn;
|
||||
import com.tencent.supersonic.common.pojo.enums.AggOperatorEnum;
|
||||
import com.tencent.supersonic.common.pojo.enums.FilterOperatorEnum;
|
||||
import com.tencent.supersonic.headless.api.pojo.request.QueryStructReq;
|
||||
import com.tencent.supersonic.headless.api.pojo.response.SemanticQueryResp;
|
||||
import lombok.extern.slf4j.Slf4j;
|
||||
import org.apache.commons.lang3.StringUtils;
|
||||
import org.apache.logging.log4j.util.Strings;
|
||||
import org.springframework.beans.factory.annotation.Value;
|
||||
import org.springframework.stereotype.Component;
|
||||
import org.springframework.util.CollectionUtils;
|
||||
|
||||
import java.util.ArrayList;
|
||||
import java.util.Arrays;
|
||||
import java.util.HashMap;
|
||||
import java.util.List;
|
||||
import java.util.Map;
|
||||
import java.util.Objects;
|
||||
import java.util.StringJoiner;
|
||||
|
||||
import static com.tencent.supersonic.common.pojo.Constants.AND_UPPER;
|
||||
import static com.tencent.supersonic.common.pojo.Constants.APOSTROPHE;
|
||||
import static com.tencent.supersonic.common.pojo.Constants.COMMA;
|
||||
import static com.tencent.supersonic.common.pojo.Constants.SPACE;
|
||||
import static com.tencent.supersonic.common.pojo.Constants.UNDERLINE_DOUBLE;
|
||||
|
||||
@Slf4j
|
||||
@Component
|
||||
public class DictQueryHelper {
|
||||
|
||||
private static final Long MAX_FREQUENCY = 99999999L;
|
||||
private SemanticInterpreter semanticInterpreter = ComponentFactory.getSemanticLayer();
|
||||
@Value("${dimension.multi.value.split:#}")
|
||||
private String dimMultiValueSplit;
|
||||
@Value("${dimension.value.show:50}")
|
||||
private Integer printDataShow;
|
||||
@Value("${dimension.max.limit:3000000}")
|
||||
private Long dimMaxLimit;
|
||||
@Value("${dimension.white.weight:60000000}")
|
||||
private Long dimensionWhiteWeight;
|
||||
|
||||
public List<String> fetchDimValueSingle(Long modelId, DefaultMetric defaultMetricDesc, Dim4Dict dim4Dict,
|
||||
User user) {
|
||||
List<String> data = new ArrayList<>();
|
||||
QueryStructReq queryStructCmd = generateQueryStructCmd(modelId, defaultMetricDesc, dim4Dict);
|
||||
try {
|
||||
SemanticQueryResp semanticQueryResp = semanticInterpreter.queryByStruct(queryStructCmd, user);
|
||||
|
||||
log.info("fetchDimValueSingle sql:{}", semanticQueryResp.getSql());
|
||||
String nature = String.format("_%d_%d", modelId, dim4Dict.getDimId());
|
||||
String dimNameRewrite = rewriteDimName(semanticQueryResp.getColumns(), dim4Dict.getBizName());
|
||||
data = generateFileData(semanticQueryResp.getResultList(), nature, dimNameRewrite,
|
||||
defaultMetricDesc.getBizName(), dim4Dict);
|
||||
if (!CollectionUtils.isEmpty(data)) {
|
||||
int size = (data.size() > printDataShow) ? printDataShow : data.size();
|
||||
log.info("data:{}", data.subList(0, size));
|
||||
} else {
|
||||
log.warn("data is empty. nature:{}", nature);
|
||||
if (Objects.nonNull(semanticQueryResp)) {
|
||||
log.warn("sql:{}", semanticQueryResp.getSql());
|
||||
}
|
||||
}
|
||||
|
||||
} catch (Exception e) {
|
||||
log.warn("fetchDimValueSingle,e:", e);
|
||||
}
|
||||
return data;
|
||||
}
|
||||
|
||||
private String rewriteDimName(List<QueryColumn> columns, String bizName) {
|
||||
// metric parser join dimension style
|
||||
String dimNameRewrite = bizName;
|
||||
|
||||
if (!CollectionUtils.isEmpty(columns)) {
|
||||
for (QueryColumn column : columns) {
|
||||
if (StringUtils.isNotEmpty(column.getNameEn())) {
|
||||
String nameEn = column.getNameEn();
|
||||
if (nameEn.endsWith(UNDERLINE_DOUBLE + bizName)) {
|
||||
dimNameRewrite = nameEn;
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
return dimNameRewrite;
|
||||
}
|
||||
|
||||
private List<String> generateFileData(List<Map<String, Object>> resultList, String nature, String dimName,
|
||||
String metricName, Dim4Dict dim4Dict) {
|
||||
List<String> data = new ArrayList<>();
|
||||
if (CollectionUtils.isEmpty(resultList)) {
|
||||
return data;
|
||||
}
|
||||
Map<String, Long> valueAndFrequencyPair = new HashMap<>(2000);
|
||||
for (Map<String, Object> line : resultList) {
|
||||
|
||||
