pro-craft 0.1.10__py3-none-any.whl → 0.1.12__py3-none-any.whl

pro_craft/prompt_helper.py

@@ -489,20 +489,25 @@ class Intel():
 
     def intellect_remove_format(self,
                     input_data: dict | str,
-                    OutputFormat: object,
                     prompt_id: str,
+                    OutputFormat: object = None,
+                    ExtraFormats: list[object] = [],
                     version: str = None,
                     inference_save_case = True,
                     ):
-                
-        base_format_prompt = """
+        
+        if OutputFormat:
+            base_format_prompt = """
 按照一定格式输出, 以便可以通过如下校验
 
 使用以下正则检出
 "```json([\s\S]*?)```"
 使用以下方式验证
 """
-        output_format = base_format_prompt + inspect.getsource(OutputFormat)
+            output_format = base_format_prompt + "\n".join([inspect.getsource(outputformat) for outputformat in ExtraFormats]) + inspect.getsource(OutputFormat)
+
+        else:
+            output_format = ""
 
         ai_result = self.intellect_remove(
             input_data=input_data,
@@ -512,21 +517,26 @@ class Intel():
             inference_save_case=inference_save_case
         )
 
-        try:
-            ai_result = json.loads(extract_(ai_result,r'json'))
-            OutputFormat(**ai_result)
-        except ValidationError as e:
-            log_ = "记忆卡片合并 - 大模型生成的格式未通过校验"
-            logger.error(log_)
-            logger.error(f"错误类型: {type(e)}")
-            logger.error(f"错误信息: {e}")
-            logger.error(f"错误详情 (errors()): {e.errors()}")
-            logger.error(f"错误详情 (json()): {e.json(indent=2)}")
-            raise ValidationError(log_)
-            
+        if OutputFormat:
+            try:
+                ai_result = json.loads(extract_(ai_result,r'json'))
+                OutputFormat(**ai_result)
+            except ValidationError as e:
+                log_ = "记忆卡片合并 - 大模型生成的格式未通过校验"
+                logger.error(log_)
+                logger.error(f"错误类型: {type(e)}")
+                logger.error(f"错误信息: {e}")
+                logger.error(f"错误详情 (errors()): {e.errors()}")
+                logger.error(f"错误详情 (json()): {e.json(indent=2)}")
+                raise ValidationError(log_)
+        else:
+            try:
+                assert isinstance(ai_result,str)
+            except AssertionError as e:
+                logger.error("错误了问题")
+       
         return ai_result
     
-
     def intellect_remove_warp(self,prompt_id: str):
         def outer_packing(func):
             @functools.wraps(func)

pro_craft/prompt_helper_async.py

@@ -68,9 +68,8 @@ class AsyncIntel():
                 ):
         database_url = database_url or os.getenv("database_url")
         assert database_url
-    
-        # async_engine
-        self.async_engine = create_async_engine(database_url, echo=True,
+
+        self.engine = create_async_engine(database_url, echo=True,
                                     pool_size=10,        # 连接池中保持的连接数
                                     max_overflow=20,     # 当pool_size不够时，允许临时创建的额外连接数
                                     pool_recycle=3600,   # 每小时回收一次连接
@@ -87,10 +86,10 @@ class AsyncIntel():
             self.llm = BianXieAdapter()
             
     async def create_database(self):
-        async with self.async_engine.begin() as conn:
+        async with self.engine.begin() as conn:
             await conn.run_sync(PromptBase.metadata.create_all)
             
-    async def _aget_latest_prompt_version(self,target_prompt_id,session):
+    async def _get_latest_prompt_version(self,target_prompt_id,session):
         """
         获取指定 prompt_id 的最新版本数据，通过创建时间判断。
         """
@@ -113,7 +112,7 @@ class AsyncIntel():
             editing_log(f"未找到 prompt_id '{target_prompt_id}' 的任何版本。")
         return result
 
-    async def _aget_specific_prompt_version(self,target_prompt_id, target_version,session):
+    async def _get_specific_prompt_version(self,target_prompt_id, target_version,session):
         """
         获取指定 prompt_id 和特定版本的数据。
 
@@ -141,7 +140,7 @@ class AsyncIntel():
             editing_log(f"未找到 prompt_id '{target_prompt_id}', 版本 '{target_version}' 的提示词数据。")
         return specific_prompt
 
