Langchain Import Issue

[Fulladorn]

System Info

Python 3.10.8
Langchain==0.0.229

AWS Sagemaker Studio w/ PyTorch 2.0.0 Python 3.10 GPU Optimized image

Who can help?

@hwchase17 or @agola11

Information

• The official example notebooks/scripts
• My own modified scripts

Related Components

• LLMs/Chat Models
• Embedding Models
• Prompts / Prompt Templates / Prompt Selectors
• Output Parsers
• Document Loaders
• Vector Stores / Retrievers
• Memory
• Agents / Agent Executors
• Tools / Toolkits
• Chains
• Callbacks/Tracing
• Async

Reproduction

Was working fine in a Jupyter Notebook in AWS Sagemaker Studio for the past few weeks but today running into an issue with no code changes... import chain issue?

!pip install langchain openai chromadb tiktoken pypdf unstructured pdf2image;

from langchain.document_loaders import TextLoader

Results in:

TypeError                                 Traceback (most recent call last)
Cell In[10], line 1
----> 1 from langchain.document_loaders import TextLoader
      2 docLoader = TextLoader('./docs/nlitest.txt', encoding='utf8')
      3 document = docLoader.load()

File /opt/conda/lib/python3.10/site-packages/langchain/__init__.py:6
      3 from importlib import metadata
      4 from typing import Optional
----> 6 from langchain.agents import MRKLChain, ReActChain, SelfAskWithSearchChain
      7 from langchain.cache import BaseCache
      8 from langchain.chains import (
      9     ConversationChain,
     10     LLMBashChain,
   (...)
     18     VectorDBQAWithSourcesChain,
     19 )

File /opt/conda/lib/python3.10/site-packages/langchain/agents/__init__.py:2
      1 """Interface for agents."""
----> 2 from langchain.agents.agent import (
      3     Agent,
      4     AgentExecutor,
      5     AgentOutputParser,
      6     BaseMultiActionAgent,
      7     BaseSingleActionAgent,
      8     LLMSingleActionAgent,
      9 )
     10 from langchain.agents.agent_toolkits import (
     11     create_csv_agent,
     12     create_json_agent,
   (...)
     21     create_vectorstore_router_agent,
     22 )
     23 from langchain.agents.agent_types import AgentType

File /opt/conda/lib/python3.10/site-packages/langchain/agents/agent.py:25
     17 from langchain.callbacks.base import BaseCallbackManager
     18 from langchain.callbacks.manager import (
     19     AsyncCallbackManagerForChainRun,
     20     AsyncCallbackManagerForToolRun,
   (...)
     23     Callbacks,
     24 )
---> 25 from langchain.chains.base import Chain
     26 from langchain.chains.llm import LLMChain
     27 from langchain.input import get_color_mapping

File /opt/conda/lib/python3.10/site-packages/langchain/chains/__init__.py:3
      1 """Chains are easily reusable components which can be linked together."""
      2 from langchain.chains.api.base import APIChain
----> 3 from langchain.chains.api.openapi.chain import OpenAPIEndpointChain
      4 from langchain.chains.combine_documents.base import AnalyzeDocumentChain
      5 from langchain.chains.combine_documents.map_reduce import MapReduceDocumentsChain

File /opt/conda/lib/python3.10/site-packages/langchain/chains/api/openapi/chain.py:17
     15 from langchain.requests import Requests
     16 from langchain.schema.language_model import BaseLanguageModel
---> 17 from langchain.tools.openapi.utils.api_models import APIOperation
     20 class _ParamMapping(NamedTuple):
     21     """Mapping from parameter name to parameter value."""

File /opt/conda/lib/python3.10/site-packages/langchain/tools/__init__.py:11
      4 from langchain.tools.azure_cognitive_services import (
      5     AzureCogsFormRecognizerTool,
      6     AzureCogsImageAnalysisTool,
      7     AzureCogsSpeech2TextTool,
      8     AzureCogsText2SpeechTool,
      9 )
     10 from langchain.tools.base import BaseTool, StructuredTool, Tool, tool
---> 11 from langchain.tools.bing_search.tool import BingSearchResults, BingSearchRun
     12 from langchain.tools.brave_search.tool import BraveSearch
     13 from langchain.tools.convert_to_openai import format_tool_to_openai_function

File /opt/conda/lib/python3.10/site-packages/langchain/tools/bing_search/__init__.py:3
      1 """Bing Search API toolkit."""
----> 3 from langchain.tools.bing_search.tool import BingSearchResults, BingSearchRun
      5 __all__ = ["BingSearchRun", "BingSearchResults"]

File /opt/conda/lib/python3.10/site-packages/langchain/tools/bing_search/tool.py:10
      5 from langchain.callbacks.manager import (
      6     AsyncCallbackManagerForToolRun,
      7     CallbackManagerForToolRun,
      8 )
      9 from langchain.tools.base import BaseTool
---> 10 from langchain.utilities.bing_search import BingSearchAPIWrapper
     13 class BingSearchRun(BaseTool):
     14     """Tool that adds the capability to query the Bing search API."""