if (CollectionUtils.isEmpty(line) || !line.containsKey(dimName)
|
||||
|| !line.containsKey(metricName)
|
||||
|| line.get(dimName) == null) {
|
||||
continue;
|
||||
}
|
||||
|
||||
String dimValue = line.get(dimName).toString();
|
||||
Object metricObject = line.get(metricName);
|
||||
if (Strings.isNotEmpty(dimValue) && Objects.nonNull(metricObject)) {
|
||||
Long metric = Math.round(Double.parseDouble(metricObject.toString()));
|
||||
mergeMultivaluedValue(valueAndFrequencyPair, dimValue, metric);
|
||||
}
|
||||
|
||||
}
|
||||
constructDataLines(valueAndFrequencyPair, nature, data, dim4Dict);
|
||||
return data;
|
||||
}
|
||||
|
||||
private void constructDataLines(Map<String, Long> valueAndFrequencyPair, String nature,
|
||||
List<String> data, Dim4Dict dim4Dict) {
|
||||
valueAndFrequencyPair.forEach((dimValue, metric) -> {
|
||||
if (metric > MAX_FREQUENCY) {
|
||||
metric = MAX_FREQUENCY;
|
||||
}
|
||||
if (Strings.isNotEmpty(dimValue) && dimValue.contains(SPACE)) {
|
||||
dimValue = dimValue.replace(SPACE, "#");
|
||||
}
|
||||
data.add(String.format("%s %s %s", dimValue, nature, metric));
|
||||
});
|
||||
|
||||
if (Objects.nonNull(dim4Dict) && !CollectionUtils.isEmpty(dim4Dict.getWhiteList())) {
|
||||
dim4Dict.getWhiteList().stream()
|
||||
.forEach(white -> data.add(String.format("%s %s %s", white, nature, dimensionWhiteWeight)));
|
||||
}
|
||||
}
|
||||
|
||||
private void mergeMultivaluedValue(Map<String, Long> valueAndFrequencyPair, String dimValue, Long metric) {
|
||||
if (Strings.isEmpty(dimValue)) {
|
||||
return;
|
||||
}
|
||||
Map<String, Long> tmp = new HashMap<>();
|
||||
if (dimValue.contains(dimMultiValueSplit)) {
|
||||
Arrays.stream(dimValue.split(dimMultiValueSplit))
|
||||
.forEach(dimValueSingle -> tmp.put(dimValueSingle, metric));
|
||||
} else {
|
||||
tmp.put(dimValue, metric);
|
||||
}
|
||||
|
||||
for (String value : tmp.keySet()) {
|
||||
long metricOld = valueAndFrequencyPair.containsKey(value) ? valueAndFrequencyPair.get(value) : 0L;
|
||||
valueAndFrequencyPair.put(value, metric + metricOld);
|
||||
}
|
||||
}
|
||||
|
||||
private QueryStructReq generateQueryStructCmd(Long modelId, DefaultMetric defaultMetricDesc, Dim4Dict dim4Dict) {
|
||||
QueryStructReq queryStructCmd = new QueryStructReq();
|
||||
|
||||
List<Filter> filters = generateFilters(dim4Dict, queryStructCmd);
|
||||
queryStructCmd.setDimensionFilters(filters);
|
||||
|
||||
List<Aggregator> aggregators = new ArrayList<>();
|
||||
aggregators.add(new Aggregator(defaultMetricDesc.getBizName(), AggOperatorEnum.SUM));
|
||||
queryStructCmd.setAggregators(aggregators);
|
||||
|
||||
List<Order> orders = new ArrayList<>();
|
||||
orders.add(new Order(defaultMetricDesc.getBizName(), Constants.DESC_UPPER));
|
||||
queryStructCmd.setOrders(orders);
|
||||
|
||||
DateConf dateInfo = new DateConf();
|
||||
dateInfo.setDateMode(DateConf.DateMode.RECENT);
|
||||
log.debug("defaultMetric unit():{}", defaultMetricDesc.getUnit());
|
||||
dateInfo.setUnit(defaultMetricDesc.getUnit());
|
||||
dateInfo.setPeriod(defaultMetricDesc.getPeriod());
|
||||
queryStructCmd.setDateInfo(dateInfo);
|
||||
|
||||
queryStructCmd.setLimit(dimMaxLimit);
|
||||
queryStructCmd.setNeedAuth(false);
|
||||
return queryStructCmd;
|
||||
|
||||
}
|
||||
|
||||
private List<Filter> generateFilters(Dim4Dict dim4Dict, QueryStructReq queryStructCmd) {
|
||||
String whereStr = generateFilter(dim4Dict);
|
||||
if (Strings.isEmpty(whereStr)) {
|
||||
return new ArrayList<>();
|
||||
}
|
||||
Filter filter = new Filter("", FilterOperatorEnum.SQL_PART, whereStr);
|
||||
List<Filter> filters = Objects.isNull(queryStructCmd.getOriginalFilter()) ? new ArrayList<>()
|
||||
: queryStructCmd.getOriginalFilter();
|
||||
filters.add(filter);
|
||||
return filters;
|
||||
}
|
||||
|
||||
private String generateFilter(Dim4Dict dim4Dict) {
|
||||
if (Objects.isNull(dim4Dict)) {
|
||||
return "";
|
||||
}
|
||||
StringJoiner joiner = new StringJoiner(SPACE + AND_UPPER + SPACE);
|
||||
|
||||
String dimName = dim4Dict.getBizName();
|
||||
if (!CollectionUtils.isEmpty(dim4Dict.getBlackList())) {
|
||||
StringJoiner joinerBlack = new StringJoiner(COMMA);