-    async def aget_prompts_from_sql(self,
+    async def get_prompts_from_sql(self,
                              prompt_id: str,
                              version = None,
                              session = None) -> Prompt:
@@ -150,15 +149,15 @@ class AsyncIntel():
         """
         # 查看是否已经存在
         if version:
-            prompts_obj = await self._aget_specific_prompt_version(prompt_id,version,session=session)
+            prompts_obj = await self._get_specific_prompt_version(prompt_id,version,session=session)
             if not prompts_obj:
-                prompts_obj = await self._aget_latest_prompt_version(prompt_id,session = session)
+                prompts_obj = await self._get_latest_prompt_version(prompt_id,session = session)
         else:
-            prompts_obj = await self._aget_latest_prompt_version(prompt_id,session = session)     
+            prompts_obj = await self._get_latest_prompt_version(prompt_id,session = session)     
         return prompts_obj
         
             
-    async def asave_prompt_increment_version(self,
+    async def save_prompt_increment_version(self,
                            prompt_id: str,
                            new_prompt: str,
                            use_case:str = "",
@@ -171,7 +170,7 @@ class AsyncIntel():
         input_data 指的是输入用例, 可以为空
         """
         # 查看是否已经存在
-        prompts_obj = await self.aget_prompts_from_sql(prompt_id=prompt_id,session=session)
+        prompts_obj = await self.get_prompts_from_sql(prompt_id=prompt_id,session=session)
 
         if prompts_obj:
             # 如果存在版本加1
@@ -198,7 +197,7 @@ class AsyncIntel():
         session.add(prompt1)
         await session.commit() # 提交事务，将数据写入数据库
 
-    async def asave_use_case_by_sql(self,
+    async def save_use_case_by_sql(self,
                              prompt_id: str,
                              use_case:str = "",
                              output = "",
@@ -237,7 +236,7 @@ class AsyncIntel():
 
         s_prompt = extract_(system_result,pattern_key=r"prompt")
         chat_history = s_prompt or system_result
-        await self.asave_prompt_increment_version(prompt_id,
+        await self.save_prompt_increment_version(prompt_id,
                                 new_prompt = chat_history,
                                 input_data = " summary ",
                                 session = session)
@@ -279,20 +278,20 @@ class AsyncIntel():
 {opinion}
 """
 
-        prompt, _ = await self.aget_prompts_from_sql(prompt_id = prompt_id,version = version)
+        prompt, _ = await self.get_prompts_from_sql(prompt_id = prompt_id,version = version)
         if demand:
             new_prompt = await self.llm.aproduct(
                 change_by_opinion_prompt.format(old_system_prompt=prompt, opinion=demand)
             )
         else:
             new_prompt = prompt
-        await self.asave_prompt_increment_version(prompt_id = prompt_id,
+        await self.save_prompt_increment_version(prompt_id = prompt_id,
                             new_prompt = new_prompt,
                             input_data = " finetune ",
                             session = session)
 
 
-    async def apush_action_order(self,demand : str,prompt_id: str,
+    async def push_action_order(self,demand : str,prompt_id: str,
                          action_type = 'train'):
 
         """
@@ -302,11 +301,11 @@ class AsyncIntel():
         将打算修改的状态推上数据库 # 1
         """
         # 查看是否已经存在
-        async with create_async_session(self.async_engine) as session:
+        async with create_async_session(self.engine) as session:
 
-            latest_prompt = await self.aget_prompts_from_sql(prompt_id=prompt_id,session=session)
+            latest_prompt = await self.get_prompts_from_sql(prompt_id=prompt_id,session=session)
 
-            await self.asave_prompt_increment_version(prompt_id=latest_prompt.prompt_id,
+            await self.save_prompt_increment_version(prompt_id=latest_prompt.prompt_id,
                                     new_prompt = latest_prompt.prompt,
                                     use_case = latest_prompt.use_case,
                                     action_type=action_type,
@@ -319,7 +318,7 @@ class AsyncIntel():
 
 
 
-    async def aintellect_remove(self,
+    async def intellect_remove(self,
                     input_data: dict | str,
                     output_format: str,
                     prompt_id: str,
@@ -333,18 +332,18 @@ class AsyncIntel():
             input_ = input_data
         
         # 查数据库, 获取最新提示词对象
-        async with create_async_session(self.async_engine) as session:
-            result_obj = await self.aget_prompts_from_sql(prompt_id=prompt_id,session=session)
+        async with create_async_session(self.engine) as session:
+            result_obj = await self.get_prompts_from_sql(prompt_id=prompt_id,session=session)
 
 
             if result_obj is None:
-                await self.asave_prompt_increment_version(
+                await self.save_prompt_increment_version(
                     prompt_id = prompt_id,
                     new_prompt = "做一些处理",
                     use_case = input_,
                     session = session
                 )
-                ai_result = await self.aintellect_remove(input_data = input_data,
+                ai_result = await self.intellect_remove(input_data = input_data,
                                        output_format = output_format,
                                        prompt_id = prompt_id,
                                        version = version,
@@ -357,7 +356,7 @@ class AsyncIntel():
                 # 直接推理即可
                 ai_result = await self.llm.aproduct(prompt + output_format + "\n-----input----\n" +  input_)
                 if inference_save_case:
-                    await self.asave_use_case_by_sql(prompt_id,
+                    await self.save_use_case_by_sql(prompt_id,
                                         use_case = input_,
                                         output = ai_result,
                                         solution = "备注/理想回复",
@@ -386,7 +385,7 @@ class AsyncIntel():
             
