search

Cinnamon / kotaemon

An open-source RAG-based tool for chatting with your documents.
https://cinnamon.github.io/kotaemon/
Apache License 2.0
14.49k stars 1.12k forks source link

[BUG] ModuleNotFoundError: No module named 'graphrag' #386

Open Harvester62 opened 1 week ago

Harvester62 commented 1 week ago

Description

I have just updated my local Kotaemon installation using the update_windows.bat successfully (see attached console messages), and immediately after that I started run_windows.bat, all without errors, then inside the UI I added a .txt document to GraphRag which was successfully indexed and embedded (no errors). But, as soon as I select that doc in the GraphRag and search within it I get an error.

I am attaching the entire console messages from the very beginning, the initial update, then the document load and search error. I hope this might be useful for the devs. kotaemon-update-065-069_graphrag_error.txt

immagine

immagine

Reproduction steps

Follow the above instructions.

Screenshots

![DESCRIPTION](LINK.png)

Logs

See attachment.

Browsers

Firefox

OS

Windows

Additional information

The problem is that even if I do not select any collection, just type "Hello" the Error arises and there is nothing I can do to prevent it. The models are all working fine (tested with the test function within Kotaemon UI).

Harvester62 commented 1 week ago

I am not a dev, so please, be patient with me. About the console messages, after several different attempts to run a query, I did notice that the messages stay always the same, and I suspect the error might be due the the lack of a Reranking solution, which I didn't set up, not having a Cohere account. Just guessing.

Would it be feasible to set a specific model (local/online) for re-ranking? What peculiarity should such a model have to re-rank results properly? These questions might be due to the lack of knowledge about how re-ranking works, sorry.

taprosoft commented 1 week ago

@Harvester62 for GraphRAG to run, make sure that you have set these environment variables in Windows Terminal correctly

# settings for GraphRAG
GRAPHRAG_API_KEY=<YOUR_OPENAI_KEY>
GRAPHRAG_LLM_MODEL=gpt-4o-mini
GRAPHRAG_EMBEDDING_MODEL=text-embedding-3-small

Would it be feasible to set a specific model (local/online) for re-ranking? What peculiarity should such a model have to re-rank results properly?

By default it should work without rerankings. For more specific suggestion please add the Console log.

Harvester62 commented 1 week ago

immagine

This is my current .env file content. I did also a SET OPENAI_API_KEY and SET GRAPHRAG_API_KEY commands inside a Windows console, just in case, but that apparently didn't fix the problem either. Here follow the console messages after restarting the application, and then loading manually a single file in the UI and query it.

Microsoft Windows [Versione 10.0.19045.4957] (c) Microsoft Corporation. Tutti i diritti sono riservati.

C:\Users\Riccardo>set OPENAI_API_KEY=sk-[MYKEY]

C:\Users\Riccardo>set GRAPHRAG_API_KEY=sk-[MYKEY]

C:\Users\Riccardo>cd \kotaemon-app

C:\kotaemon-app>.\scripts\run_windows.bat

Setting up Git

Git is installed at C:\kotaemon-app\install_dir\Git git version 2.46.0.windows.1 Git is added to PATH for this session

Setting up Miniconda

Conda is installed at C:\kotaemon-app\install_dir\conda Conda version: conda 24.7.1

Creating conda environment

Conda environment exists at C:\kotaemon-app\install_dir\env Activate conda environment at C:\kotaemon-app\install_dir\env

Installing Kotaemon

Dependencies are already installed

Setting up a local model

Local model not found: llama3.1:8b

Downloading and extracting PDF.js

Directory already exists: C:\kotaemon-app\libs\ktem\ktem\assets\prebuilt\pdfjs-4.0.379-dist

Launching Kotaemon in your browser, please wait...

Starting Kotaemon UI... (prebuilt PDF.js is at C:\kotaemon-app\libs\ktem\ktem\assets\prebuilt\pdfjs-4.0.379-dist) [nltk_data] Downloading package punkt_tab to C:\kotaemon- [nltk_data] app\install_dir\env\lib\site- [nltk_data] packages\llama_index\core_static/nltk_cache... [nltk_data] Package punkt_tab is already up-to-date! GraphRAG dependencies not installed. GraphRAG retriever pipeline will not work properly. User "admin" already exists Setting up quick upload event Running on local URL: http://127.0.0.1:7860