|
||||
dim4Dict.getBlackList().stream().forEach(black -> joinerBlack.add(APOSTROPHE + black + APOSTROPHE));
|
||||
joiner.add(String.format("(%s not in (%s))", dimName, joinerBlack.toString()));
|
||||
}
|
||||
|
||||
if (!CollectionUtils.isEmpty(dim4Dict.getRuleList())) {
|
||||
dim4Dict.getRuleList().stream().forEach(rule -> joiner.add(rule));
|
||||
}
|
||||
|
||||
return joiner.toString();
|
||||
}
|
||||
}
|
||||
@@ -1,82 +0,0 @@
|
||||
package com.tencent.supersonic.chat.core.utils;
|
||||
|
||||
import static com.hankcs.hanlp.HanLP.Config.CustomDictionaryPath;
|
||||
|
||||
import com.hankcs.hanlp.dictionary.DynamicCustomDictionary;
|
||||
|
||||
import java.io.File;
|
||||
import java.io.IOException;
|
||||
import java.util.ArrayList;
|
||||
import java.util.List;
|
||||
|
||||
import lombok.extern.slf4j.Slf4j;
|
||||
|
||||
@Slf4j
|
||||
public class FileHelper {
|
||||
|
||||
public static final String FILE_SPILT = File.separator;
|
||||
|
||||
public static void deleteCacheFile(String[] path) throws IOException {
|
||||
|
||||
String customPath = getCustomPath(path);
|
||||
File customFolder = new File(customPath);
|
||||
|
||||
File[] customSubFiles = getFileList(customFolder, ".bin");
|
||||
|
||||
for (File file : customSubFiles) {
|
||||
try {
|
||||
file.delete();
|
||||
log.info("customPath:{},delete file:{}", customPath, file);
|
||||
} catch (Exception e) {
|
||||
log.error("delete " + file, e);
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
private static File[] getFileList(File customFolder, String suffix) {
|
||||
File[] customSubFiles = customFolder.listFiles(file -> {
|
||||
if (file.isDirectory()) {
|
||||
return false;
|
||||
}
|
||||
if (file.getName().toLowerCase().endsWith(suffix)) {
|
||||
return true;
|
||||
}
|
||||
return false;
|
||||
});
|
||||
return customSubFiles;
|
||||
}
|
||||
|
||||
private static String getCustomPath(String[] path) {
|
||||
return path[0].substring(0, path[0].lastIndexOf(FILE_SPILT)) + FILE_SPILT;
|
||||
}
|
||||
|
||||
/**
|
||||
* reset path
|
||||
*
|
||||
* @param customDictionary
|
||||
*/
|
||||
public static void resetCustomPath(DynamicCustomDictionary customDictionary) {
|
||||
String[] path = CustomDictionaryPath;
|
||||
|
||||
String customPath = getCustomPath(path);
|
||||
File customFolder = new File(customPath);
|
||||
|
||||
File[] customSubFiles = getFileList(customFolder, ".txt");
|
||||
|
||||
List<String> fileList = new ArrayList<>();
|
||||
|
||||
for (File file : customSubFiles) {
|
||||
if (file.isFile()) {
|
||||
fileList.add(file.getAbsolutePath());
|
||||
}
|
||||
}
|
||||
|
||||
log.debug("CustomDictionaryPath:{}", fileList);
|
||||
CustomDictionaryPath = fileList.toArray(new String[0]);
|
||||
customDictionary.path = (CustomDictionaryPath == null || CustomDictionaryPath.length == 0) ? path
|
||||
: CustomDictionaryPath;
|
||||
if (CustomDictionaryPath == null || CustomDictionaryPath.length == 0) {
|
||||
CustomDictionaryPath = path;
|
||||
}
|
||||
}
|
||||
}
|
||||
@@ -1,215 +0,0 @@
|
||||
package com.tencent.supersonic.chat.core.utils;
|
||||
|
||||
import static com.hankcs.hanlp.HanLP.Config.CustomDictionaryPath;
|
||||
|
||||
import com.hankcs.hanlp.HanLP;
|
||||
import com.hankcs.hanlp.corpus.tag.Nature;
|
||||
import com.hankcs.hanlp.dictionary.CoreDictionary;
|
||||
import com.hankcs.hanlp.dictionary.DynamicCustomDictionary;
|
||||
import com.hankcs.hanlp.seg.Segment;
|
||||
import com.hankcs.hanlp.seg.common.Term;
|
||||
import com.tencent.supersonic.chat.core.knowledge.DictWord;
|
||||
import com.tencent.supersonic.chat.core.knowledge.HadoopFileIOAdapter;
|
||||
import com.tencent.supersonic.chat.core.knowledge.MapResult;
|
||||
import com.tencent.supersonic.chat.core.knowledge.MultiCustomDictionary;
|
||||
import com.tencent.supersonic.chat.core.knowledge.SearchService;
|
||||
import com.tencent.supersonic.common.pojo.enums.DictWordType;
|
||||
import java.io.File;
|
||||
import java.io.FileNotFoundException;
|
||||
import java.io.IOException;
|
||||
import java.util.ArrayList;
|
||||
import java.util.Arrays;
|
||||
import java.util.List;
|
