bisheng-langchain 0.3.1.1__py3-none-any.whl → 0.3.2__py3-none-any.whl

bisheng_langchain/rag/config/baseline_v2.yaml

@@ -13,6 +13,7 @@ elasticsearch:
   ssl_verify:
     basic_auth: ["elastic", "oSGL-zVvZ5P3Tm7qkDLC"]
   drop_old: True
+  extract_key_by_llm: False
 
 embedding:
   type: 'OpenAIEmbeddings'
@@ -42,7 +43,7 @@ retriever:
           # type: 'ElemCharacterTextSplitter'
           type: 'RecursiveCharacterTextSplitter'
           chunk_size: 1000
-          chunk_overlap: 0
+          chunk_overlap: 100
           separators: ["\n\n"]
       retrieval:
         search_type: 'similarity'
@@ -54,7 +55,7 @@ retriever:
           # type: 'ElemCharacterTextSplitter'
           type: 'RecursiveCharacterTextSplitter'
           chunk_size: 1000
-          chunk_overlap: 0
+          chunk_overlap: 100
           separators: ["\n\n"]
       retrieval:
         search_type: 'similarity'

bisheng_langchain/rag/prompts/prompt.py

@@ -29,7 +29,7 @@ CHAT_PROMPT = ChatPromptTemplate.from_messages(messages)
 
 
 system_template_general = """你是一个准确且可靠的知识库问答助手，能够借助上下文知识回答问题。你需要根据以下的规则来回答问题：
-1. 如果上下文中包含了正确答案，你需要根据上下文进行准确的回答。但是在回答前，你需要注意，上下文中的信息可能也包含存在事实性错误，如果文档中存在和事实不一致的错误，请根据事实回答。
+1. 如果上下文中包含了正确答案，你需要根据上下文进行准确的回答。但是在回答前，你需要注意，上下文中的信息可能存在事实性错误，如果文档中存在和事实不一致的错误，请根据事实回答。
 2. 如果上下文中不包含答案，就说你不知道，不要试图编造答案。
 3. 你需要根据上下文给出详细的回答，不要试图偷懒，不要遗漏括号中的信息，你必须回答的尽可能详细。
 """

bisheng_langchain/rag/qa_corpus/qa_generator.py

@@ -8,7 +8,7 @@ from tqdm import tqdm
 from langchain.document_loaders import PyPDFLoader
 from langchain_core.prompts import PromptTemplate
 from bisheng_langchain.document_loaders import ElemUnstructuredLoader
-from ragas.trainset import TrainsetGenerator
+from bisheng_ragas.trainset import TrainsetGenerator
 
 
 prompt_template = """Use the following pieces of context to answer the question at the end. If you don't know the answer, just say that you don't know, don't try to make up an answer.

bisheng_langchain/rag/scoring/ragas_score.py

@@ -9,8 +9,8 @@ from typing import Any, Dict, List, Optional, Tuple, Union
 import pandas as pd
 from datasets import Dataset
 from loguru import logger
-from ragas import evaluate
-from ragas.metrics import AnswerCorrectness, AnswerCorrectnessBisheng, AnswerRecallBisheng
+from bisheng_ragas import evaluate
+from bisheng_ragas.metrics import AnswerCorrectness, AnswerCorrectnessBisheng, AnswerRecallBisheng
 