File /opt/conda/lib/python3.10/site-packages/langchain/utilities/__init__.py:3
      1 """General utilities."""
      2 from langchain.requests import TextRequestsWrapper
----> 3 from langchain.utilities.apify import ApifyWrapper
      4 from langchain.utilities.arxiv import ArxivAPIWrapper
      5 from langchain.utilities.awslambda import LambdaWrapper

File /opt/conda/lib/python3.10/site-packages/langchain/utilities/apify.py:5
      1 from typing import Any, Callable, Dict, Optional
      3 from pydantic import BaseModel, root_validator
----> 5 from langchain.document_loaders import ApifyDatasetLoader
      6 from langchain.document_loaders.base import Document
      7 from langchain.utils import get_from_dict_or_env

File /opt/conda/lib/python3.10/site-packages/langchain/document_loaders/__init__.py:44
     39 from langchain.document_loaders.duckdb_loader import DuckDBLoader
     40 from langchain.document_loaders.email import (
     41     OutlookMessageLoader,
     42     UnstructuredEmailLoader,
     43 )
---> 44 from langchain.document_loaders.embaas import EmbaasBlobLoader, EmbaasLoader
     45 from langchain.document_loaders.epub import UnstructuredEPubLoader
     46 from langchain.document_loaders.evernote import EverNoteLoader

File /opt/conda/lib/python3.10/site-packages/langchain/document_loaders/embaas.py:54
     50     bytes: str
     51     """The base64 encoded bytes of the document to extract text from."""
---> 54 class BaseEmbaasLoader(BaseModel):
     55     """Base class for embedding a model into an Embaas document extraction API."""
     57     embaas_api_key: Optional[str] = None

File /opt/conda/lib/python3.10/site-packages/pydantic/main.py:204, in pydantic.main.ModelMetaclass.__new__()

File /opt/conda/lib/python3.10/site-packages/pydantic/fields.py:488, in pydantic.fields.ModelField.infer()

File /opt/conda/lib/python3.10/site-packages/pydantic/fields.py:419, in pydantic.fields.ModelField.__init__()

File /opt/conda/lib/python3.10/site-packages/pydantic/fields.py:539, in pydantic.fields.ModelField.prepare()

File /opt/conda/lib/python3.10/site-packages/pydantic/fields.py:801, in pydantic.fields.ModelField.populate_validators()

File /opt/conda/lib/python3.10/site-packages/pydantic/validators.py:696, in find_validators()

File /opt/conda/lib/python3.10/site-packages/pydantic/validators.py:585, in pydantic.validators.make_typeddict_validator()

File /opt/conda/lib/python3.10/site-packages/pydantic/annotated_types.py:35, in pydantic.annotated_types.create_model_from_typeddict()

File /opt/conda/lib/python3.10/site-packages/pydantic/main.py:972, in pydantic.main.create_model()

File /opt/conda/lib/python3.10/site-packages/pydantic/main.py:204, in pydantic.main.ModelMetaclass.__new__()

File /opt/conda/lib/python3.10/site-packages/pydantic/fields.py:488, in pydantic.fields.ModelField.infer()

File /opt/conda/lib/python3.10/site-packages/pydantic/fields.py:419, in pydantic.fields.ModelField.__init__()

File /opt/conda/lib/python3.10/site-packages/pydantic/fields.py:534, in pydantic.fields.ModelField.prepare()

File /opt/conda/lib/python3.10/site-packages/pydantic/fields.py:638, in pydantic.fields.ModelField._type_analysis()

File /opt/conda/lib/python3.10/typing.py:1158, in _SpecialGenericAlias.__subclasscheck__(self, cls)
   1156     return issubclass(cls.__origin__, self.__origin__)
   1157 if not isinstance(cls, _GenericAlias):
-> 1158     return issubclass(cls, self.__origin__)
   1159 return super().__subclasscheck__(cls)

TypeError: issubclass() arg 1 must be a class ```

### Expected behavior

The module should import with no error.

[Fulladorn]

I saw 2 similar issues on Discord, Langchain was working for a few weeks and then today is presenting the above error.

[luke-j0hnson]

I am also experiencing the above error

[dosubot]

Answer generated by a 🤖

Answer

I found related issues that might be helpful. I did my best to summarize the solutions, but I recommend looking at the related issues yourself.

Closed Issues

ImportError: cannot import name 'LLMChain' from partially initialized module 'langchain'

The suggested solution is: Rename the file from requests.py to requests_chain.py to resolve the issue.

Unable to import from langchain.document_loaders

No solution found.

Open Issues

Issue: Can not import the Langchain modules.

The suggested solution is: The user should upgrade their Python version to 3.9 and then install LangChain version 0.0.219. The issue seems to be due to a version mismatch in the Pydantic module.