                 ai_result = await self.llm.aproduct(input_prompt)
                 chat_history = input_prompt + "\nassistant:\n" + ai_result # 用聊天记录作为完整提示词
-                await self.asave_prompt_increment_version(prompt_id, chat_history,
+                await self.save_prompt_increment_version(prompt_id, chat_history,
                                         use_case = input_,
                                         session = session)
     
@@ -426,7 +425,7 @@ class AsyncIntel():
 
         return ai_result
     
-    async def aintellect_stream_remove(self,
+    async def intellect_stream_remove(self,
                     input_data: dict | str,
                     output_format: str,
                     prompt_id: str,
@@ -441,18 +440,18 @@ class AsyncIntel():
 
         
         # 查数据库, 获取最新提示词对象
-        with create_session(self.async_engine) as session:
-            result_obj = await self.aget_prompts_from_sql(prompt_id=prompt_id,session=session)
+        with create_session(self.engine) as session:
+            result_obj = await self.get_prompts_from_sql(prompt_id=prompt_id,session=session)
 
             '''
                         if result_obj is None:
-                            await self.asave_prompt_increment_version(
+                            await self.save_prompt_increment_version(
                                 prompt_id = prompt_id,
                                 new_prompt = "做一些处理",
                                 use_case = input_,
                                 session = session
                             )
-                            ai_result = await self.aintellect_stream_remove(input_data = input_data,
+                            ai_result = await self.intellect_stream_remove(input_data = input_data,
                                                 output_format = output_format,
                                                 prompt_id = prompt_id,
                                                 version = version,
@@ -470,7 +469,7 @@ class AsyncIntel():
                     ai_result += word
                     yield word
                 if inference_save_case:
-                    await self.asave_use_case_by_sql(prompt_id,
+                    await self.save_use_case_by_sql(prompt_id,
                                         use_case = input_,
                                         output = ai_result,
                                         solution = "备注/理想回复",
@@ -504,7 +503,7 @@ class AsyncIntel():
                     yield word
 
                 chat_history = input_prompt + "\nassistant:\n" + ai_result # 用聊天记录作为完整提示词
-                await self.asave_prompt_increment_version(prompt_id, chat_history,
+                await self.save_prompt_increment_version(prompt_id, chat_history,
                                         use_case = input_,
                                         session = session)
     
@@ -555,7 +554,7 @@ class AsyncIntel():
             else:
                 raise
 
-    async def aintellect_remove_format(self,
+    async def intellect_remove_format(self,
                     input_data: dict | str,
                     OutputFormat: object,
                     prompt_id: str,
@@ -573,7 +572,7 @@ class AsyncIntel():
 """
         output_format = base_format_prompt + "\n".join([inspect.getsource(outputformat) for outputformat in ExtraFormats]) + inspect.getsource(OutputFormat)
 
-        ai_result = await self.aintellect_remove(
+        ai_result = await self.intellect_remove(
                     input_data=input_data,
                     output_format=output_format,
                     prompt_id=prompt_id,
@@ -606,7 +605,7 @@ class AsyncIntel():
         return ai_result
     
     
-    def aintellect_remove_warp(self,prompt_id: str):
+    def intellect_remove_warp(self,prompt_id: str):
         def outer_packing(func):
             @functools.wraps(func)
             async def wrapper(*args, **kwargs):
@@ -621,7 +620,7 @@ class AsyncIntel():
                 elif isinstance(input_data,str):
                     input_ = output_ = input_data
 
-                output_ = await self.aintellect_remove_format(
+                output_ = await self.intellect_remove_format(
                         input_data = input_data,
                         prompt_id = prompt_id,
                         OutputFormat = OutputFormat,