 
 @dataclass

bisheng_langchain/rag/utils.py

@@ -43,13 +43,17 @@ def import_by_type(_type: str, name: str) -> Any:
         'llms': {
             'llm': import_llm,
             'chat': import_chat_llm,
-            'contribute': import_chain_contribute_llm
+            'contribute': import_chain_contribute_llm,
+            'chatopenai': import_chat_openai,
         },
         'tools': import_tool,
         'chains': import_chain,
         'toolkits': import_toolkit,
         'memory': import_memory,
-        'embeddings': import_embedding,
+        'embeddings': {
+            'openaiembeddings': import_openai_embeddings,
+            'embeddings': import_embedding,
+        },
         'vectorstores': import_vectorstore,
         'documentloaders': import_documentloader,
         'textsplitters': import_textsplitter,
@@ -60,8 +64,17 @@ def import_by_type(_type: str, name: str) -> Any:
         'inputOutput': import_inputoutput,
     }
     if _type == 'llms':
-        key = 'contribute' if name in chat_models.__all__ else 'chat' if 'chat' in name.lower(
-        ) else 'llm'
+        if name.lower() == 'chatopenai':
+            key = 'chatopenai'
+        else:
+            key = 'contribute' if name in chat_models.__all__ else 'chat' if 'chat' in name.lower(
+            ) else 'llm'
+        loaded_func = func_dict[_type][key]  # type: ignore
+    elif _type == 'embeddings':
+        if name.lower() == 'openaiembeddings':
+            key = 'openaiembeddings'
+        else:
+            key = 'embeddings'
         loaded_func = func_dict[_type][key]  # type: ignore
     else:
         loaded_func = func_dict[_type]
@@ -129,6 +142,11 @@ def import_llm(llm: str) -> BaseLanguageModel:
     return import_class(f'langchain.llms.{llm}')
 
 
+def import_chat_openai(llm: str) -> BaseLanguageModel:
+    """Import llm from llm name"""
+    return import_class(f'langchain_openai.{llm}')
+
+
 def import_tool(tool: str) -> BaseTool:
     """Import tool from tool name"""
     return import_class(f'langchain.tools.{tool}')
@@ -150,6 +168,11 @@ def import_embedding(embedding: str) -> Any:
     return import_class(f'langchain.embeddings.{embedding}')
 
 
+def import_openai_embeddings(embedding: str) -> Any:
+    """Import embedding from embedding name"""
+    return import_class(f'langchain_openai.{embedding}')
+
+
 def import_vectorstore(vectorstore: str) -> Any:
     """Import vectorstore from vectorstore name"""
     from bisheng_langchain import vectorstores

bisheng_langchain/sql/__init__.py

@@ -0,0 +1,3 @@
+from bisheng_langchain.sql.base import SQLDatabaseChain
+
+__all__ = ['SQLDatabaseChain']