ERROR-Import Langchain : TypeError: dataclass_transform() got an unexpected keyword argument 'field_specifiers'

The suggested solution is:

Create a new environment with Python version 3.9.12. Then install langchain (pip install langchain) and chromadb (pip install chromadb). This should resolve the issue.

---

This response is meant to be useful, save you time, and share context. It is not meant to be a precise solution, but rather a starting point for your own research.

Help me be more useful! Please leave a 👍 if this is helpful and 👎 if it is irrelevant.

[Fulladorn]

@luke-j0hnson

Can you run pip uninstall -y langchain and then reinstall and see if your issue persists? It seemed to work for me...

Edit: Worked for 1 notebook, but not the other... hmm

[luke-j0hnson]

Didn’t solve the issue in my notebook either. Code has worked for weeks stopped working tonight with the same error

[softzer0]

Same here, using latest Python (3.11.4) in Docker environment and Langchain 0.0.229 (latest in PyPI). Tried to upgrade Pydantic to the latest version, as suggested here, but I've got this error:

The conflict is caused by:
     The user requested pydantic~=2.0.2
     langchain 0.0.229 depends on pydantic<2 and >=1

Then I've tried to upgrade to 1.10.11, but I use ChromaDB (0.3.27) which depends on version 1.9 exclusively.
Also, manually installing typing-inspect and typing_extensions as it was also suggested in the linked StackOverflow question didn't help either.

[Fulladorn]

@luke-j0hnson Can you duplicate the notebook and try and run it again?

[Fulladorn]

Same here, using latest Python (3.11.4) in Docker environment and Langchain 0.0.229 (latest in PyPI). Tried to upgrade Pydantic to the latest version, as suggested here, but I've got this error:

The conflict is caused by:
     The user requested pydantic~=2.0.2
     langchain 0.0.229 depends on pydantic<2 and >=1

Then I've tried to upgrade to 1.10.11, but I use ChromaDB which depends on version 1.9 exclusively. Also, manually installing typing-inspect and typing_extensions as it was also suggested in the linked StackOverflow question didn't help either.

Is your development environment a Jupyter Notebook?

[softzer0]

No, it's a standard one, sorry that I forgot to mention that only difference.

[Fulladorn]

I'm not sure why my duplicated notebook is working now and hasn't been changed at all...

[lukafilipxvic]

Try using chromadb==0.3.26
There is a problem with Chroma 0.3.27 and Pydantic. chroma-core/chroma#785

[wiskasEpiphi]

I had the same issue and I solved it installing chromadb==0.3.26 and chromadb==0.3.26 versions.

[priyamsaha1012]

pip install pydantic==1.10.8 working

[softzer0]

I've tried with ChromaDB 0.3.26 as suggested but that didn't work either :( Seems like Pydantic needs to be upgraded too, but ChromaDB requirement prevents it to be upgraded to the version greater than 1.9.

Update: Putting pydantic==1.10.8 after chromadb==0.3.26 in requirements.txt did the trick, thanks!

[taranjeet]

embedchain is also facing some issue due to dependency on chroma.
Fixed by locking the package version in setup.py for now: mem0ai/mem0@34fd418

[bartoncreek]

FWIW, following directions above of @lukafilipxvic and @softzer0, my below is working, and my three-day nightmare is over.

[tool.poetry.dependencies]
python = "^3.10"
pydantic = "1.10.8"
chromadb = "0.3.26"
#chromadb = ">=0.3.27"
#pydantic = "1.9.0"
langchain = "^0.0.231"
llama-index = "^0.7.5"
openai = "^0.27.8"
ipykernel = "^6.24.0"
python-dotenv = "^1.0.0"

[baskaryan]

believe this is same as #7548

[dosubot]

Hi, @Fulladorn! I'm Dosu, and I'm here to help the LangChain team manage their backlog. I wanted to let you know that we are marking this issue as stale.

Based on my understanding, the issue you reported is related to the module langchain.document_loaders not being importable due to a TypeError. There have been some suggestions in the comments, such as renaming a file, upgrading Python, and installing a specific version of LangChain. Additionally, there is a discussion about conflicts with Pydantic and ChromaDB versions.

Before we close this issue, we wanted to check with you if it is still relevant to the latest version of the LangChain repository. If it is, please let us know by commenting on the issue. Otherwise, feel free to close the issue yourself, or it will be automatically closed in 7 days.

Thank you for your understanding and contribution to the LangChain project!

[Jamaludiin]

Please assist me for this problem

ImportError: cannot import name 'Ollama' from 'langchain.llms'

More details the problem also written in this way

"from langchain.llms import Ollama
ImportError: cannot import name 'Ollama' from 'langchain.llms' (/Library/Frameworks/Python.framework/Versions/3.10/lib/python3.10/site-packages/langchain/llms/init.py)"

[ChamaliVishmani]

use 'langchain_community.llms' to import Ollama

from langchain_community.llms import Ollama

[dk1coding1zone]

note that langchain is not supporting python version 3.8 below try to use 3.8+