||||
import java.util.stream.Collectors;
|
||||
import lombok.extern.slf4j.Slf4j;
|
||||
import org.apache.commons.lang3.StringUtils;
|
||||
import org.springframework.util.CollectionUtils;
|
||||
import org.springframework.util.ResourceUtils;
|
||||
|
||||
/**
|
||||
* HanLP helper
|
||||
*/
|
||||
@Slf4j
|
||||
public class HanlpHelper {
|
||||
|
||||
public static final String FILE_SPILT = File.separator;
|
||||
public static final String SPACE_SPILT = "#";
|
||||
private static volatile DynamicCustomDictionary CustomDictionary;
|
||||
private static volatile Segment segment;
|
||||
|
||||
static {
|
||||
// reset hanlp config
|
||||
try {
|
||||
resetHanlpConfig();
|
||||
} catch (FileNotFoundException e) {
|
||||
log.error("resetHanlpConfig error", e);
|
||||
}
|
||||
}
|
||||
|
||||
public static Segment getSegment() {
|
||||
if (segment == null) {
|
||||
synchronized (HanlpHelper.class) {
|
||||
if (segment == null) {
|
||||
segment = HanLP.newSegment()
|
||||
.enableIndexMode(true).enableIndexMode(4)
|
||||
.enableCustomDictionary(true).enableCustomDictionaryForcing(true).enableOffset(true)
|
||||
.enableJapaneseNameRecognize(false).enableNameRecognize(false)
|
||||
.enableAllNamedEntityRecognize(false)
|
||||
.enableJapaneseNameRecognize(false).enableNumberQuantifierRecognize(false)
|
||||
.enablePlaceRecognize(false)
|
||||
.enableOrganizationRecognize(false).enableCustomDictionary(getDynamicCustomDictionary());
|
||||
}
|
||||
}
|
||||
}
|
||||
return segment;
|
||||
}
|
||||
|
||||
public static DynamicCustomDictionary getDynamicCustomDictionary() {
|
||||
if (CustomDictionary == null) {
|
||||
synchronized (HanlpHelper.class) {
|
||||
if (CustomDictionary == null) {
|
||||
CustomDictionary = new MultiCustomDictionary(CustomDictionaryPath);
|
||||
}
|
||||
}
|
||||
}
|
||||
return CustomDictionary;
|
||||
}
|
||||
|
||||
/***
|
||||
* reload custom dictionary
|
||||
*/
|
||||
public static boolean reloadCustomDictionary() throws IOException {
|
||||
|
||||
log.info("reloadCustomDictionary start");
|
||||
|
||||
final long startTime = System.currentTimeMillis();
|
||||
|
||||
if (CustomDictionaryPath == null || CustomDictionaryPath.length == 0) {
|
||||
return false;
|
||||
}
|
||||
if (HanLP.Config.IOAdapter instanceof HadoopFileIOAdapter) {
|
||||
// 1.delete hdfs file
|
||||
HdfsFileHelper.deleteCacheFile(CustomDictionaryPath);
|
||||
// 2.query txt files，update CustomDictionaryPath
|
||||
HdfsFileHelper.resetCustomPath(getDynamicCustomDictionary());
|
||||
} else {
|
||||
FileHelper.deleteCacheFile(CustomDictionaryPath);
|
||||
FileHelper.resetCustomPath(getDynamicCustomDictionary());
|
||||
}
|
||||
// 3.clear trie
|
||||
SearchService.clear();
|
||||
|
||||
boolean reload = getDynamicCustomDictionary().reload();
|
||||
log.info("reloadCustomDictionary end ,cost:{},reload:{}", System.currentTimeMillis() - startTime, reload);
|
||||
return reload;
|
||||
}
|
||||
|
||||
private static void resetHanlpConfig() throws FileNotFoundException {
|
||||
if (HanLP.Config.IOAdapter instanceof HadoopFileIOAdapter) {
|
||||
return;
|
||||
}
|
||||
String hanlpPropertiesPath = getHanlpPropertiesPath();
|
||||
|
||||
CustomDictionaryPath = Arrays.stream(CustomDictionaryPath).map(path -> hanlpPropertiesPath + FILE_SPILT + path)
|
||||
.toArray(String[]::new);
|
||||
log.info("hanlpPropertiesPath:{},CustomDictionaryPath:{}", hanlpPropertiesPath, CustomDictionaryPath);
|
||||
|
||||
HanLP.Config.CoreDictionaryPath = hanlpPropertiesPath + FILE_SPILT + HanLP.Config.BiGramDictionaryPath;
|
||||
HanLP.Config.CoreDictionaryTransformMatrixDictionaryPath = hanlpPropertiesPath + FILE_SPILT
|
||||
+ HanLP.Config.CoreDictionaryTransformMatrixDictionaryPath;
|
||||
HanLP.Config.BiGramDictionaryPath = hanlpPropertiesPath + FILE_SPILT + HanLP.Config.BiGramDictionaryPath;
|
||||
HanLP.Config.CoreStopWordDictionaryPath =
|
||||
hanlpPropertiesPath + FILE_SPILT + HanLP.Config.CoreStopWordDictionaryPath;
|
||||
HanLP.Config.CoreSynonymDictionaryDictionaryPath = hanlpPropertiesPath + FILE_SPILT
|
||||
+ HanLP.Config.CoreSynonymDictionaryDictionaryPath;
|
||||