To create a public link, set share=True in launch(). User-id: None, can see public conversations: False User-id: 1, can see public conversations: True Overriding with default loaders use_quick_index_mode False reader_mode default Using reader TxtReader() User-id: 1, can see public conversations: True Session reasoning type simple Session LLM openai Reasoning class <class 'ktem.reasoning.simple.FullQAPipeline'> Traceback (most recent call last): File "C:\kotaemon-app\install_dir\env\lib\site-packages\gradio\queueing.py", line 536, in process_events response = await route_utils.call_process_api( File "C:\kotaemon-app\install_dir\env\lib\site-packages\gradio\route_utils.py", line 276, in call_process_api output = await app.get_blocks().process_api( File "C:\kotaemon-app\install_dir\env\lib\site-packages\gradio\blocks.py", line 1923, in process_api result = await self.call_function( File "C:\kotaemon-app\install_dir\env\lib\site-packages\gradio\blocks.py", line 1520, in call_function prediction = await utils.async_iteration(iterator) File "C:\kotaemon-app\install_dir\env\lib\site-packages\gradio\utils.py", line 663, in async_iteration return await iterator.anext() File "C:\kotaemon-app\install_dir\env\lib\site-packages\gradio\utils.py", line 656, in anext return await anyio.to_thread.run_sync( File "C:\Users\Riccardo\AppData\Roaming\Python\Python310\site-packages\anyio\to_thread.py", line 31, in run_sync return await get_asynclib().run_sync_in_worker_thread( File "C:\Users\Riccardo\AppData\Roaming\Python\Python310\site-packages\anyio_backends_asyncio.py", line 937, in run_sync_in_worker_thread return await future File "C:\Users\Riccardo\AppData\Roaming\Python\Python310\site-packages\anyio_backends_asyncio.py", line 867, in run result = context.run(func, args) File "C:\kotaemon-app\install_dir\env\lib\site-packages\gradio\utils.py", line 639, in run_sync_iterator_async return next(iterator) File "C:\kotaemon-app\install_dir\env\lib\site-packages\gradio\utils.py", line 801, in gen_wrapper response = next(iterator) File "C:\kotaemon-app\install_dir\env\lib\site-packages\ktem\pages\chat__init.py", line 857, in chat_fn pipeline, reasoning_state = self.create_pipeline( File "C:\kotaemon-app\install_dir\env\lib\site-packages\ktem\pages\chat\init__.py", line 824, in create_pipeline iretrievers = index.get_retriever_pipelines( File "C:\kotaemon-app\install_dir\env\lib\site-packages\ktem\index\file\index.py", line 423, in get_retriever_pipelines obj = cls.get_pipeline(stripped_settings, self.config, selected_ids) File "C:\kotaemon-app\install_dir\env\lib\site-packages\ktem\index\file\pipelines.py", line 287, in get_pipeline "reranking", reranking_models_manager.get_default_name() File "C:\kotaemon-app\install_dir\env\lib\site-packages\ktem\rerankings\manager.py", line 108, in get_default_name raise ValueError("No models in pool") ValueError: No models in pool Traceback (most recent call last): File "C:\kotaemon-app\install_dir\env\lib\site-packages\gradio\queueing.py", line 536, in process_events response = await route_utils.call_process_api( File "C:\kotaemon-app\install_dir\env\lib\site-packages\gradio\route_utils.py", line 276, in call_process_api output = await app.get_blocks().process_api( File "C:\kotaemon-app\install_dir\env\lib\site-packages\gradio\blocks.py", line 1923, in process_api result = await self.call_function( File "C:\kotaemon-app\install_dir\env\lib\site-packages\gradio\blocks.py", line 1508, in call_function prediction = await anyio.to_thread.run_sync( # type: ignore File "C:\Users\Riccardo\AppData\Roaming\Python\Python310\site-packages\anyio\to_thread.py", line 31, in run_sync return await get_asynclib().run_sync_in_worker_thread( File "C:\Users\Riccardo\AppData\Roaming\Python\Python310\site-packages\anyio_backends_asyncio.py", line 937, in run_sync_in_worker_thread return await future File "C:\Users\Riccardo\AppData\Roaming\Python\Python310\site-packages\anyio_backends_asyncio.py", line 867, in run result = context.run(func, args) File "C:\kotaemon-app\install_dir\env\lib\site-packages\gradio\utils.py", line 818, in wrapper response = f(*args, kwargs) File "C:\kotaemon-app\install_dir\env\lib\site-packages\ktem\pages\chat__init.py", line 958, in check_and_suggest_name_conv suggested_name = suggest_pipeline(chat_history).text File "C:\kotaemon-app\install_dir\env\lib\site-packages\theflow\base.py", line 1097, in call raise e from None File "C:\kotaemon-app\install_dir\env\lib\site-packages\theflow\base.py", line 1088, in call output = self.fl.exec(func, args, kwargs) File "C:\kotaemon-app\install_dir\env\lib\site-packages\theflow\backends\base.py", line 151, in exec return run(*args, **kwargs) File "C:\kotaemon-app\install_dir\env\lib\site-packages\theflow\middleware.py", line 144, in call raise e from None File "C:\kotaemon-app\install_dir\env\lib\site-packages\theflow\middleware.py", line 141, in call__ _output = self.next_call(*args, kwargs) File "C:\kotaemon-app\install_dir\env\lib\site-packages\theflow\middleware.py", line 117, in call return self.next_call(*args, kwargs) File "C:\kotaemon-app\install_dir\env\lib\site-packages\theflow\base.py", line 1017, in _runx return self.run(*args, *kwargs) File "C:\kotaemon-app\install_dir\env\lib\site-packages\ktem\reasoning\prompt_optimization\suggest_conversation_name.py", line 32, in run messages.append(AIMessage(content=ai)) File "C:\kotaemon-app\install_dir\env\lib\site-packages\kotaemon\base\schema.py", line 62, in init super().init(args, kwargs) TypeError: AIMessage.init() missing 1 required positional argument: 'content'

That "no models in pool", and the last line complaining about a missing content are suspicious. Which models? What content is missing?

taprosoft commented 5 days ago

Can you check the solution mentioned in https://github.com/Cinnamon/kotaemon/issues/394