Refactor translator module (#1932)

* [improvement][Chat] Support agent permission management #1143

* [improvement][chat]Iterate LLM prompts of parsing and correction.

* [improvement][headless]Clean code logic of headless core.

* (fix) (chat) 记忆管理更新不生效 (#1912)

* [improvement][headless-fe] Added null-check conditions to the data formatting function.

* [improvement][headless]Clean code logic of headless translator.

* [improvement][headless-fe] Added permissions management for agents.

* [improvement][headless-fe] Unified the assistant's permission settings interaction to match the system style.

* [improvement](Dict)Support returns dict task list of dimensions by page

* [improvement][headless-fe] Revised the interaction for semantic modeling routing and implemented the initial version of metric management switching.

* [improvement][launcher]Set system property `s2.test` in junit tests in order to facilitate conditional breakpoints.

* [improvement][headless] add validateAndQuery interface in SqlQueryApiController

* [improvement][launcher]Use API to get element ID avoiding hard-code.

* [improvement][launcher]Support DuckDB database and refactor translator code structure.

---------

Co-authored-by: lxwcodemonkey <jolunoluo@tencent.com>
Co-authored-by: tristanliu <tristanliu@tencent.com>
Co-authored-by: daikon12 <1059907724@qq.com>
Co-authored-by: lexluo09 <39718951+lexluo09@users.noreply.github.com>
This commit is contained in:
Jun Zhang
2024-11-30 00:27:33 +08:00
committed by GitHub
parent 224c114d20
commit 593597fe26
155 changed files with 3757 additions and 3551 deletions

View File

@@ -108,9 +108,9 @@ public abstract class S2BaseDemo implements CommandLineRunner {
}
}
abstract void doRun();
protected abstract void doRun();
abstract boolean checkNeedToRun();
protected abstract boolean checkNeedToRun();
protected DatabaseResp addDatabaseIfNotExist() {
List<DatabaseResp> databaseList = databaseService.getDatabaseList(defaultUser);
@@ -119,8 +119,8 @@ public abstract class S2BaseDemo implements CommandLineRunner {
}
String url = dataSourceProperties.getUrl();
DatabaseReq databaseReq = new DatabaseReq();
databaseReq.setName("H2数据库DEMO");
databaseReq.setDescription("样例数据库实例仅用于体验,正式使用请切换持久化数据库");
databaseReq.setName("S2数据库DEMO");
databaseReq.setDescription("样例数据库实例仅用于体验");
if (StringUtils.isNotBlank(url)
&& url.toLowerCase().contains(DataType.MYSQL.getFeature().toLowerCase())) {
databaseReq.setType(DataType.MYSQL.getFeature());