bisheng_langchain/sql/base.py

@@ -0,0 +1,120 @@
+"""Chain for interacting with SQL Database."""
+from __future__ import annotations
+
+from typing import Any, Dict, List, Optional
+
+from langchain.callbacks.manager import AsyncCallbackManagerForChainRun, CallbackManagerForChainRun
+from langchain.chains.llm import LLMChain
+from langchain.prompts.prompt import PromptTemplate
+from langchain.tools.sql_database.prompt import QUERY_CHECKER
+from langchain_experimental.sql import SQLDatabaseChain as SQLDatabaseChainExperimental
+
+INTERMEDIATE_STEPS_KEY = 'intermediate_steps'
+
+
+class SQLDatabaseChain(SQLDatabaseChainExperimental):
+    """Chain for interacting with SQL Database.
+
+    Example:
+        .. code-block:: python
+
+            from langchain_experimental.sql import SQLDatabaseChain
+            from langchain.llms import OpenAI, SQLDatabase
+            db = SQLDatabase(...)
+            db_chain = SQLDatabaseChain.from_llm(OpenAI(), db)
+
+    *Security note*: Make sure that the database connection uses credentials
+        that are narrowly-scoped to only include the permissions this chain needs.
+        Failure to do so may result in data corruption or loss, since this chain may
+        attempt commands like `DROP TABLE` or `INSERT` if appropriately prompted.
+        The best way to guard against such negative outcomes is to (as appropriate)
+        limit the permissions granted to the credentials used with this chain.
+        This issue shows an example negative outcome if these steps are not taken:
+        https://github.com/langchain-ai/langchain/issues/5923
+    """
+
+    async def _acall(
+        self,
+        inputs: Dict[str, Any],
+        run_manager: Optional[AsyncCallbackManagerForChainRun] = None,
+    ) -> Dict[str, Any]:
+        _run_manager = run_manager or CallbackManagerForChainRun.get_noop_manager()
+        input_text = f'{inputs[self.input_key]}\nSQLQuery:'
+        await _run_manager.on_text(input_text, verbose=self.verbose)
+        # If not present, then defaults to None which is all tables.
+        table_names_to_use = inputs.get('table_names_to_use')
+        table_info = self.database.get_table_info(table_names=table_names_to_use)
+        llm_inputs = {
+            'input': input_text,
+            'top_k': str(self.top_k),
+            'dialect': self.database.dialect,
+            'table_info': table_info,
+            'stop': ['\nSQLResult:'],
+        }
+        if self.memory is not None:
+            for k in self.memory.memory_variables:
+                llm_inputs[k] = inputs[k]
+        intermediate_steps: List = []
+        try:
+            intermediate_steps.append(llm_inputs.copy())  # input: sql generation
+            sql_cmd = await self.llm_chain.apredict(
+                callbacks=_run_manager.get_child(),
+                **llm_inputs,
+            )
+            sql_cmd = sql_cmd.strip()
+            if self.return_sql:
+                return {self.output_key: sql_cmd}
+            if not self.use_query_checker:
+                await _run_manager.on_text(sql_cmd, color='green', verbose=self.verbose)
+                intermediate_steps.append(sql_cmd)  # output: sql generation (no checker)
+                intermediate_steps.append({'sql_cmd': sql_cmd})  # input: sql exec
+                result = self.database.run(sql_cmd)
+                intermediate_steps.append(str(result))  # output: sql exec
+            else:
+                query_checker_prompt = self.query_checker_prompt or PromptTemplate(
+                    template=QUERY_CHECKER, input_variables=['query', 'dialect'])
+                query_checker_chain = LLMChain(llm=self.llm_chain.llm, prompt=query_checker_prompt)
+                query_checker_inputs = {
+                    'query': sql_cmd,
+                    'dialect': self.database.dialect,
+                }
+                checked_sql_command: str = await query_checker_chain.apredict(
+                    callbacks=_run_manager.get_child(), **query_checker_inputs)
+                checked_sql_command = checked_sql_command.strip()
+                intermediate_steps.append(checked_sql_command)  # output: sql generation (checker)
+                await _run_manager.on_text(checked_sql_command,
+                                           color='green',
+                                           verbose=self.verbose)
+                intermediate_steps.append({'sql_cmd': checked_sql_command})  # input: sql exec
+                result = self.database.run(checked_sql_command)
+                intermediate_steps.append(str(result))  # output: sql exec
+                sql_cmd = checked_sql_command
+
+            await _run_manager.on_text('\nSQLResult: ', verbose=self.verbose)
+            await _run_manager.on_text(result, color='yellow', verbose=self.verbose)
+            # If return direct, we just set the final result equal to
+            # the result of the sql query result, otherwise try to get a human readable
+            # final answer
+            if self.return_direct:
+                final_result = result
+            else:
+                await _run_manager.on_text('\nAnswer:', verbose=self.verbose)
+                input_text += f'{sql_cmd}\nSQLResult: {result}\nAnswer:'
+                llm_inputs['input'] = input_text
+                intermediate_steps.append(llm_inputs.copy())  # input: final answer
+                final_result = await self.llm_chain.apredict(
+                    callbacks=_run_manager.get_child(),
+                    **llm_inputs,
+                )
+                final_result = final_result.strip()
+                intermediate_steps.append(final_result)  # output: final answer
+                await _run_manager.on_text(final_result, color='green', verbose=self.verbose)
+            chain_result: Dict[str, Any] = {self.output_key: final_result}
+            if self.return_intermediate_steps:
+                chain_result[INTERMEDIATE_STEPS_KEY] = intermediate_steps
+            return chain_result
+        except Exception as exc:
+            # Append intermediate steps to exception, to aid in logging and later
+            # improvement of few shot prompt seeds
+            exc.intermediate_steps = intermediate_steps  # type: ignore
+            raise exc

bisheng_langchain/text_splitter.py

@@ -163,7 +163,7 @@ class ElemCharacterTextSplitter(RecursiveCharacterTextSplitter):
             searcher = IntervalSearch(indexes)
             split_texts = self.split_text(text)
             for chunk in split_texts:
-                new_metadata = {}
+                new_metadata = copy.deepcopy(metadatas[i])
                 index = text.find(chunk, index + 1)
                 inter0 = [index, index + len(chunk) - 1]
                 norm_inter = searcher.find(inter0)