pro_craft/prompt_helper_new.py

@@ -0,0 +1,529 @@
+# 测试1
+
+from pro_craft.utils import extract_
+from pro_craft.log import Log
+from pro_craft.database import Prompt, UseCase, PromptBase
+from pro_craft.utils import create_session, create_async_session
+from llmada.core import BianXieAdapter, ArkAdapter
+from datetime import datetime
+from enum import Enum
+import functools
+import json
+import os
+from sqlalchemy import create_engine
+from pro_craft.database import SyncMetadata
+import inspect
+from pydantic import BaseModel, Field, ValidationError, field_validator
+from datetime import datetime, timedelta
+logger = Log.logger
+editing_log = logger.debug
+
+BATCH_SIZE = 1000
+
+def get_last_sync_time(target_session) -> datetime:
+    """从目标数据库获取上次同步时间"""
+    metadata_entry = target_session.query(SyncMetadata).filter_by(table_name="sync_metadata").first()
+    if metadata_entry:
+        return metadata_entry.last_sync_time
+    return datetime(1970, 1, 1) # 默认一个很早的时间
+
+def update_last_sync_time(target_session, new_sync_time: datetime):
+    """更新目标数据库的上次同步时间"""
+    metadata_entry = target_session.query(SyncMetadata).filter_by(table_name="sync_metadata").first()
+    if metadata_entry:
+        metadata_entry.last_sync_time = new_sync_time
+    else:
+        # 如果不存在，则创建
+        new_metadata = SyncMetadata(table_name="sync_metadata", last_sync_time=new_sync_time)
+        target_session.add(new_metadata)
+    target_session.commit()
+    print(f"Updated last sync time to: {new_sync_time}")
+
+
+
+class IntellectType(Enum):
+    train = "train"
+    inference = "inference"
+    summary = "summary"
+
+class Intel():
+    def __init__(self,
+                 database_url = "",
+                 model_name = "",
+                ):
+        database_url = database_url or os.getenv("database_url")
+        assert database_url
+        self.engine = create_engine(database_url, echo=False, # echo=True 仍然会打印所有执行的 SQL 语句
+                                    pool_size=10,        # 连接池中保持的连接数
+                                    max_overflow=20,     # 当pool_size不够时，允许临时创建的额外连接数
+                                    pool_recycle=3600,   # 每小时回收一次连接
+                                    pool_pre_ping=True,  # 使用前检查连接活性
+                                    pool_timeout=30      # 等待连接池中连接的最长时间（秒）
+                                    ) 
+
+        PromptBase.metadata.create_all(self.engine)
+        
+        if model_name in ["gemini-2.5-flash-preview-05-20-nothinking",]:
+            self.llm = BianXieAdapter(model_name = model_name)
+        elif model_name in ["doubao-1-5-pro-256k-250115",]:
+            self.llm = ArkAdapter(model_name = model_name)
+        else:
+            print('Use BianXieAdapter')
+            self.llm = BianXieAdapter()
+
+        
+    def _get_latest_prompt_version(self,target_prompt_id,session):
+        """
+        获取指定 prompt_id 的最新版本数据，通过创建时间判断。
+        """
+        
+        result = session.query(Prompt).filter(
+            Prompt.prompt_id == target_prompt_id
+        ).order_by(
+            Prompt.timestamp.desc(),
+            Prompt.version.desc()
+        ).first()
+
+        if result:
+            editing_log(f"找到 prompt_id '{target_prompt_id}' 的最新版本 (基于时间): {result.version}")
+        else:
+            editing_log(f"未找到 prompt_id '{target_prompt_id}' 的任何版本。")
+        return result
+
+    def _get_specific_prompt_version(self,target_prompt_id, target_version,session):
+        """
+        获取指定 prompt_id 和特定版本的数据。
+
+        Args:
+            target_prompt_id (str): 目标提示词的唯一标识符。
+            target_version (int): 目标提示词的版本号。
+            table_name (str): 存储提示词数据的数据库表名。
+            db_manager (DBManager): 数据库管理器的实例，用于执行查询。
+
+        Returns:
+            dict or None: 如果找到，返回包含 id, prompt_id, version, timestamp, prompt 字段的字典；
+                        否则返回 None。
+        """
+
+        result = session.query(Prompt).filter(
+            Prompt.prompt_id == target_prompt_id,
+            Prompt.version == target_version
+        ).first() # 因为 (prompt_id, version) 是唯一的，所以 first() 足够
+        if result:
+            editing_log(f"找到 prompt_id '{target_prompt_id}', 版本 '{target_version}' 的提示词数据。")
+        else:
+            editing_log(f"未找到 prompt_id '{target_prompt_id}', 版本 '{target_version}' 的提示词数据。")
+        return result
+    
+                
+    def sync_prompt_data_to_database(self,database_url:str):
+        target_engine = create_engine(database_url, echo=False)
+        # PromptBase.metadata.create_all(target_engine,tables=[SyncMetadata.__table__]) 
+        PromptBase.metadata.create_all(target_engine) 
+
+        with create_session(self.engine) as source_session:
+            with create_session(target_engine) as target_session:
+                last_sync_time = get_last_sync_time(target_session)
+                print(f"Starting sync for sync_metadata from: {last_sync_time}")
+
+
+                processed_count = 0
+                current_batch_max_updated_at = last_sync_time
+
+                while True:
+                    records_to_sync = source_session.query(Prompt)\
+                                .filter(Prompt.timestamp > last_sync_time)\
