feat: add support examples (including Q&A SQL pairs, synonym pairs, a…

README.md

@@ -486,9 +486,9 @@ Closes #123"
 - 5、支持将智能问数项目对外提供OpenAPI的服务；
 - 6、支持将智能问数项目对外提供MCP的服务；
 - 7、支持seed命令可以将CSV文件初始化加载入数据库；
-- 8、基于LLM的数据探查辅助生成语义模型；（TODO）
-- 9、数据模型、语义模型、智能问数的单元测试；（TODO）
-- 10、SQL问答对、同义词、业务知识等向量化入库与检索；（TODO）
+- 8、SQL问答对、同义词、业务知识等向量化入库与检索；
+- 9、基于LLM的数据探查辅助生成语义模型；（TODO）
+- 10、数据模型、语义模型、智能问数的单元测试；（TODO）
 - 11、指标的配置（构建语义模型后可以更进一步添加指标）；（TODO）
 
 

dat-cli/src/main/resources/project_init_template/models/examples/knowledge_examples1.yaml

@@ -0,0 +1,8 @@
+examples:
+  knowledge:
+    - |
+      COVID-19（新型冠状病毒肺炎）是由SARS-CoV-2病毒引起的呼吸道传染病。
+      主要传播途径为飞沫和密切接触，在相对封闭的环境中可经气溶胶传播。
+      典型症状包括发热、干咳、乏力等，部分患者可能丧失嗅觉或味觉。值得注意的是，许多感染者症状轻微甚至无症状，但仍具有传染性。
+      预防措施至关重要：接种疫苗、在人员密集场所佩戴口罩、勤洗手、多通风。若出现症状，应及时进行检测并自我隔离，保护他人。
+      虽然大多数患者可康复，但对老年人和有基础疾病者威胁较大，仍需保持警惕。

dat-cli/src/main/resources/project_init_template/models/examples/knowledge_examples2.yaml

@@ -0,0 +1,8 @@
+examples:
+  knowledge:
+    - |
+      COVID 预防是关键：
+      接种疫苗：有效降低重症和死亡风险。
+      良好卫生：勤洗手、戴口罩、常通风。
+      保持距离：避免人群聚集。
+      虽然目前毒株毒性减弱，但基础防护对保护高危人群（如老年人、有基础疾病者）仍至关重要。若出现症状请及时就医。

dat-cli/src/main/resources/project_init_template/models/examples/sql_examples.yaml

@@ -0,0 +1,12 @@
+examples:
+  sql_pairs:
+    - question: 各个国家的平均病例数
+      sql: |
+        SELECT country_covid_cases.country, AVG(country_covid_cases.cases_total) AS average_cases 
+        FROM country_covid_cases 
+        GROUP BY country_covid_cases.country
+    - question: 各个国家的平均病例数
+      sql: |
+        SELECT country_covid_cases.country, AVG(country_covid_cases.cases_total) AS average_cases 
+        FROM country_covid_cases 
+        GROUP BY country_covid_cases.country

dat-cli/src/main/resources/project_init_template/models/examples/synonyms_examples.yaml

@@ -0,0 +1,15 @@
+examples:
+  synonyms_pairs:
+    - word: COVID-19
+      synonyms:
+        - 新型冠状病毒肺炎
+        - 2019冠状病毒病
+        - 新冠
+        - 新冠肺炎
+        - 冠状病毒/新冠
+    - word: UnitedStates
+      synonyms:
+        - USA
+        - 美国
+        - 美利坚合众国
+        - The United States of America

dat-core/src/main/java/ai/dat/core/contentstore/ContentStore.java

@@ -62,7 +62,7 @@ default void removeSql(String id) {
 
     void removeAllSqls();
 
-    // ---------------名词和同义词对-------------------
+    // ---------------词和同义词对-------------------
 
     default String addSyn(WordSynonymPair synonymPair) {
         return addSyns(List.of(synonymPair)).get(0);

dat-core/src/main/java/ai/dat/core/contentstore/data/WordSynonymPair.java

@@ -20,47 +20,28 @@
 @JsonInclude(JsonInclude.Include.NON_NULL)
 public class WordSynonymPair {
     @NonNull
-    private final String noun;
+    private final String word;
 