HanLP.Config.PersonDictionaryPath = hanlpPropertiesPath + FILE_SPILT + HanLP.Config.PersonDictionaryPath;
|
||||
HanLP.Config.PersonDictionaryTrPath = hanlpPropertiesPath + FILE_SPILT + HanLP.Config.PersonDictionaryTrPath;
|
||||
|
||||
HanLP.Config.PinyinDictionaryPath = hanlpPropertiesPath + FILE_SPILT + HanLP.Config.PinyinDictionaryPath;
|
||||
HanLP.Config.TranslatedPersonDictionaryPath = hanlpPropertiesPath + FILE_SPILT
|
||||
+ HanLP.Config.TranslatedPersonDictionaryPath;
|
||||
HanLP.Config.JapanesePersonDictionaryPath = hanlpPropertiesPath + FILE_SPILT
|
||||
+ HanLP.Config.JapanesePersonDictionaryPath;
|
||||
HanLP.Config.PlaceDictionaryPath = hanlpPropertiesPath + FILE_SPILT + HanLP.Config.PlaceDictionaryPath;
|
||||
HanLP.Config.PlaceDictionaryTrPath = hanlpPropertiesPath + FILE_SPILT + HanLP.Config.PlaceDictionaryTrPath;
|
||||
HanLP.Config.OrganizationDictionaryPath = hanlpPropertiesPath + FILE_SPILT
|
||||
+ HanLP.Config.OrganizationDictionaryPath;
|
||||
HanLP.Config.OrganizationDictionaryTrPath = hanlpPropertiesPath + FILE_SPILT
|
||||
+ HanLP.Config.OrganizationDictionaryTrPath;
|
||||
HanLP.Config.CharTypePath = hanlpPropertiesPath + FILE_SPILT + HanLP.Config.CharTypePath;
|
||||
HanLP.Config.CharTablePath = hanlpPropertiesPath + FILE_SPILT + HanLP.Config.CharTablePath;
|
||||
HanLP.Config.PartOfSpeechTagDictionary =
|
||||
hanlpPropertiesPath + FILE_SPILT + HanLP.Config.PartOfSpeechTagDictionary;
|
||||
HanLP.Config.WordNatureModelPath = hanlpPropertiesPath + FILE_SPILT + HanLP.Config.WordNatureModelPath;
|
||||
HanLP.Config.MaxEntModelPath = hanlpPropertiesPath + FILE_SPILT + HanLP.Config.MaxEntModelPath;
|
||||
HanLP.Config.NNParserModelPath = hanlpPropertiesPath + FILE_SPILT + HanLP.Config.NNParserModelPath;
|
||||
HanLP.Config.PerceptronParserModelPath =
|
||||
hanlpPropertiesPath + FILE_SPILT + HanLP.Config.PerceptronParserModelPath;
|
||||
HanLP.Config.CRFSegmentModelPath = hanlpPropertiesPath + FILE_SPILT + HanLP.Config.CRFSegmentModelPath;
|
||||
HanLP.Config.HMMSegmentModelPath = hanlpPropertiesPath + FILE_SPILT + HanLP.Config.HMMSegmentModelPath;
|
||||
HanLP.Config.CRFCWSModelPath = hanlpPropertiesPath + FILE_SPILT + HanLP.Config.CRFCWSModelPath;
|
||||
HanLP.Config.CRFPOSModelPath = hanlpPropertiesPath + FILE_SPILT + HanLP.Config.CRFPOSModelPath;
|
||||
HanLP.Config.CRFNERModelPath = hanlpPropertiesPath + FILE_SPILT + HanLP.Config.CRFNERModelPath;
|
||||
HanLP.Config.PerceptronCWSModelPath = hanlpPropertiesPath + FILE_SPILT + HanLP.Config.PerceptronCWSModelPath;
|
||||
HanLP.Config.PerceptronPOSModelPath = hanlpPropertiesPath + FILE_SPILT + HanLP.Config.PerceptronPOSModelPath;
|
||||
HanLP.Config.PerceptronNERModelPath = hanlpPropertiesPath + FILE_SPILT + HanLP.Config.PerceptronNERModelPath;
|
||||
}
|
||||
|
||||
public static String getHanlpPropertiesPath() throws FileNotFoundException {
|
||||
return ResourceUtils.getFile("classpath:hanlp.properties").getParent();
|
||||
}
|
||||
|
||||
public static boolean addToCustomDictionary(DictWord dictWord) {
|
||||
log.info("dictWord:{}", dictWord);
|
||||
return getDynamicCustomDictionary().insert(dictWord.getWord(), dictWord.getNatureWithFrequency());
|
||||
}
|
||||
|
||||
public static void removeFromCustomDictionary(DictWord dictWord) {
|
||||
log.info("dictWord:{}", dictWord);
|
||||
CoreDictionary.Attribute attribute = getDynamicCustomDictionary().get(dictWord.getWord());
|
||||
if (attribute == null) {
|
||||
return;
|
||||
}
|
||||
log.info("get attribute:{}", attribute);
|
||||
getDynamicCustomDictionary().remove(dictWord.getWord());
|
||||
StringBuilder sb = new StringBuilder();
|
||||
List<Nature> natureList = new ArrayList<>();
|
||||
for (int i = 0; i < attribute.nature.length; i++) {
|
||||
if (!attribute.nature[i].toString().equals(dictWord.getNature())) {
|
||||
sb.append(attribute.nature[i].toString() + " ");
|
||||
sb.append(attribute.frequency[i] + " ");
|
||||
natureList.add((attribute.nature[i]));
|
||||
}
|
||||
}
|
||||
String natureWithFrequency = sb.toString();
|
||||
int len = natureWithFrequency.length();
|
||||
log.info("filtered natureWithFrequency:{}", natureWithFrequency);
|
||||