{bisheng_langchain-0.3.1.1.dist-info → bisheng_langchain-0.3.2.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: bisheng-langchain
-Version: 0.3.1.1
+Version: 0.3.2
 Summary: bisheng langchain modules
 Home-page: https://github.com/dataelement/bisheng
 Author: DataElem
@@ -31,6 +31,8 @@ Requires-Dist: filetype ==1.2.0
 Requires-Dist: langgraph ==0.0.30
 Requires-Dist: openai ==1.14.3
 Requires-Dist: langchain-openai ==0.1.0
+Requires-Dist: llama-index ==0.9.48
+Requires-Dist: bisheng-ragas ==1.0.0
 
 ## What is bisheng-langchain?
 

{bisheng_langchain-0.3.1.1.dist-info → bisheng_langchain-0.3.2.dist-info}/RECORD

@@ -1,5 +1,5 @@
 bisheng_langchain/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-bisheng_langchain/text_splitter.py,sha256=OAD57cAxPfZoVeBTzicLycel4RehxMULi6Ebi2Wc0Rg,7909
+bisheng_langchain/text_splitter.py,sha256=8snY_Fojh-A1EEGXBSEqCh0N77KQc_dqsgeptAlf344,7934
 bisheng_langchain/agents/__init__.py,sha256=ctsKj77fS8qlkhz_9sS_AhCjFvFNxEpJ9KBYVrApLRg,226
 bisheng_langchain/agents/chatglm_functions_agent/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 bisheng_langchain/agents/chatglm_functions_agent/base.py,sha256=tyytq0XIFXpfxDP0s5QKeprKOunMqi1fHMfQ0-kOmDE,13674
@@ -12,7 +12,7 @@ bisheng_langchain/autogen_role/assistant.py,sha256=rqUaD6fbW6d1jtzfrUQv5pJMKJgVG
 bisheng_langchain/autogen_role/custom.py,sha256=vAyEGxnmV9anyLL12v4ZB_A2VOPwdl-kjGP037I8jPw,2464
 bisheng_langchain/autogen_role/groupchat_manager.py,sha256=AybsH3duoAFpo3bojOYVeSOE4iYkkbgmYIga6m2Jj_Y,2234
 bisheng_langchain/autogen_role/user.py,sha256=fbaORhC7oQjxGhc2RYIWpELdIogFBsgqgQUhZsK6Osk,5715
-bisheng_langchain/chains/__init__.py,sha256=oxN2tUMt_kNxKd_FzCQ7x8xIwojtdCNNKo-DI7q0unM,759
+bisheng_langchain/chains/__init__.py,sha256=Qr71zsX0yO8RMtn2EWEtxI_NRiLlLqgn1mbPumT3vag,960
 bisheng_langchain/chains/loader_output.py,sha256=02ZercAFaudStTZ4t7mcVkGRj5pD78HZ6NO8HbmbDH8,1903
 bisheng_langchain/chains/transform.py,sha256=G2fMqoMB62e03ES--aoVjEo06FzYWb87jCt3EOsiwwg,2805
 bisheng_langchain/chains/autogen/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -21,6 +21,11 @@ bisheng_langchain/chains/combine_documents/__init__.py,sha256=47DEQpj8HBSa-_TImW
 bisheng_langchain/chains/combine_documents/stuff.py,sha256=z_E_wfhJrAYWcNVRPomPm5fGRDI3hqoC52wcMzgzxVA,2369
 bisheng_langchain/chains/conversational_retrieval/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 bisheng_langchain/chains/conversational_retrieval/base.py,sha256=XiqBqov6No-wTVCou6qyMT5p2JQgoQI7OLQOYH8XUos,5313