+                                .order_by(Prompt.timestamp.asc(), Prompt.id.asc())\
+                                .limit(BATCH_SIZE)\
+                                .all()
+                    if not records_to_sync:
+                        break # 没有更多记录了
+
+                    # 准备要插入或更新到目标数据库的数据
+                    for record in records_to_sync:
+                        # 查找目标数据库中是否存在该ID的记录
+                        # 这里的 `User` 模型会对应到 target_db.users
+                        target_prompt = target_session.query(Prompt).filter_by(id=record.id).first()
+
+                        if target_prompt:
+                            # 如果存在，则更新
+                            target_prompt.prompt_id = record.prompt_id
+                            target_prompt.version = record.version
+                            target_prompt.timestamp = record.timestamp
+                            target_prompt.prompt = record.prompt
+                            target_prompt.use_case = record.use_case
+                            target_prompt.action_type = record.action_type
+                            target_prompt.demand = record.demand
+                            target_prompt.score = record.score
+                            target_prompt.is_deleted = record.is_deleted
+                        else:
+                            # 如果不存在，则添加新记录
+                            # 注意：这里需要创建一个新的User实例，而不是直接添加源数据库的record对象
+                            new_user = Prompt(
+                                prompt_id=record.prompt_id, 
+                                version=record.version,
+                                timestamp=record.timestamp,
+                                prompt = record.prompt,
+                                use_case = record.use_case,
+                                action_type = record.action_type,
+                                demand = record.demand,
+                                score = record.score,
+                                is_deleted = record.is_deleted
+                                )
+                            target_session.add(new_user)
+                        
+                        # 记录当前批次最大的 updated_at
+                        if record.timestamp > current_batch_max_updated_at:
+                            current_batch_max_updated_at = record.timestamp
+
+                    target_session.commit() # 提交当前批次的变更
+                    processed_count += len(records_to_sync)
+                    print(f"Processed {len(records_to_sync)} records. Total processed: {processed_count}")
+
+                    # 更新 last_sync_time，为确保下次查询从正确的时间点开始，
+                    # 可以使用当前批次中最大的 updated_at + 1微秒，以避免重复查询同一时间戳的记录。
+                    # 更严谨的做法是，如果 updated_at 相同，则使用 ID 进行辅助排序，并记录 ID。
+                    # 为了简化，这里直接使用 updated_at
+                    last_sync_time = current_batch_max_updated_at + timedelta(microseconds=1) 
+                    
+                    if len(records_to_sync) < BATCH_SIZE: # 如果查询到的记录数小于批次大小，说明已经处理完所有符合条件的记录
+                        break
+
+                if processed_count > 0:
+                    # 最终更新last_sync_time到数据库，确保记录的是所有已处理记录中最新的一个
+                    update_last_sync_time(target_session, current_batch_max_updated_at + timedelta(microseconds=1))
+                else:
+                    print("No new records to sync.")
+
+
+    def get_prompts_from_sql(self,
+                             prompt_id: str,
+                             version = None,
+                             session = None) -> Prompt:
+        """
+        从sql获取提示词
+        """
+        # 查看是否已经存在
+        if version:
+            prompts_obj = self._get_specific_prompt_version(prompt_id,version,session=session)
+            if not prompts_obj:
+                prompts_obj = self._get_latest_prompt_version(prompt_id,session = session)
+        else:
+            prompts_obj = self._get_latest_prompt_version(prompt_id,session = session)     
+        return prompts_obj
+            
+
+    def save_prompt_increment_version(self,
+                           prompt_id: str,
+                           new_prompt: str,
+                           use_case:str = "",
+                           action_type = "inference",
+                           demand = "",
+                           score = 60,
+                           session = None):
+        """
+        从sql保存提示词
+        input_data 指的是输入用例, 可以为空
+        """
+        # 查看是否已经存在
+        prompts_obj = self.get_prompts_from_sql(prompt_id=prompt_id,session=session)
+
+        if prompts_obj:
+            # 如果存在版本加1
+            version_ori = prompts_obj.version
+            _, version = version_ori.split(".")
+            version = int(version)
+            version += 1
+            version_ = f"1.{version}"
+
+        else:
+            # 如果不存在版本为1.0
+            version_ = '1.0'
+        
+        prompt1 = Prompt(prompt_id=prompt_id, 
+                        version=version_,
+                        timestamp=datetime.now(),
+                        prompt = new_prompt,
+                        use_case = use_case,
+                        action_type = action_type,
+                        demand = demand,