     @NonNull
     private final List<String> synonyms;
 
-    private String description;
-
     @JsonCreator
-    private WordSynonymPair(@JsonProperty("noun") @NonNull String noun,
+    private WordSynonymPair(@JsonProperty("word") @NonNull String word,
                             @JsonProperty("synonyms") @NonNull List<String> synonyms) {
-        this.noun = noun;
-        this.synonyms = synonyms;
-    }
-
-    @JsonCreator
-    private WordSynonymPair(@JsonProperty("noun") @NonNull String noun,
-                            @JsonProperty("synonyms") @NonNull List<String> synonyms,
-                            @JsonProperty("description") String description) {
-        this.noun = noun;
+        this.word = word;
         this.synonyms = synonyms;
-        this.description = description;
-    }
-
-    public static WordSynonymPair from(@NonNull String noun, @NonNull String synonym, String description) {
-        return new WordSynonymPair(noun, Collections.singletonList(synonym), description);
-    }
-
-    public static WordSynonymPair from(@NonNull String noun, @NonNull List<String> synonyms, String description) {
-        return new WordSynonymPair(noun, synonyms, description);
     }
 
-    public static WordSynonymPair from(@NonNull String noun, @NonNull String synonym) {
-        return new WordSynonymPair(noun, Collections.singletonList(synonym));
+    public static WordSynonymPair from(@NonNull String word, @NonNull String synonym) {
+        return new WordSynonymPair(word, Collections.singletonList(synonym));
     }
 
-    public static WordSynonymPair from(@NonNull String noun, @NonNull List<String> synonyms) {
-        return new WordSynonymPair(noun, synonyms);
+    public static WordSynonymPair from(@NonNull String word, @NonNull List<String> synonyms) {
+        return new WordSynonymPair(word, synonyms);
     }
 
-    public static WordSynonymPair from(@NonNull String noun, @NonNull String... synonyms) {
-        return new WordSynonymPair(noun, Arrays.asList(synonyms));
+    public static WordSynonymPair from(@NonNull String word, @NonNull String... synonyms) {
+        return new WordSynonymPair(word, Arrays.asList(synonyms));
     }
 
     public void add(@NonNull String synonym) {

dat-core/src/main/resources/prompts/default/intent_classification_user_prompt_template.txt

@@ -12,9 +12,9 @@ SQL: {{ item.sql }}
 {% endif %}
 
 {% if synonyms %}
-### NOUN AND SYNONYMS ###
+### WORD AND SYNONYMS ###
 {% for item in synonyms %}
-Noun: {{ item.noun }} Synonyms: {{ item.synonyms|join(', ') }}
+Word: {{ item.noun }} Synonyms: {{ item.synonyms|join(', ') }}
 {% endfor %}
 {% endif %}
 

dat-core/src/main/resources/prompts/default/sql_generation_reasoning_user_prompt_template.txt

@@ -12,9 +12,9 @@ SQL: {{ item.sql }}
 {% endif %}
 
 {% if synonyms %}
-### NOUN AND SYNONYMS ###
+### WORD AND SYNONYMS ###
 {% for item in synonyms %}
-Noun: {{ item.noun }} Synonyms: {{ item.synonyms|join(', ') }}
+Word: {{ item.noun }} Synonyms: {{ item.synonyms|join(', ') }}
 {% endfor %}
 {% endif %}
 

dat-core/src/main/resources/prompts/default/sql_generation_reasoning_with_followup_user_prompt_template.txt

@@ -12,9 +12,9 @@ SQL: {{ item.sql }}
 {% endif %}
 
 {% if synonyms %}
-### NOUN AND SYNONYMS ###
+### WORD AND SYNONYMS ###
 {% for item in synonyms %}
-Noun: {{ item.noun }} Synonyms: {{ item.synonyms|join(', ') }}
+Word: {{ item.noun }} Synonyms: {{ item.synonyms|join(', ') }}
 {% endfor %}
 {% endif %}
 

dat-core/src/main/resources/prompts/default/sql_generation_user_prompt_template.txt