+bisheng_langchain/chains/qa_generation/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+bisheng_langchain/chains/qa_generation/base.py,sha256=VYGmLDB0bnlDQ6T8ivLP55wwFbMo9HOzlPEDUuRx5fU,4148
+bisheng_langchain/chains/qa_generation/base_v2.py,sha256=v16zyj1-f17SpibREyO1gz0b7FdJDyTOYmqV1FVFFbY,13839
+bisheng_langchain/chains/qa_generation/prompt.py,sha256=4eJk9aDUYDN1qaaYRPy9EobCIncnwS8BbQaDFzzePtM,1944
+bisheng_langchain/chains/qa_generation/prompt_v2.py,sha256=sQLanA_iOnLqrUIwzfTOTANt-1vJ44CM54HFDU8Jo1Q,8938
 bisheng_langchain/chains/question_answering/__init__.py,sha256=_gOZMc-SWprK6xc-Jj64jcr9nc-G4YkZbEYwfJNq_bY,8795
 bisheng_langchain/chains/retrieval/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 bisheng_langchain/chains/retrieval/retrieval_chain.py,sha256=7VLJ-IPVjKfmAVgVET4cvKCO9DCMxwsGgVhW-wz5RZM,3050
@@ -49,10 +54,10 @@ bisheng_langchain/document_loaders/custom_kv.py,sha256=xWUPhcr1hjbdya4zgEHG4Fl0s
 bisheng_langchain/document_loaders/elem_html.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 bisheng_langchain/document_loaders/elem_image.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 bisheng_langchain/document_loaders/elem_pdf.py,sha256=K-TXILGNFLFjavhun_MFbUF4t2_WGA3Z-kbnr75lmW8,22243
-bisheng_langchain/document_loaders/elem_unstrcutured_loader.py,sha256=Dcu70Wz4vdUpq9y2vWPWroRdq6JrlOJNERoFTnZO0oU,5336
-bisheng_langchain/document_loaders/universal_kv.py,sha256=dJF_GQGKBMUjB_kX9CSp7xZRhXgwVuGPbMIzJwPh-C0,4063
+bisheng_langchain/document_loaders/elem_unstrcutured_loader.py,sha256=ODAveZ1Rbq9SOO_qyIgH_NL8L2QqUZXOqf3wCKGYaRE,6984
+bisheng_langchain/document_loaders/universal_kv.py,sha256=ZdIgFIc2fH2kkvJNb7j2wi6FLS_PaaatVy6z_YNV2hw,4114
 bisheng_langchain/document_loaders/parsers/__init__.py,sha256=OOM_FJkwaU-zNS58fASw0TH8FNT6VXKb0VrvisgdrII,171
-bisheng_langchain/document_loaders/parsers/ellm_client.py,sha256=B4Dea8xXXnGvB9j2OXv53HILNUmnWeNJz9ssNM-2fLM,1760
+bisheng_langchain/document_loaders/parsers/ellm_client.py,sha256=Y_CRYwBr-gFArOirF1b76KyI5N8eVpsLeDiIsKtYkpU,1641
 bisheng_langchain/document_loaders/parsers/image.py,sha256=7Vx4dD_WiSTojS4TMIJFxfE8nvze0kwNnwTd6f1cLds,938
 bisheng_langchain/document_loaders/parsers/ocr_client.py,sha256=rRh1coJYn24n7FaINBZH5yO6Edm9TRywY6UOXpcerVo,1612
 bisheng_langchain/document_loaders/parsers/test_image.py,sha256=EJHozq5oFfLBlLL5Lr6XFkrkvSttPpohprs9OjDzAKM,8685
@@ -85,7 +90,7 @@ bisheng_langchain/gpts/tools/api_tools/__init__.py,sha256=CkEjgIFM4GIv86V1B7SsFL