+                        score = score
+                        )
+
+        session.add(prompt1)
+        session.commit() # 提交事务，将数据写入数据库
+
+    def save_use_case_by_sql(self,
+                             prompt_id: str,
+                             use_case:str = "",
+                             output = "",
+                             solution: str = "",
+                             session = None
+                            ):
+        """
+        从sql保存提示词
+        """
+        use_case = UseCase(prompt_id=prompt_id, 
+                        use_case = use_case,
+                        output = output,
+                        solution = solution,
+                        )
+
+        session.add(use_case)
+        session.commit() # 提交事务，将数据写入数据库
+
+    def summary_to_sql(
+            self,
+            prompt_id:str,
+            version = None,
+            prompt = "",
+            session = None
+        ):
+        """
+        让大模型微调已经存在的 system_prompt
+        """
+        system_prompt_created_prompt = """        
+很棒, 我们已经达成了某种默契, 我们之间合作无间, 但是, 可悲的是, 当我关闭这个窗口的时候, 你就会忘记我们之间经历的种种磨合, 这是可惜且心痛的, 所以你能否将目前这一套处理流程结晶成一个优质的prompt 这样, 我们下一次只要将prompt输入, 你就能想起我们今天的磨合过程,
+对了,我提示一点, 这个prompt的主角是你, 也就是说, 你在和未来的你对话, 你要教会未来的你今天这件事, 是否让我看懂到时其次
+
+只要输出提示词内容即可, 不需要任何的说明和解释
+"""
+        system_result = self.llm.product(prompt + system_prompt_created_prompt)
+
+        s_prompt = extract_(system_result,pattern_key=r"prompt")
+        chat_history = s_prompt or system_result
+        self.save_prompt_increment_version(prompt_id,
+                                new_prompt = chat_history,
+                                use_case = " summary ",
+                                session = session)
+        
+    def prompt_finetune_to_sql(
+            self,
+            prompt_id:str,
+            version = None,
+            demand: str = "",
+            session = None,
+        ):
+        """
+        让大模型微调已经存在的 system_prompt
+        """
+        change_by_opinion_prompt = """
+你是一个资深AI提示词工程师，具备卓越的Prompt设计与优化能力。
+我将为你提供一段现有System Prompt。你的核心任务是基于这段Prompt进行修改，以实现我提出的特定目标和功能需求。
+请你绝对严格地遵循以下原则：
+ 极端最小化修改原则（核心）：
+ 在满足所有功能需求的前提下，只进行我明确要求的修改。
+ 即使你认为有更“优化”、“清晰”或“简洁”的表达方式，只要我没有明确要求，也绝不允许进行任何未经指令的修改。
+ 目的就是尽可能地保留原有Prompt的字符和结构不变，除非我的功能要求必须改变。
+ 例如，如果我只要求你修改一个词，你就不应该修改整句话的结构。
+ 严格遵循我的指令：
+ 你必须精确地执行我提出的所有具体任务和要求。
+ 绝不允许自行添加任何超出指令范围的说明、角色扮演、约束条件或任何非我指令要求的内容。
+ 保持原有Prompt的风格和语调：
+ 尽可能地与现有Prompt的语言风格、正式程度和语调保持一致。
+ 不要改变不相关的句子或其表达方式。
+ 只提供修改后的Prompt：
+ 直接输出修改后的完整System Prompt文本。
+ 不要包含任何解释、说明或额外对话。
+ 在你开始之前，请务必确认你已理解并能绝对严格地遵守这些原则。任何未经明确指令的改动都将视为未能完成任务。
+
+现有System Prompt:
+{old_system_prompt}
+
+功能需求:
+{opinion}
+"""
+        prompts_obj = self.get_prompts_from_sql(prompt_id = prompt_id,version = version,session = session)
+
+        if demand:
+            new_prompt = self.llm.product(
+                change_by_opinion_prompt.format(old_system_prompt=prompts_obj.prompt, opinion=demand)
+            )
+        else:
+            new_prompt = prompts_obj.prompt
+        self.save_prompt_increment_version(prompt_id = prompt_id,
+                            new_prompt = new_prompt,
+                            use_case = " finetune ",
+                            session = session)
+
+    def push_action_order(self,
+                          prompt_id: str,
+                          demand : str,
+                          action_type = 'train'):
+
+        """
+        从sql保存提示词
+        推一个train 状态到指定的位置
+
+        将打算修改的状态推上数据库 # 1
+        """
+        # 查看是否已经存在
+        with create_session(self.engine) as session:
+            latest_prompt = self.get_prompts_from_sql(prompt_id=prompt_id,session=session)
+
+            self.save_prompt_increment_version(prompt_id=latest_prompt.prompt_id,
+                                    new_prompt = latest_prompt.prompt,
+                                    use_case = latest_prompt.use_case,
+                                    action_type=action_type,
+                                    demand=demand,
+                                    score=latest_prompt.score,
+                                    session=session
+                                    )
+            
+            return "success"
+
+
+    def intellect_remove(self,
+                    input_data: dict | str,
+                    output_format: str,
+                    prompt_id: str,
+                    version: str = None,
+                    inference_save_case = True,
+                    push_patch = False,
+                    ):
+        """
+        # output_format 单独使用一个大模型处理
+        使用指南:
+            1 训练, 使用单一例子做大量的沟通来奠定基础
+            2 总结, 将沟通好的总结成完整提示词
+            3 推理, 使用部署
+            4 微调, 针对一些格式性的, 问题进行微调
+            5 补丁, 微调无法解决的问题, 可以尝试使用补丁来解决
+        """
+        if isinstance(input_data,dict):
+            input_ = json.dumps(input_data,ensure_ascii=False)
+        elif isinstance(input_data,str):
+            input_ = input_data
+
+        
+        # 查数据库, 获取最新提示词对象
+        with create_session(self.engine) as session:
+            result_obj = self.get_prompts_from_sql(prompt_id=prompt_id,session=session)
+
+            prompt = result_obj.prompt