@@ -12,9 +12,9 @@ SQL: {{ item.sql }}
 {% endif %}
 
 {% if synonyms %}
-### NOUN AND SYNONYMS ###
+### WORD AND SYNONYMS ###
 {% for item in synonyms %}
-Noun: {{ item.noun }} Synonyms: {{ item.synonyms|join(', ') }}
+Word: {{ item.noun }} Synonyms: {{ item.synonyms|join(', ') }}
 {% endfor %}
 {% endif %}
 

dat-core/src/main/resources/prompts/default/sql_generation_with_followup_user_prompt_template.txt

@@ -18,9 +18,9 @@ SQL:
 {% endif %}
 
 {% if synonyms %}
-### NOUN AND SYNONYMS ###
+### WORD AND SYNONYMS ###
 {% for item in synonyms %}
-Noun: {{ item.noun }} Synonyms: {{ item.synonyms|join(', ') }}
+Word: {{ item.noun }} Synonyms: {{ item.synonyms|join(', ') }}
 {% endfor %}
 {% endif %}
 

dat-sdk/src/main/java/ai/dat/boot/BuildStateManager.java

@@ -23,7 +23,7 @@
  * @Date 2025/7/17
  */
 @Slf4j
-public class BuildStateManager {
+class BuildStateManager {
 
     private static final String STATE_FILE_PREFIX = "build_state_";
     private static final String STATE_FILE_SUFFIX = ".json";

dat-sdk/src/main/java/ai/dat/boot/ChangeKnowledgeCacheUtil.java

@@ -0,0 +1,54 @@
+package ai.dat.boot;
+
+import lombok.NonNull;
+
+import java.util.Collections;
+import java.util.HashMap;
+import java.util.List;
+import java.util.Map;
+
+/**
+ * 变化的业务知识缓存工具
+ *
+ * @Author JunjieM
+ * @Date 2025/9/26
+ */
+class ChangeKnowledgeCacheUtil {
+
+    private ChangeKnowledgeCacheUtil() {
+    }
+
+    // Map<projectId, Map<yamlFileRelativePath, List<String>>>
+    private final static Map<String, Map<String, List<String>>> CACHE = new HashMap<>();
+
+    public static void add(@NonNull String projectId,
+                           @NonNull String relativePath,
+                           @NonNull List<String> knowledge) {
+        Map<String, List<String>> listMap = new HashMap<>();
+        if (CACHE.containsKey(projectId)) {
+            listMap = CACHE.get(projectId);
+        }
+        listMap.put(relativePath, knowledge);
+        CACHE.put(projectId, listMap);
+    }
+
+    public static List<String> get(@NonNull String projectId,
+                                   @NonNull String relativePath) {
+        Map<String, List<String>> listMap = new HashMap<>();
+        if (CACHE.containsKey(projectId)) {
+            listMap = CACHE.get(projectId);
+        }
+        return listMap.get(relativePath);
+    }
+
+    public static Map<String, List<String>> get(String projectId) {
+        if (CACHE.containsKey(projectId)) {
+            return CACHE.get(projectId);
+        }
+        return Collections.emptyMap();
+    }
+
+    public static Map<String, List<String>> remove(String projectId) {
+        return CACHE.remove(projectId);
+    }
+}