if (StringUtils.isNotBlank(natureWithFrequency)) {
|
||||
getDynamicCustomDictionary().add(dictWord.getWord(), natureWithFrequency.substring(0, len - 1));
|
||||
}
|
||||
SearchService.remove(dictWord, natureList.toArray(new Nature[0]));
|
||||
}
|
||||
|
||||
public static <T extends MapResult> void transLetterOriginal(List<T> mapResults) {
|
||||
if (CollectionUtils.isEmpty(mapResults)) {
|
||||
return;
|
||||
}
|
||||
for (T mapResult : mapResults) {
|
||||
if (MultiCustomDictionary.isLowerLetter(mapResult.getName())) {
|
||||
if (CustomDictionary.contains(mapResult.getName())) {
|
||||
CoreDictionary.Attribute attribute = CustomDictionary.get(mapResult.getName());
|
||||
if (attribute != null && attribute.original != null) {
|
||||
mapResult.setName(attribute.original);
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
public static List<Term> getTerms(String text) {
|
||||
return getSegment().seg(text.toLowerCase()).stream()
|
||||
.filter(term -> term.getNature().startsWith(DictWordType.NATURE_SPILT))
|
||||
.collect(Collectors.toList());
|
||||
}
|
||||
|
||||
}
|
||||
@@ -1,84 +0,0 @@
|
||||
package com.tencent.supersonic.chat.core.utils;
|
||||
|
||||
import static com.hankcs.hanlp.HanLP.Config.CustomDictionaryPath;
|
||||
|
||||
import com.hankcs.hanlp.dictionary.DynamicCustomDictionary;
|
||||
import com.hankcs.hanlp.utility.Predefine;
|
||||
import java.io.IOException;
|
||||
import java.net.URI;
|
||||
import java.util.ArrayList;
|
||||
import java.util.List;
|
||||
|
||||
import lombok.extern.slf4j.Slf4j;
|
||||
import org.apache.hadoop.conf.Configuration;
|
||||
import org.apache.hadoop.fs.FileStatus;
|
||||
import org.apache.hadoop.fs.FileSystem;
|
||||
import org.apache.hadoop.fs.Path;
|
||||
|
||||
/**
|
||||
* Hdfs File Helper
|
||||
*/
|
||||
@Slf4j
|
||||
public class HdfsFileHelper {
|
||||
|
||||
/***
|
||||
* delete cache file
|
||||
* @param path
|
||||
* @throws IOException
|
||||
*/
|
||||
public static void deleteCacheFile(String[] path) throws IOException {
|
||||
FileSystem fs = FileSystem.get(URI.create(path[0]), new Configuration());
|
||||
String cacheFilePath = path[0] + Predefine.BIN_EXT;
|
||||
log.info("delete cache file:{}", cacheFilePath);
|
||||
try {
|
||||
fs.delete(new Path(cacheFilePath), false);
|
||||
} catch (Exception e) {
|
||||
log.error("delete:" + cacheFilePath, e);
|
||||
}
|
||||
int customBase = cacheFilePath.lastIndexOf(FileHelper.FILE_SPILT);
|
||||
String customPath = cacheFilePath.substring(0, customBase) + FileHelper.FILE_SPILT + "*.bin";
|
||||
List<String> fileList = getFileList(fs, new Path(customPath));
|
||||
for (String file : fileList) {
|
||||
try {
|
||||
fs.delete(new Path(file), false);
|
||||
log.info("delete cache file:{}", file);
|
||||
} catch (Exception e) {
|
||||
log.error("delete " + file, e);
|
||||
}
|
||||
}
|
||||
log.info("fileList:{}", fileList);
|
||||
}
|
||||
|
||||
/**
|
||||
* reset path
|
||||
*
|
||||
* @param customDictionary
|
||||
* @throws IOException
|
||||
*/
|
||||
public static void resetCustomPath(DynamicCustomDictionary customDictionary) throws IOException {
|
||||
String[] path = CustomDictionaryPath;
|
||||
FileSystem fs = FileSystem.get(URI.create(path[0]), new Configuration());
|
||||
String cacheFilePath = path[0] + Predefine.BIN_EXT;
|
||||
int customBase = cacheFilePath.lastIndexOf(FileHelper.FILE_SPILT);
|
||||
String customPath = cacheFilePath.substring(0, customBase) + FileHelper.FILE_SPILT + "*.txt";
|
||||
log.info("customPath:{}", customPath);
|
||||
List<String> fileList = getFileList(fs, new Path(customPath));
|
||||
log.info("CustomDictionaryPath:{}", fileList);
|
||||
CustomDictionaryPath = fileList.toArray(new String[0]);
|
||||
customDictionary.path = (CustomDictionaryPath == null || CustomDictionaryPath.length == 0) ? path
|
||||
: CustomDictionaryPath;
|
||||
if (CustomDictionaryPath == null || CustomDictionaryPath.length == 0) {
|
||||
CustomDictionaryPath = path;
|
||||
}
|
||||
}
|
||||
|
||||
public static List<String> getFileList(FileSystem fs, Path folderPath) throws IOException {
|
||||
List<String> paths = new ArrayList();
|
||||
