build: Upgrade to LangChain 0.2 (#997)

Author: cecheta

code/backend/batch/utilities/document_loading/web.py

@@ -1,6 +1,5 @@
 from typing import List
 import re
-from langchain.docstore.document import Document
 from langchain_community.document_loaders import WebBaseLoader
 from .document_loading_base import DocumentLoadingBase
 from ..common.source_document import SourceDocument
@@ -11,7 +10,7 @@ def __init__(self) -> None:
         super().__init__()
 
     def load(self, document_url: str) -> List[SourceDocument]:
-        documents: List[Document] = WebBaseLoader(document_url).load()
+        documents = WebBaseLoader(document_url).load()
         for document in documents:
             document.page_content = re.sub("\n{3,}", "\n\n", document.page_content)
             # Remove half non-ascii character from start/end of doc content

code/backend/batch/utilities/helpers/azure_search_helper.py

@@ -1,6 +1,6 @@
 import logging
 from typing import Union
-from langchain.vectorstores.azuresearch import AzureSearch
+from langchain_community.vectorstores import AzureSearch
 from azure.core.credentials import AzureKeyCredential
 from azure.identity import DefaultAzureCredential
 from azure.search.documents import SearchClient

code/backend/batch/utilities/helpers/document_chunking_helper.py

@@ -1,6 +1,6 @@
 from typing import List
-from langchain.docstore.document import Document
 
+from ..common.source_document import SourceDocument
 from ..document_chunking.chunking_strategy import ChunkingSettings, ChunkingStrategy
 from ..document_chunking.strategies import get_document_chunker
 
@@ -12,8 +12,8 @@ def __init__(self) -> None:
         pass
 
     def chunk(
-        self, documents: List[Document], chunking: ChunkingSettings
-    ) -> List[Document]:
+        self, documents: List[SourceDocument], chunking: ChunkingSettings
+    ) -> List[SourceDocument]:
         chunker = get_document_chunker(chunking.chunking_strategy.value)
         if chunker is None:
             raise Exception(

code/backend/batch/utilities/helpers/document_loading_helper.py

@@ -1,5 +1,6 @@
 from typing import List
-from langchain.docstore.document import Document
+
+from ..common.source_document import SourceDocument
 from ..document_loading import LoadingSettings
 from ..document_loading.strategies import get_document_loader
 
@@ -8,7 +9,7 @@ class DocumentLoading:
     def __init__(self) -> None:
         pass
 
-    def load(self, document_url: str, loading: LoadingSettings) -> List[Document]:
+    def load(self, document_url: str, loading: LoadingSettings) -> List[SourceDocument]:
         loader = get_document_loader(loading.loading_strategy.value)
         if loader is None:
             raise Exception(

code/backend/batch/utilities/orchestrator/lang_chain_agent.py

@@ -3,7 +3,7 @@
 from langchain.agents import Tool
 from langchain.memory import ConversationBufferMemory
 from langchain.agents import ZeroShotAgent, AgentExecutor
-from langchain.chains import LLMChain
+from langchain.chains.llm import LLMChain
 from langchain_community.callbacks import get_openai_callback
 
 from .orchestrator_base import OrchestratorBase

poetry.lock

@@ -18,8 +18,8 @@ azure-storage-blob = "12.20.0"
 azure-identity = "1.16.0"
 flask = {extras = ["async"], version = "^3.0.3"}
 openai = "1.30.3"
-langchain = "0.1.20"
-langchain-community = "0.0.38"
+langchain = "0.2.1"
+langchain-community = "0.2.1"
 langchain-openai = "0.1.7"
 requests = "2.32.2"
 tiktoken = "0.7.0"