(improvement)(chat) Clean up unused code and optimize some code and make the tests pass (#595)

This commit is contained in:
lexluo09
2024-01-04 18:22:40 +08:00
committed by GitHub
parent 023e84c420
commit ade96c3adc
18 changed files with 38 additions and 177 deletions

View File

@@ -20,10 +20,8 @@ import java.util.TreeSet;
import java.util.stream.Collectors;
import lombok.extern.slf4j.Slf4j;
import org.apache.commons.lang3.StringUtils;
import org.springframework.stereotype.Service;
import org.springframework.util.CollectionUtils;
@Service
@Slf4j
public class SearchService {

View File

@@ -5,7 +5,6 @@ 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.ModelResolver;
import com.tencent.supersonic.common.util.ContextUtils;
import java.util.List;
import java.util.Map;
import java.util.Objects;
import lombok.extern.slf4j.Slf4j;
@@ -52,12 +51,6 @@ public class ComponentFactory {
return modelResolver;
}
private static <T> List<T> init(Class<T> factoryType, List list) {
list.addAll(SpringFactoriesLoader.loadFactories(factoryType,
Thread.currentThread().getContextClassLoader()));
return list;
}
private static <T> T init(Class<T> factoryType) {
return SpringFactoriesLoader.loadFactories(factoryType,
Thread.currentThread().getContextClassLoader()).get(0);

View File

@@ -0,0 +1,15 @@
package com.tencent.supersonic.chat.core.mapper;
import com.hankcs.hanlp.algorithm.EditDistance;
import org.junit.Assert;
import org.junit.jupiter.api.Test;
class LoadRemoveServiceTest {
@Test
void edit() {
int compute = EditDistance.compute("", "在你的身边");
Assert.assertEquals(compute, 4);
}
}

View File

@@ -0,0 +1,38 @@
package com.tencent.supersonic.chat.core.parser.aggregate;
import com.tencent.supersonic.chat.core.parser.sql.rule.AggregateTypeParser;
import com.tencent.supersonic.common.pojo.enums.AggregateTypeEnum;
import org.junit.Assert;
import org.junit.jupiter.api.Test;
class AggregateTypeParserTest {
@Test
void getAggregateParser() {
AggregateTypeParser aggregateParser = new AggregateTypeParser();
AggregateTypeEnum aggregateType = aggregateParser.resolveAggregateType("supsersonic产品访问次数最大值");
Assert.assertEquals(aggregateType, AggregateTypeEnum.MAX);
aggregateType = aggregateParser.resolveAggregateType("supsersonic产品pv");
Assert.assertEquals(aggregateType, AggregateTypeEnum.COUNT);
aggregateType = aggregateParser.resolveAggregateType("supsersonic产品uv");
Assert.assertEquals(aggregateType, AggregateTypeEnum.DISTINCT);
aggregateType = aggregateParser.resolveAggregateType("supsersonic产品访问次数最大值");
Assert.assertEquals(aggregateType, AggregateTypeEnum.MAX);
aggregateType = aggregateParser.resolveAggregateType("supsersonic产品访问次数最小值");
Assert.assertEquals(aggregateType, AggregateTypeEnum.MIN);
aggregateType = aggregateParser.resolveAggregateType("supsersonic产品访问次数平均值");
Assert.assertEquals(aggregateType, AggregateTypeEnum.AVG);
aggregateType = aggregateParser.resolveAggregateType("supsersonic产品访问次数topN");
Assert.assertEquals(aggregateType, AggregateTypeEnum.TOPN);
aggregateType = aggregateParser.resolveAggregateType("supsersonic产品访问次数汇总");
Assert.assertEquals(aggregateType, AggregateTypeEnum.SUM);
}
}