FileStatus[] fileStatuses = fs.globStatus(folderPath);
|
||||
for (int i = 0; i < fileStatuses.length; i++) {
|
||||
FileStatus fileStatus = fileStatuses[i];
|
||||
paths.add(fileStatus.getPath().toString());
|
||||
}
|
||||
return paths;
|
||||
}
|
||||
}
|
||||
@@ -1,180 +0,0 @@
|
||||
package com.tencent.supersonic.chat.core.utils;
|
||||
|
||||
import com.hankcs.hanlp.corpus.tag.Nature;
|
||||
import com.hankcs.hanlp.seg.common.Term;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElementType;
|
||||
import com.tencent.supersonic.chat.core.knowledge.ViewInfoStat;
|
||||
import com.tencent.supersonic.common.pojo.enums.DictWordType;
|
||||
import lombok.extern.slf4j.Slf4j;
|
||||
import org.apache.commons.lang3.StringUtils;
|
||||
|
||||
import java.util.ArrayList;
|
||||
import java.util.Comparator;
|
||||
import java.util.HashMap;
|
||||
import java.util.List;
|
||||
import java.util.Map;
|
||||
import java.util.Objects;
|
||||
import java.util.stream.Collectors;
|
||||
|
||||
/**
|
||||
* nature parse helper
|
||||
*/
|
||||
@Slf4j
|
||||
public class NatureHelper {
|
||||
|
||||
public static SchemaElementType convertToElementType(String nature) {
|
||||
DictWordType dictWordType = DictWordType.getNatureType(nature);
|
||||
if (Objects.isNull(dictWordType)) {
|
||||
return null;
|
||||
}
|
||||
SchemaElementType result = null;
|
||||
switch (dictWordType) {
|
||||
case METRIC:
|
||||
result = SchemaElementType.METRIC;
|
||||
break;
|
||||
case DIMENSION:
|
||||
result = SchemaElementType.DIMENSION;
|
||||
break;
|
||||
case ENTITY:
|
||||
result = SchemaElementType.ENTITY;
|
||||
break;
|
||||
case VIEW:
|
||||
result = SchemaElementType.VIEW;
|
||||
break;
|
||||
case VALUE:
|
||||
result = SchemaElementType.VALUE;
|
||||
break;
|
||||
case TAG:
|
||||
result = SchemaElementType.TAG;
|
||||
break;
|
||||
default:
|
||||
break;
|
||||
}
|
||||
return result;
|
||||
}
|
||||
|
||||
private static boolean isViewOrEntity(Term term, Integer model) {
|
||||
return (DictWordType.NATURE_SPILT + model).equals(term.nature.toString()) || term.nature.toString()
|
||||
.endsWith(DictWordType.ENTITY.getType());
|
||||
}
|
||||
|
||||
public static Integer getViewByNature(Nature nature) {
|
||||
if (nature.startsWith(DictWordType.NATURE_SPILT)) {
|
||||
String[] dimensionValues = nature.toString().split(DictWordType.NATURE_SPILT);
|
||||
if (StringUtils.isNumeric(dimensionValues[1])) {
|
||||
return Integer.valueOf(dimensionValues[1]);
|
||||
}
|
||||
}
|
||||
return 0;
|
||||
}
|
||||
|
||||
public static Long getViewId(String nature) {
|
||||
try {
|
||||
String[] split = nature.split(DictWordType.NATURE_SPILT);
|
||||
if (split.length <= 1) {
|
||||
return null;
|
||||
}
|
||||
return Long.valueOf(split[1]);
|
||||
} catch (NumberFormatException e) {
|
||||
log.error("", e);
|
||||
}
|
||||
return null;
|
||||
}
|
||||
|
||||
public static boolean isDimensionValueViewId(String nature) {
|
||||
if (StringUtils.isEmpty(nature)) {
|
||||
return false;
|
||||
}
|
||||
if (!nature.startsWith(DictWordType.NATURE_SPILT)) {
|
||||
return false;
|
||||
}
|
||||
String[] split = nature.split(DictWordType.NATURE_SPILT);
|
||||
if (split.length <= 1) {
|
||||
return false;
|
||||
}
|
||||
return !nature.endsWith(DictWordType.METRIC.getType()) && !nature.endsWith(DictWordType.DIMENSION.getType())
|
||||
&& StringUtils.isNumeric(split[1]);
|
||||
}
|
||||
|
||||
public static ViewInfoStat getViewStat(List<Term> terms) {
|
||||
return ViewInfoStat.builder()
|
||||
.viewCount(getViewCount(terms))
|
||||
.dimensionViewCount(getDimensionCount(terms))
|
||||
.metricViewCount(getMetricCount(terms))
|
||||
.dimensionValueViewCount(getDimensionValueCount(terms))
|
||||
.build();
|
||||
}
|
||||
|
||||
private static long getViewCount(List<Term> terms) {
|
||||
return terms.stream().filter(term -> isViewOrEntity(term, getViewByNature(term.nature))).count();
|
||||
}
|
||||
|
||||
private static long getDimensionValueCount(List<Term> terms) {
|
||||
return terms.stream().filter(term -> isDimensionValueViewId(term.nature.toString())).count();
|
||||
}
|
||||
|
||||
private static long getDimensionCount(List<Term> terms) {