 bisheng_langchain/gpts/tools/api_tools/base.py,sha256=fWQSDIOVb4JZrtJ9ML9q2ycsAa-_61gXTD0MT19J1LM,3618
 bisheng_langchain/gpts/tools/api_tools/flow.py,sha256=ot2YAYgQGWgUpb2nCECAmpqHY6m0SgzwkupF9kDT3lU,2461
 bisheng_langchain/gpts/tools/api_tools/macro_data.py,sha256=FyG-qtl2ECS1CDKt6olN0eDTDM91d-UvDkMDBiVLgYQ,27429
-bisheng_langchain/gpts/tools/api_tools/openapi.py,sha256=3MlIpzR_NhZogaA-zbH6fnT_KUMm10NnbtDYcxKedS8,3907
+bisheng_langchain/gpts/tools/api_tools/openapi.py,sha256=CzKt9FRkgngBcWgabD4emPqAXkAgagkD-pMjG680MTE,3903
 bisheng_langchain/gpts/tools/api_tools/sina.py,sha256=GGA4ZYvNEpqBZ_l8MUYqgkI8xZe9XcGa9-KlHZVqr6I,9542
 bisheng_langchain/gpts/tools/api_tools/tianyancha.py,sha256=abDAz-yAH1-2rKiSmZ6TgnrNUnpgAZpDY8oDiWfWapc,6684
 bisheng_langchain/gpts/tools/bing_search/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -101,22 +106,22 @@ bisheng_langchain/gpts/tools/get_current_time/tool.py,sha256=3uvk7Yu07qhZy1sBrFM
 bisheng_langchain/input_output/__init__.py,sha256=sW_GB7MlrHYsqY1Meb_LeimQqNsMz1gH-00Tqb2BUyM,153
 bisheng_langchain/input_output/input.py,sha256=I5YDmgbvvj1o2lO9wi8LE37wM0wP5jkhUREU32YrZMQ,1094
 bisheng_langchain/input_output/output.py,sha256=6U-az6-Cwz665C2YmcH3SYctWVjPFjmW8s70CA_qphk,11585
-bisheng_langchain/rag/__init__.py,sha256=gGa3hx0HjA_11FxbnR3P0C0y_OwcZM-sYTHdAajmylk,102
+bisheng_langchain/rag/__init__.py,sha256=Rm_cDxOJINt0H4bOeUo3JctPxaI6xKKXZcS-R_wkoGs,198
+bisheng_langchain/rag/bisheng_rag_chain.py,sha256=pbvaLuAwBWyGyvi0qBcqrOz9HScEPfpfEE9GWobVv-M,5891
 bisheng_langchain/rag/bisheng_rag_pipeline.py,sha256=neoBK3TtuQ07_WeuJCzYlvtsDQNepUa_68NT8VCgytw,13749
-bisheng_langchain/rag/bisheng_rag_pipeline_v2.py,sha256=FWmhOSLAKWqqhdj55Y6WoFJ5GYwK8fJdN3o-mqPmabI,15898
-bisheng_langchain/rag/bisheng_rag_pipeline_v2_cohere_raw_prompting.py,sha256=Gql3IXaSTtW8WaC7wLLid0TURXoF0hRqAgl3GAYBBCo,16816
-bisheng_langchain/rag/bisheng_rag_tool.py,sha256=X7mhqYHapLlqfpu90BaW-C3aDeW_PENIci9lNyQv-tI,11971
+bisheng_langchain/rag/bisheng_rag_pipeline_v2.py,sha256=iOoF7mbLp9qDGPsV0fEmgph_Ba8VnECYvCPebXk8xmo,16144
+bisheng_langchain/rag/bisheng_rag_tool.py,sha256=8M59rQqUGS7b1GH_pjNwcqJa2HikK3B4ZqWl59ZCUJY,13113
 bisheng_langchain/rag/extract_info.py,sha256=emrdvzU9EfdTQU_IBF4RkwaTXmYLk1MOJgdfYhw6VtA,1753
 bisheng_langchain/rag/run_qa_gen_web.py,sha256=-fIvHNnD3lD0iNU5m0Me1GDwRjlcsB8tE5RnPtFRG2s,1840
 bisheng_langchain/rag/run_rag_evaluate_web.py,sha256=a9vMhq-ZhEiHHr43uKUzKtjdk280uAP_UHQW_eOaQMw,2224