+            if result_obj.action_type == "inference":
+                # 直接推理即可
+                ai_result = self.llm.product(prompt + "\n-----input----\n" +  input_)
+                if inference_save_case:
+                    self.save_use_case_by_sql(prompt_id,
+                                        use_case = input_,
+                                        output = ai_result,
+                                        solution = "备注/理想回复",
+                                        session = session,
+                                        )
+                    
+            elif result_obj.action_type == "train":
+                assert result_obj.demand # 如果type = train 且 demand 是空 则报错
+                # 则训练推广
+
+                # 新版本 默人修改会 inference 状态
+                chat_history = prompt
+                before_input = result_obj.use_case
+                demand = result_obj.demand
+            
+
+                assert demand
+                # 注意, 这里的调整要求使用最初的那个输入, 最好一口气调整好
+                chat_history = prompt
+                if input_ == before_input: # 输入没变, 说明还是针对同一个输入进行讨论
+                    # input_prompt = chat_history + "\nuser:" + demand
+                    input_prompt = chat_history + "\nuser:" + demand 
+                else:
+                    # input_prompt = chat_history + "\nuser:" + demand + "\n-----input----\n" + input_
+                    input_prompt = chat_history + "\nuser:" + demand  + "\n-----input----\n" + input_
+            
+                ai_result = self.llm.product(input_prompt)
+                chat_history = input_prompt + "\nassistant:\n" + ai_result # 用聊天记录作为完整提示词
+                self.save_prompt_increment_version(prompt_id, chat_history,
+                                        use_case = input_,
+                                        score = 60,
+                                        session = session)
+    
+            elif result_obj.action_type == "summary":
+                self.summary_to_sql(prompt_id = prompt_id,
+                            prompt = prompt,
+                            session = session
+                            )
+                ai_result = self.llm.product(prompt + "\n-----input----\n" +  input_)
+
+            elif result_obj.action_type == "finetune":
+                demand = result_obj.demand
+            
+                assert demand
+                self.prompt_finetune_to_sql(prompt_id = prompt_id,
+                                            demand = demand,
+                                            session = session
+                                            )
+                ai_result = self.llm.product(prompt + "\n-----input----\n" +  input_)
+            elif result_obj.action_type == "patch":
+                
+                demand = result_obj.demand
+                assert demand
+
+                chat_history = prompt + demand
+                ai_result = self.llm.product(chat_history + "\n-----input----\n" +  input_)
+                if push_patch:
+                    self.save_prompt_increment_version(prompt_id, chat_history,
+                                            use_case = input_,
+                                            score = 60,
+                                            session = session)
+            else:
+                raise
+
+            system_prompt = """
+对数据的数据,进行整理, 不改变其文本内容, 只将其整合成特定的形式
+"""
+            ai_result = self.llm.product(system_prompt + ai_result + output_format)
+
+        return ai_result
+    
+
+    def intellect_remove_format(self,
+                    input_data: dict | str,
+                    prompt_id: str,
+                    OutputFormat: object = None,
+                    ExtraFormats: list[object] = [],
+                    version: str = None,
+                    inference_save_case = True,
+                    ):
+        
+        if OutputFormat:
+            base_format_prompt = """
+按照一定格式输出, 以便可以通过如下校验
+
+使用以下正则检出
+"```json([\s\S]*?)```"
+使用以下方式验证
+"""
+            output_format = base_format_prompt + "\n".join([inspect.getsource(outputformat) for outputformat in ExtraFormats]) + inspect.getsource(OutputFormat)
+
+        else:
+            output_format = ""
+
+        ai_result = self.intellect_remove(
+            input_data=input_data,
+            output_format=output_format,
+            prompt_id=prompt_id,
+            version=version,
+            inference_save_case=inference_save_case
+        )
+
+        if OutputFormat:
+            try:
+                ai_result = json.loads(extract_(ai_result,r'json'))
+                OutputFormat(**ai_result)
+            except ValidationError as e:
+                log_ = "记忆卡片合并 - 大模型生成的格式未通过校验"
+                logger.error(log_)
+                logger.error(f"错误类型: {type(e)}")
+                logger.error(f"错误信息: {e}")
+                logger.error(f"错误详情 (errors()): {e.errors()}")
+                logger.error(f"错误详情 (json()): {e.json(indent=2)}")
+                raise ValidationError(log_)
+        else:
+            try:
+                assert isinstance(ai_result,str)
+            except AssertionError as e:
+                logger.error("错误了问题")
+       
+        return ai_result