|
||||
return terms.stream().filter(term -> term.nature.startsWith(DictWordType.NATURE_SPILT) && term.nature.toString()
|
||||
.endsWith(DictWordType.DIMENSION.getType())).count();
|
||||
}
|
||||
|
||||
private static long getMetricCount(List<Term> terms) {
|
||||
return terms.stream().filter(term -> term.nature.startsWith(DictWordType.NATURE_SPILT) && term.nature.toString()
|
||||
.endsWith(DictWordType.METRIC.getType())).count();
|
||||
}
|
||||
|
||||
/**
|
||||
* Get the number of types of class parts of speech
|
||||
* modelId -> (nature , natureCount)
|
||||
*
|
||||
* @param terms
|
||||
* @return
|
||||
*/
|
||||
public static Map<Long, Map<DictWordType, Integer>> getViewToNatureStat(List<Term> terms) {
|
||||
Map<Long, Map<DictWordType, Integer>> modelToNature = new HashMap<>();
|
||||
terms.stream().filter(
|
||||
term -> term.nature.startsWith(DictWordType.NATURE_SPILT)
|
||||
).forEach(term -> {
|
||||
DictWordType dictWordType = DictWordType.getNatureType(String.valueOf(term.nature));
|
||||
Long model = getViewId(String.valueOf(term.nature));
|
||||
|
||||
Map<DictWordType, Integer> natureTypeMap = new HashMap<>();
|
||||
natureTypeMap.put(dictWordType, 1);
|
||||
|
||||
Map<DictWordType, Integer> original = modelToNature.get(model);
|
||||
if (Objects.isNull(original)) {
|
||||
modelToNature.put(model, natureTypeMap);
|
||||
} else {
|
||||
Integer count = original.get(dictWordType);
|
||||
if (Objects.isNull(count)) {
|
||||
count = 1;
|
||||
} else {
|
||||
count = count + 1;
|
||||
}
|
||||
original.put(dictWordType, count);
|
||||
}
|
||||
});
|
||||
return modelToNature;
|
||||
}
|
||||
|
||||
public static List<Long> selectPossibleViews(List<Term> terms) {
|
||||
Map<Long, Map<DictWordType, Integer>> modelToNatureStat = getViewToNatureStat(terms);
|
||||
Integer maxViewTypeSize = modelToNatureStat.entrySet().stream()
|
||||
.max(Comparator.comparingInt(o -> o.getValue().size())).map(entry -> entry.getValue().size())
|
||||
.orElse(null);
|
||||
if (Objects.isNull(maxViewTypeSize) || maxViewTypeSize == 0) {
|
||||
return new ArrayList<>();
|
||||
}
|
||||
return modelToNatureStat.entrySet().stream().filter(entry -> entry.getValue().size() == maxViewTypeSize)
|
||||
.map(entry -> entry.getKey()).collect(Collectors.toList());
|
||||
}
|
||||
|
||||
public static Long getElementID(String nature) {
|
||||
String[] split = nature.split(DictWordType.NATURE_SPILT);
|
||||
if (split.length >= 3) {
|
||||
return Long.valueOf(split[2]);
|
||||
}
|
||||
return 0L;
|
||||
}
|
||||
|
||||
}
|
||||
@@ -1,7 +1,7 @@
|
||||
package com.tencent.supersonic.chat.core.utils;
|
||||
|
||||
import com.google.common.collect.Lists;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElement;
|
||||
import com.tencent.supersonic.headless.api.pojo.SchemaElement;
|
||||
import com.tencent.supersonic.chat.api.pojo.SemanticParseInfo;
|
||||
import com.tencent.supersonic.chat.core.query.QueryManager;
|
||||
import com.tencent.supersonic.common.pojo.Aggregator;
|
||||
|
||||
@@ -1,7 +1,7 @@
|
||||
package com.tencent.supersonic.chat.core.utils;
|
||||
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElementMatch;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElementType;
|
||||
import com.tencent.supersonic.headless.api.pojo.SchemaElementType;
|
||||
import java.util.List;
|
||||
import java.util.Set;
|
||||
import java.util.stream.Collectors;
|
||||
|
||||
@@ -1,2 +0,0 @@
|
||||
com.tencent.supersonic.chat.core.knowledge.FileHandler=\
|
||||
com.tencent.supersonic.chat.core.knowledge.LocalFileHandler
|
||||
@@ -1,7 +1,7 @@
|
||||
package com.tencent.supersonic.chat.core.s2sql;
|
||||
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaElement;
|
||||
import com.tencent.supersonic.chat.api.pojo.SchemaValueMap;
|
||||
import com.tencent.supersonic.headless.api.pojo.SchemaElement;
|
||||
import com.tencent.supersonic.headless.api.pojo.SchemaValueMap;
|
||||
import com.tencent.supersonic.chat.api.pojo.SemanticSchema;
|
||||
import org.junit.jupiter.api.Test;
|
||||
import org.mockito.Mockito;
|
||||
|
||||
Reference in New Issue
Block a user