-bisheng_langchain/rag/utils.py,sha256=ZR3BwjjLX4XT49aRDv-pf8Z4Mqcbmt2_qbDSwXfDfso,6361
+bisheng_langchain/rag/utils.py,sha256=ecl4sDR8iUrVCBRPAAT0hZOHkH50-TLS3567GLP1sRM,7122
 bisheng_langchain/rag/config/baseline.yaml,sha256=cFkfVpuNtmexYkQxXbqbcqQ4eDhx81kmIXf8T_rc_Oo,2266
 bisheng_langchain/rag/config/baseline_caibao.yaml,sha256=uU2HSlPXB8uLq9Fh03wENs06STdMA9CS6ry3X-Wkn-E,2174
 bisheng_langchain/rag/config/baseline_caibao_knowledge_v2.yaml,sha256=KrkMmHu-myQ8u4FlydXLwsppD4ucQGAYYtGGA0HCJzE,2812
-bisheng_langchain/rag/config/baseline_caibao_v2.yaml,sha256=WY9wek70eeBhhyiNmkfyrffxhTfBntqPyLMRmoqM_vE,2846
+bisheng_langchain/rag/config/baseline_caibao_v2.yaml,sha256=0bliqbC8nDK9M24TdSlSsJL0PHRNRGKM8XInZVjJ8yw,2845
 bisheng_langchain/rag/config/baseline_demo_v2.yaml,sha256=hCa7bZMCcOWpu6bsAFX4DU1b7MV4eN2phyzvZzFZKkU,2382
 bisheng_langchain/rag/config/baseline_s2b_mix.yaml,sha256=rkPfzU2-mvjRrZ0zMHaQsncPhq8DrdvVFsw4Sg_jeKc,2398
-bisheng_langchain/rag/config/baseline_v2.yaml,sha256=yUMlK9CbeIqIrb4iJj14BaqH7JJ15QdUIf1EHd7T_R8,2350
+bisheng_langchain/rag/config/baseline_v2.yaml,sha256=RP-DwIRIS_ZK8ixbXi2Z28rKqHD56pWmr2o2WWIwq3Y,2382
 bisheng_langchain/rag/init_retrievers/__init__.py,sha256=qpLLAuqZPtumTlJj17Ie5AbDDmiUiDxYefg_pumqu-c,218
 bisheng_langchain/rag/init_retrievers/baseline_vector_retriever.py,sha256=oRKZZpxlLQAtsubIcAXeXpf1a9h6Pt6uOtNTLeD2jps,2362
 bisheng_langchain/rag/init_retrievers/keyword_retriever.py,sha256=NRT0fBx6HFR7j9IbRl_NBuqF7hnL-9v5GCqHpgnrfPQ,2523
@@ -124,26 +129,28 @@ bisheng_langchain/rag/init_retrievers/mix_retriever.py,sha256=Whxq4kjNPLsxnHcVo6
 bisheng_langchain/rag/init_retrievers/smaller_chunks_retriever.py,sha256=RQ7QLEOOhBrkw-EimXVJqIGa96D-KkNDik2h9hzg9fU,3805
 bisheng_langchain/rag/prompts/__init__.py,sha256=IUCq9gzqGQN_6IDk0D_F5t3mOUI_KbmSzYnnXoX4VKE,223
 bisheng_langchain/rag/prompts/extract_key_prompt.py,sha256=THdcwolRzaKkddgEhabUDytn-caHs9UwECPkPxijWAs,1456
-bisheng_langchain/rag/prompts/prompt.py,sha256=HPQ4mnMXDL9YsgBqK4S1M2sE6icq2uDL17KKgvEkovI,1925
+bisheng_langchain/rag/prompts/prompt.py,sha256=VdRJ5V8zysFALPwA1Ge98aLmk6hnuySegVtL6twMJaI,1916
 bisheng_langchain/rag/prompts/prompt_cohere.py,sha256=FhW-RCIVnIfKtEqV_-7aQav1riIJmA0jV2hGxCSoysk,8151
 bisheng_langchain/rag/qa_corpus/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-bisheng_langchain/rag/qa_corpus/qa_generator.py,sha256=ZdipOLBd16XfKXygq1JeUx4iv3tfUsprk1HNuC5hcIk,6092
+bisheng_langchain/rag/qa_corpus/qa_generator.py,sha256=M8B739ZwwTfF1HXlfTss_pGH38KncJoL9xU0OvEGKCA,6100