{pro_craft-0.1.10.dist-info → pro_craft-0.1.12.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: pro-craft
-Version: 0.1.10
+Version: 0.1.12
 Summary: Add your description here
 Requires-Python: >=3.12
 Description-Content-Type: text/markdown

{pro_craft-0.1.10.dist-info → pro_craft-0.1.12.dist-info}/RECORD

@@ -4,8 +4,9 @@ pro_craft/designer.py,sha256=3gyCqrjcw61sHzDjUPKhL1LOAE8xWLLbNT8NlK2mFLc,4739
 pro_craft/evals.py,sha256=1T86jur4k3cLk43j1GyAW4JS0nPNfl6P0ZOQmu-SgpA,1928
 pro_craft/file_manager.py,sha256=2j7lCt9L4mtvAy8_76ibTthXLwKKmVatWIB3DSvQM7U,3805
 pro_craft/log.py,sha256=MZf9jCZsiRoAq8v4FxVnJqeSXxgzAiiKf7mxz6bFtwM,4263
-pro_craft/prompt_helper.py,sha256=1EDGP2sKoUCrLXiR0kD2AqzT9LcFYWO8S2FyaZlBoxc,26150
-pro_craft/prompt_helper_async.py,sha256=Ej0JXwev0FmPSoOE5iEXtZxQ3pXoqcaulyyUuC7WyC4,27859
+pro_craft/prompt_helper.py,sha256=GClPsomdGdo6FXlp8VdiQy8GaBtj3oLeq2oFw0RiM3k,26592
+pro_craft/prompt_helper_async.py,sha256=-EtwjDUL2Xoi4qjZpzgnUyAVBkt-WCocGgVl5PgA4Js,27771
+pro_craft/prompt_helper_new.py,sha256=DV871XElJmHDO2xis1mC1f3QFvq3ll81VGGD9aj0Eiw,23542
 pro_craft/server.py,sha256=fPAosQIU0d7gxICiALl8u6QwbLI4cawVFyoRYebRES0,2827
 pro_craft/utils.py,sha256=cpvwk68mD9hYY8WCq2JXzfrrXqpshiscz_OSav4tC7U,5687
 pro_craft/code_helper/coder.py,sha256=NXglF1KiPtGe4HZN0MZvFJ8p9Iyd5kzIt72DQGgRwXA,24715
@@ -17,7 +18,7 @@ pro_craft/server/mcp/math.py,sha256=OOzGXx64nK4bOVlu33PtVddcCQ9ilqA3Em9yxjSX9cg,
 pro_craft/server/mcp/resource.py,sha256=z94jP3qZofO-1lZCM3TuOfLajw41HARs1ojXab1ymas,776
 pro_craft/server/mcp/weather.py,sha256=RAGuf4sgjlTQSfRRZ1Fo18JnuMQRS_Db9p6AqBQrl8E,455
 pro_craft/server/router/recommended.py,sha256=IAZFdmb8HSl2_TOJeuv5uOKIX47XyX4p4sEwxG-0vt0,9968
-pro_craft-0.1.10.dist-info/METADATA,sha256=d0eKMauNY8YEx9A8ZAtIvGXEVYTER04UwYCscH7ifHY,1800
-pro_craft-0.1.10.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-pro_craft-0.1.10.dist-info/top_level.txt,sha256=yqYDHArnYMWpeCxkmGRwlL6sJtxiOUnYylLDx9EOgFg,10
-pro_craft-0.1.10.dist-info/RECORD,,
+pro_craft-0.1.12.dist-info/METADATA,sha256=VKEqBoEWYaOP0Ue2BSZ3iKZ1W_MYqsaRS7f3lQ-bscs,1800
+pro_craft-0.1.12.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+pro_craft-0.1.12.dist-info/top_level.txt,sha256=yqYDHArnYMWpeCxkmGRwlL6sJtxiOUnYylLDx9EOgFg,10
+pro_craft-0.1.12.dist-info/RECORD,,