dat-sdk/src/main/java/ai/dat/boot/ChangeQuestionSqlPairsCacheUtil.java

@@ -0,0 +1,55 @@
+package ai.dat.boot;
+
+import ai.dat.core.contentstore.data.QuestionSqlPair;
+import lombok.NonNull;
+
+import java.util.Collections;
+import java.util.HashMap;
+import java.util.List;
+import java.util.Map;
+
+/**
+ * 变化的问答SQL对缓存工具
+ *
+ * @Author JunjieM
+ * @Date 2025/9/26
+ */
+class ChangeQuestionSqlPairsCacheUtil {
+
+    private ChangeQuestionSqlPairsCacheUtil() {
+    }
+
+    // Map<projectId, Map<yamlFileRelativePath, List<QuestionSqlPair>>>
+    private final static Map<String, Map<String, List<QuestionSqlPair>>> CACHE = new HashMap<>();
+
+    public static void add(@NonNull String projectId,
+                           @NonNull String relativePath,
+                           @NonNull List<QuestionSqlPair> questionSqlPairs) {
+        Map<String, List<QuestionSqlPair>> listMap = new HashMap<>();
+        if (CACHE.containsKey(projectId)) {
+            listMap = CACHE.get(projectId);
+        }
+        listMap.put(relativePath, questionSqlPairs);
+        CACHE.put(projectId, listMap);
+    }
+
+    public static List<QuestionSqlPair> get(@NonNull String projectId,
+                                            @NonNull String relativePath) {
+        Map<String, List<QuestionSqlPair>> listMap = new HashMap<>();
+        if (CACHE.containsKey(projectId)) {
+            listMap = CACHE.get(projectId);
+        }
+        return listMap.get(relativePath);
+    }
+
+    public static Map<String, List<QuestionSqlPair>> get(String projectId) {
+        if (CACHE.containsKey(projectId)) {
+            return CACHE.get(projectId);
+        }
+        return Collections.emptyMap();
+    }
+
+    public static Map<String, List<QuestionSqlPair>> remove(String projectId) {
+        return CACHE.remove(projectId);
+    }
+}

dat-sdk/src/main/java/ai/dat/boot/ChangeSemanticModelsCacheUtil.java

@@ -19,6 +19,7 @@ class ChangeSemanticModelsCacheUtil {
     private ChangeSemanticModelsCacheUtil() {
     }
 
+    // Map<projectId, Map<yamlFileRelativePath, List<SemanticModel>>>
     private final static Map<String, Map<String, List<SemanticModel>>> CACHE = new HashMap<>();
 
     public static void add(@NonNull String projectId,

dat-sdk/src/main/java/ai/dat/boot/ChangeWordSynonymPairsCacheUtil.java

@@ -0,0 +1,55 @@
+package ai.dat.boot;
+
+import ai.dat.core.contentstore.data.WordSynonymPair;
+import lombok.NonNull;
+
+import java.util.Collections;
+import java.util.HashMap;
+import java.util.List;
+import java.util.Map;
+
+/**
+ * 变化的同义词对缓存工具
+ *
+ * @Author JunjieM
+ * @Date 2025/9/26
+ */
+class ChangeWordSynonymPairsCacheUtil {
+
+    private ChangeWordSynonymPairsCacheUtil() {
+    }
+
+    // Map<projectId, Map<yamlFileRelativePath, List<WordSynonymPair>>>
+    private final static Map<String, Map<String, List<WordSynonymPair>>> CACHE = new HashMap<>();
+
+    public static void add(@NonNull String projectId,
+                           @NonNull String relativePath,
+                           @NonNull List<WordSynonymPair> wordSynonymPairs) {
+        Map<String, List<WordSynonymPair>> listMap = new HashMap<>();
+        if (CACHE.containsKey(projectId)) {
+            listMap = CACHE.get(projectId);
+        }
+        listMap.put(relativePath, wordSynonymPairs);
+        CACHE.put(projectId, listMap);
+    }
+
+    public static List<WordSynonymPair> get(@NonNull String projectId,
+                                            @NonNull String relativePath) {
+        Map<String, List<WordSynonymPair>> listMap = new HashMap<>();
+        if (CACHE.containsKey(projectId)) {
+            listMap = CACHE.get(projectId);
+        }
+        return listMap.get(relativePath);
+    }
+
+    public static Map<String, List<WordSynonymPair>> get(String projectId) {
+        if (CACHE.containsKey(projectId)) {
+            return CACHE.get(projectId);
+        }
+        return Collections.emptyMap();
+    }
+
+    public static Map<String, List<WordSynonymPair>> remove(String projectId) {
+        return CACHE.remove(projectId);
+    }
+}