View File

@@ -7,26 +7,12 @@ import com.tencent.supersonic.chat.server.agent.Agent;
import com.tencent.supersonic.chat.server.agent.AgentToolType;
import com.tencent.supersonic.chat.server.agent.DatasetTool;
import com.tencent.supersonic.chat.server.agent.ToolConfig;
import com.tencent.supersonic.chat.server.processor.execute.DataInterpretProcessor;
import com.tencent.supersonic.common.pojo.ChatApp;
import com.tencent.supersonic.common.pojo.JoinCondition;
import com.tencent.supersonic.common.pojo.ModelRela;
import com.tencent.supersonic.common.pojo.enums.AggOperatorEnum;
import com.tencent.supersonic.common.pojo.enums.AppModule;
import com.tencent.supersonic.common.pojo.enums.FilterOperatorEnum;
import com.tencent.supersonic.common.pojo.enums.TimeMode;
import com.tencent.supersonic.common.pojo.enums.TypeEnums;
import com.tencent.supersonic.common.pojo.enums.*;
import com.tencent.supersonic.common.util.ChatAppManager;
import com.tencent.supersonic.headless.api.pojo.AggregateTypeDefaultConfig;
import com.tencent.supersonic.headless.api.pojo.DataSetDetail;
import com.tencent.supersonic.headless.api.pojo.DataSetModelConfig;
import com.tencent.supersonic.headless.api.pojo.Dim;
import com.tencent.supersonic.headless.api.pojo.DimensionTimeTypeParams;
import com.tencent.supersonic.headless.api.pojo.Identify;
import com.tencent.supersonic.headless.api.pojo.Measure;
import com.tencent.supersonic.headless.api.pojo.ModelDetail;
import com.tencent.supersonic.headless.api.pojo.QueryConfig;
import com.tencent.supersonic.headless.api.pojo.TimeDefaultConfig;
import com.tencent.supersonic.headless.api.pojo.*;
import com.tencent.supersonic.headless.api.pojo.enums.DimensionType;
import com.tencent.supersonic.headless.api.pojo.enums.IdentifyType;
import com.tencent.supersonic.headless.api.pojo.request.DataSetReq;
@@ -40,11 +26,7 @@ import lombok.extern.slf4j.Slf4j;
import org.springframework.core.annotation.Order;
import org.springframework.stereotype.Component;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.Collections;
import java.util.List;
import java.util.Map;
import java.util.*;
@Component
@Slf4j
@@ -59,8 +41,8 @@ public class S2CompanyDemo extends S2BaseDemo {
ModelResp model_brand = addModel_2(domain, demoDatabase);
ModelResp model_brand_revenue = addModel_3(domain, demoDatabase);
addModelRela(domain, model_company, model_brand, "company_id");
addModelRela(domain, model_brand, model_brand_revenue, "brand_id");
addModelRela(domain, model_brand, model_company, "company_id");
addModelRela(domain, model_brand_revenue, model_brand, "brand_id");
DataSetResp dataset = addDataSet(domain);
addAgent(dataset.getId());
@@ -70,7 +52,7 @@ public class S2CompanyDemo extends S2BaseDemo {
}
@Override
boolean checkNeedToRun() {
protected boolean checkNeedToRun() {
List<DomainResp> domainList = domainService.getDomainList();
for (DomainResp domainResp : domainList) {
if (domainResp.getBizName().equalsIgnoreCase("corporate")) {
@@ -124,8 +106,7 @@ public class S2CompanyDemo extends S2BaseDemo {
modelDetail.setMeasures(measures);
modelDetail.setQueryType("sql_query");
modelDetail.setSqlQuery("SELECT company_id,company_name,headquarter_address,"
+ "company_established_time,founder,ceo,annual_turnover,employee_count FROM company");
modelDetail.setSqlQuery("SELECT * FROM company");
modelReq.setModelDetail(modelDetail);
ModelResp companyModel = modelService.createModel(modelReq, defaultUser);
@@ -164,8 +145,7 @@ public class S2CompanyDemo extends S2BaseDemo {
modelDetail.setMeasures(measures);
modelDetail.setQueryType("sql_query");
modelDetail.setSqlQuery("SELECT brand_id,brand_name,brand_established_time,"
+ "company_id,legal_representative,registered_capital FROM brand");
modelDetail.setSqlQuery("SELECT * FROM brand");
modelReq.setModelDetail(modelDetail);
ModelResp brandModel = modelService.createModel(modelReq, defaultUser);
@@ -205,8 +185,7 @@ public class S2CompanyDemo extends S2BaseDemo {
modelDetail.setMeasures(measures);
modelDetail.setQueryType("sql_query");
modelDetail.setSqlQuery("SELECT year_time,brand_id,revenue,profit,"
+ "revenue_growth_year_on_year,profit_growth_year_on_year FROM brand_revenue");
modelDetail.setSqlQuery("SELECT * FROM brand_revenue");
modelReq.setModelDetail(modelDetail);
return modelService.createModel(modelReq, defaultUser);
}
@@ -245,7 +224,7 @@ public class S2CompanyDemo extends S2BaseDemo {
modelRelaReq.setDomainId(domain.getId());
modelRelaReq.setFromModelId(fromModel.getId());
modelRelaReq.setToModelId(toModel.getId());
modelRelaReq.setJoinType("left join");
modelRelaReq.setJoinType("inner join");
modelRelaReq.setJoinConditions(joinConditions);
modelRelaService.save(modelRelaReq, defaultUser);
}
@@ -272,7 +251,6 @@ public class S2CompanyDemo extends S2BaseDemo {
Map<String, ChatApp> chatAppConfig =
Maps.newHashMap(ChatAppManager.getAllApps(AppModule.CHAT));
chatAppConfig.values().forEach(app -> app.setChatModelId(demoChatModel.getId()));
chatAppConfig.get(DataInterpretProcessor.APP_KEY).setEnable(true);
agent.setChatAppConfig(chatAppConfig);
agentService.createAgent(agent, defaultUser);

View File

@@ -57,7 +57,7 @@ public class S2SingerDemo extends S2BaseDemo {
}
@Override
boolean checkNeedToRun() {
protected boolean checkNeedToRun() {
List<DomainResp> domainList = domainService.getDomainList();
for (DomainResp domainResp : domainList) {
if (domainResp.getBizName().equalsIgnoreCase("singer")) {