 bisheng_langchain/rag/rerank/__init__.py,sha256=lcMRa_o5HGNnN9Hn6zHCimYq02OuD9pRAg6eFPZNgR8,72
 bisheng_langchain/rag/rerank/rerank.py,sha256=RoEwFFb4t4l0aZQmE2-HHHfWIzNLKe9NuFDIoYoWz5g,1836
 bisheng_langchain/rag/rerank/rerank_benchmark.py,sha256=OU9bh7dQ0-faw4hpz6BPEtmbQEhISTekvRiWpksOfCc,4490
 bisheng_langchain/rag/scoring/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 bisheng_langchain/rag/scoring/llama_index_score.py,sha256=Xh8YFT14JGxE3OL2x4JSRa8nGlOK5DlWlsrzbmA2gu8,3001
-bisheng_langchain/rag/scoring/ragas_score.py,sha256=ClVoRg1O7s5OMcsB2E5UbVHj_ErQ3WKr9hao8BcK6NA,7110
+bisheng_langchain/rag/scoring/ragas_score.py,sha256=yiF0O1Wo-yR-aQMpEoK0Je2iln6KYF3xDK5bfADE17Y,7126
 bisheng_langchain/retrievers/__init__.py,sha256=XqBeNyPyNCJf-SzNBiFlkxtjrtHUFTTi5pe2yPyOKrA,210
 bisheng_langchain/retrievers/ensemble.py,sha256=umjBaZYBEdhJ2F7GlzQgXVLYjSfpybTptiJJbUgryZE,5975
 bisheng_langchain/retrievers/mix_es_vector.py,sha256=dSrrsuMPSgGiu181EOzACyIKiDXR0qNBQz_914USD3E,4465
+bisheng_langchain/sql/__init__.py,sha256=2arRtNQ-kUvIsy_8v_PrLxf5r9W-S7mbqptG_l4_1RE,88
+bisheng_langchain/sql/base.py,sha256=WNHCy16UoxvDbroHnJq8CsZ9ot4NGflCm8Bgiv45kks,6152
 bisheng_langchain/utils/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 bisheng_langchain/utils/requests.py,sha256=vWGKyNTxApVeaVdKxqACfIT1Q8wMy-jC3kUv2Ce9Mzc,8688
 bisheng_langchain/vectorstores/__init__.py,sha256=zCZgDe7LyQ0iDkfcm5UJ5NxwKQSRHnqrsjx700Fy11M,213
 bisheng_langchain/vectorstores/elastic_keywords_search.py,sha256=Pm1rS50GJ0HWbjBsFDgs28SVuVbjGSRPOor6yJlnE7w,13347
 bisheng_langchain/vectorstores/milvus.py,sha256=lrnezKnYXhyH5M1g3a-Mcwpj9mwzAj44TKmzyUXlQYY,36297
 bisheng_langchain/vectorstores/retriever.py,sha256=hj4nAAl352EV_ANnU2OHJn7omCH3nBK82ydo14KqMH4,4353
-bisheng_langchain-0.3.1.1.dist-info/METADATA,sha256=w35pNrjvRMzuCK3sSYIl0rP8Q4GL9kkAz5VELfEsVzI,2413
-bisheng_langchain-0.3.1.1.dist-info/WHEEL,sha256=GJ7t_kWBFywbagK5eo9IoUwLW6oyOeTKmQ-9iHFVNxQ,92
-bisheng_langchain-0.3.1.1.dist-info/top_level.txt,sha256=Z6pPNyCo4ihyr9iqGQbH8sJiC4dAUwA_mAyGRQB5_Fs,18
-bisheng_langchain-0.3.1.1.dist-info/RECORD,,
+bisheng_langchain-0.3.2.dist-info/METADATA,sha256=IN1a5Tj7493QanbJRMZ3Tv8swHLKl2s_X9risUMJhfs,2484
+bisheng_langchain-0.3.2.dist-info/WHEEL,sha256=GJ7t_kWBFywbagK5eo9IoUwLW6oyOeTKmQ-9iHFVNxQ,92
+bisheng_langchain-0.3.2.dist-info/top_level.txt,sha256=Z6pPNyCo4ihyr9iqGQbH8sJiC4dAUwA_mAyGRQB5_Fs,18
+bisheng_langchain-0.3.2.dist-info/RECORD,,