View File

@@ -0,0 +1,55 @@
package com.tencent.supersonic.chat.core.s2sql;
import com.tencent.supersonic.chat.core.parser.sql.llm.LLMResponseService;
import com.tencent.supersonic.chat.core.query.llm.s2sql.LLMSqlResp;
import com.tencent.supersonic.chat.core.query.llm.s2sql.LLMResp;
import java.util.HashMap;
import java.util.Map;
import org.junit.Assert;
import org.junit.jupiter.api.Test;
class LLMResponseServiceTest {
@Test
void deduplicationSqlWeight() {
String sql1 = "SELECT a,b,c,d FROM table1 WHERE column1 = 1 AND column2 = 2 order by a";
String sql2 = "SELECT d,c,b,a FROM table1 WHERE column2 = 2 AND column1 = 1 order by a";
LLMResp llmResp = new LLMResp();
Map<String, LLMSqlResp> sqlWeight = new HashMap<>();
sqlWeight.put(sql1, LLMSqlResp.builder().sqlWeight(0.20).build());
sqlWeight.put(sql2, LLMSqlResp.builder().sqlWeight(0.80).build());
llmResp.setSqlRespMap(sqlWeight);
LLMResponseService llmResponseService = new LLMResponseService();
Map<String, LLMSqlResp> deduplicationSqlResp = llmResponseService.getDeduplicationSqlResp(llmResp);
Assert.assertEquals(deduplicationSqlResp.size(), 1);
sql1 = "SELECT a,b,c,d FROM table1 WHERE column1 = 1 AND column2 = 2 order by a";
sql2 = "SELECT d,c,b,a FROM table1 WHERE column2 = 2 AND column1 = 1 order by a";
LLMResp llmResp2 = new LLMResp();
Map<String, LLMSqlResp> sqlWeight2 = new HashMap<>();
sqlWeight2.put(sql1, LLMSqlResp.builder().sqlWeight(0.20).build());
sqlWeight2.put(sql2, LLMSqlResp.builder().sqlWeight(0.80).build());
llmResp2.setSqlRespMap(sqlWeight2);
deduplicationSqlResp = llmResponseService.getDeduplicationSqlResp(llmResp2);
Assert.assertEquals(deduplicationSqlResp.size(), 1);
sql1 = "SELECT a,b,c,d,e FROM table1 WHERE column1 = 1 AND column2 = 2 order by a";
sql2 = "SELECT d,c,b,a FROM table1 WHERE column2 = 2 AND column1 = 1 order by a";
LLMResp llmResp3 = new LLMResp();
Map<String, LLMSqlResp> sqlWeight3 = new HashMap<>();
sqlWeight3.put(sql1, LLMSqlResp.builder().sqlWeight(0.20).build());
sqlWeight3.put(sql2, LLMSqlResp.builder().sqlWeight(0.80).build());
llmResp3.setSqlRespMap(sqlWeight3);
deduplicationSqlResp = llmResponseService.getDeduplicationSqlResp(llmResp3);
Assert.assertEquals(deduplicationSqlResp.size(), 2);
}
}

View File

@@ -0,0 +1,55 @@
package com.tencent.supersonic.chat.core.s2sql;
import static org.mockito.Mockito.when;
import com.tencent.supersonic.chat.api.pojo.SchemaElement;
import com.tencent.supersonic.chat.api.pojo.SchemaValueMap;
import com.tencent.supersonic.chat.api.pojo.SemanticSchema;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.List;
import org.junit.jupiter.api.Test;
import org.mockito.Mockito;
class LLMSqlParserTest {
@Test
void setFilter() {
SemanticSchema mockSemanticSchema = Mockito.mock(SemanticSchema.class);
List<SchemaElement> dimensions = new ArrayList<>();
List<SchemaValueMap> schemaValueMaps = new ArrayList<>();
SchemaValueMap value1 = new SchemaValueMap();
value1.setBizName("杰伦");
value1.setTechName("周杰伦");
value1.setAlias(Arrays.asList("周杰倫", "Jay Chou", "周董", "周先生"));
schemaValueMaps.add(value1);
SchemaElement schemaElement = SchemaElement.builder()
.bizName("singer_name")
.name("歌手名")
.model(2L)
.schemaValueMaps(schemaValueMaps)
.build();
dimensions.add(schemaElement);
SchemaElement schemaElement2 = SchemaElement.builder()
.bizName("publish_time")
.name("发布时间")
.model(2L)
.build();
dimensions.add(schemaElement2);
when(mockSemanticSchema.getDimensions()).thenReturn(dimensions);
List<SchemaElement> metrics = new ArrayList<>();
SchemaElement metric = SchemaElement.builder()
.bizName("play_count")
.name("播放量")
.model(2L)
.build();
metrics.add(metric);
when(mockSemanticSchema.getMetrics()).thenReturn(metrics);
}
}