View File

@@ -46,9 +46,10 @@ public class S2SmallTalkDemo extends S2BaseDemo {
}
@Override
boolean checkNeedToRun() {
protected boolean checkNeedToRun() {
List<String> agentNames =
agentService.getAgents().stream().map(Agent::getName).collect(Collectors.toList());
return !agentNames.contains("闲聊助手");
}
}

View File

@@ -79,8 +79,8 @@ public class S2VisitsDemo extends S2BaseDemo {
ModelResp userModel = addModel_1(s2Domain, demoDatabase);
ModelResp pvUvModel = addModel_2(s2Domain, demoDatabase);
ModelResp stayTimeModel = addModel_3(s2Domain, demoDatabase);
addModelRela(s2Domain, userModel, pvUvModel, "user_name");
addModelRela(s2Domain, userModel, stayTimeModel, "user_name");
addModelRela(s2Domain, pvUvModel, userModel, "user_name");
addModelRela(s2Domain, stayTimeModel, userModel, "user_name");
// create metrics and dimensions
DimensionResp departmentDimension = getDimension("department", userModel);
@@ -146,7 +146,8 @@ public class S2VisitsDemo extends S2BaseDemo {
agent.setStatus(1);
agent.setEnableSearch(1);
agent.setExamples(Lists.newArrayList("近15天超音数访问次数汇总", "按部门统计超音数的访问人数", "对比alice和lucy的停留时长",
"过去30天访问次数最高的部门top3", "近1个月总访问次数超过100次的部门有几个", "过去半个月每个核心用户的总停留时长"));
"过去30天访问次数最高的部门top3", "近1个月总访问次数超过100次的部门有几个", "过去半个月每个核心用户的总停留时长",
"今年以来访问次数最高的一天是哪一天"));
// configure tools
ToolConfig toolConfig = new ToolConfig();
@@ -198,6 +199,7 @@ public class S2VisitsDemo extends S2BaseDemo {
List<Dim> dimensions = new ArrayList<>();
dimensions.add(new Dim("部门", "department", DimensionType.categorical, 1));
// dimensions.add(new Dim("用户", "user_name", DimensionType.categorical, 1));
modelDetail.setDimensions(dimensions);
List<Field> fields = Lists.newArrayList();
fields.add(Field.builder().fieldName("user_name").dataType("Varchar").build());
@@ -382,9 +384,9 @@ public class S2VisitsDemo extends S2BaseDemo {
metricReq.setDescription("访问的用户个数");
metricReq.setAlias("UV,访问人数");
MetricDefineByFieldParams metricTypeParams = new MetricDefineByFieldParams();
metricTypeParams.setExpr("count(distinct user_id)");
metricTypeParams.setExpr("count(distinct user_name)");
List<FieldParam> fieldParams = new ArrayList<>();
fieldParams.add(new FieldParam("user_id"));
fieldParams.add(new FieldParam("user_name"));
metricTypeParams.setFields(fieldParams);
metricReq.setMetricDefineByFieldParams(metricTypeParams);
metricReq.setMetricDefineType(MetricDefineType.FIELD);

View File

@@ -26,15 +26,18 @@ com.tencent.supersonic.headless.chat.parser.llm.DataSetResolver=\
com.tencent.supersonic.headless.core.translator.converter.QueryConverter=\
com.tencent.supersonic.headless.core.translator.converter.DefaultDimValueConverter,\
com.tencent.supersonic.headless.core.translator.converter.SqlVariableParseConverter,\
com.tencent.supersonic.headless.core.translator.converter.CalculateAggConverter,\
com.tencent.supersonic.headless.core.translator.converter.ParserDefaultConverter
com.tencent.supersonic.headless.core.translator.converter.SqlVariableConverter,\
com.tencent.supersonic.headless.core.translator.converter.MetricRatioConverter,\
com.tencent.supersonic.headless.core.translator.converter.SqlQueryConverter,\
com.tencent.supersonic.headless.core.translator.converter.StructQueryConverter
com.tencent.supersonic.headless.core.translator.QueryOptimizer=\
com.tencent.supersonic.headless.core.translator.DetailQueryOptimizer
com.tencent.supersonic.headless.core.translator.optimizer.QueryOptimizer=\
com.tencent.supersonic.headless.core.translator.optimizer.DetailQueryOptimizer,\
com.tencent.supersonic.headless.core.translator.optimizer.DbDialectOptimizer,\
com.tencent.supersonic.headless.core.translator.optimizer.ResultLimitOptimizer
com.tencent.supersonic.headless.core.translator.QueryParser=\
com.tencent.supersonic.headless.core.translator.calcite.CalciteQueryParser
com.tencent.supersonic.headless.core.translator.parser.QueryParser=\
com.tencent.supersonic.headless.core.translator.parser.calcite.CalciteQueryParser
com.tencent.supersonic.headless.core.executor.QueryExecutor=\
com.tencent.supersonic.headless.core.executor.JdbcExecutor

View File

@@ -11,4 +11,24 @@ spring:
h2:
console:
path: /h2-console/semantic
enabled: true
enabled: true
### Comment out following lines if using MySQL
#spring:
# datasource:
# driver-class-name: com.mysql.cj.jdbc.Driver
# url: jdbc:mysql://localhost:3306/s2_database?user=root
# username: root
# password:
# sql:
# enabled: true
# mode: always
# username: root
# password:
# init:
# schema-locations: classpath:db/schema-mysql.sql,classpath:db/schema-mysql-demo.sql
# data-locations: classpath:db/data-mysql.sql,classpath:db/data-mysql-demo.sql
# h2:
# console:
# path: /h2-console/semantic
# enabled: true

View File

@@ -396,5 +396,5 @@ ALTER TABLE s2_agent DROP COLUMN `enable_memory_review`;
alter table s2_agent add column `enable_feedback` tinyint DEFAULT 1;
--20241116
alter table s2_agent add column `admin` varchar(1000);
alter table s2_agent add column `viewer` varchar(1000);
alter table s2_agent add column `admin` varchar(1000) COLLATE utf8_unicode_ci DEFAULT NULL;
alter table s2_agent add column `viewer` varchar(1000) COLLATE utf8_unicode_ci DEFAULT NULL;

View File

@@ -1,4 +1,4 @@
-------S2VisitsDemo
-- S2VisitsDemo
insert into s2_user_department (user_name, department) values ('jack','HR');
insert into s2_user_department (user_name, department) values ('tom','sales');
insert into s2_user_department (user_name, department) values ('lucy','marketing');
@@ -1019,7 +1019,7 @@ INSERT INTO s2_stay_time_statis (imp_date, user_name, stay_hours, page) VALUES (
INSERT INTO s2_stay_time_statis (imp_date, user_name, stay_hours, page) VALUES (DATE_SUB(CURRENT_DATE(), INTERVAL 15 DAY), 'lucy', '0.8124302447925607', 'p4');
INSERT INTO s2_stay_time_statis (imp_date, user_name, stay_hours, page) VALUES (DATE_SUB(CURRENT_DATE(), INTERVAL 8 DAY), 'lucy', '0.039935860913407284', 'p2');
-------S2ArtistDemo
-- S2ArtistDemo
INSERT INTO singer (singer_name, act_area, song_name, genre, js_play_cnt, down_cnt, favor_cnt)
VALUES ('周杰伦', '港台', '青花瓷', '国风', 1000000, 1000000, 1000000);

View File

@@ -1,4 +1,4 @@
-------S2VisitsDemo
-- S2VisitsDemo
CREATE TABLE IF NOT EXISTS `s2_user_department` (
`user_name` varchar(200) NOT NULL,
`department` varchar(200) NOT NULL
@@ -27,7 +27,7 @@ CREATE TABLE IF NOT EXISTS `singer` (
`favor_cnt` bigint DEFAULT NULL
)ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_unicode_ci;
-------S2ArtistDemo
-- S2ArtistDemo
CREATE TABLE IF NOT EXISTS `genre` (
`g_name` varchar(20) NOT NULL , -- genre name
`rating` INT ,

View File

@@ -15,6 +15,8 @@ CREATE TABLE IF NOT EXISTS `s2_agent` (
`created_at` datetime DEFAULT NULL,
`updated_by` varchar(100) COLLATE utf8_unicode_ci DEFAULT NULL,
`updated_at` datetime DEFAULT NULL,
`admin` varchar(1000) COLLATE utf8_unicode_ci DEFAULT NULL,
`viewer` varchar(1000) COLLATE utf8_unicode_ci DEFAULT NULL,
PRIMARY KEY (`id`)
) ENGINE=InnoDB DEFAULT CHARSET=utf8 COLLATE=utf8_unicode_ci;
@@ -539,7 +541,7 @@ CREATE TABLE IF NOT EXISTS `s2_term` (
PRIMARY KEY (`id`)
) ENGINE = InnoDB DEFAULT CHARSET = utf8 COMMENT ='术语表';
CREATE TABLE `s2_user_token` (
CREATE TABLE IF NOT EXISTS `s2_user_token` (
`id` bigint NOT NULL AUTO_INCREMENT,
`name` VARCHAR(255) NOT NULL,
`user_name` VARCHAR(255) NOT NULL,