Cinnamon
[BUG] kotaemon 0.9.2 requires umap-learn==0.5.5, but you have umap-learn 0.5.7 which is incompatible.
Description
Hi,
On conda python 3.11 env kotaemon install and run work well.
In second time, I would like to install nanoGraphRAG but dependency problem and starting crash.
Thanks for your work, Kind regards, David
Attempting uninstall: umap-learn
Found existing installation: umap-learn 0.5.5
Uninstalling umap-learn-0.5.5:
Successfully uninstalled umap-learn-0.5.5
ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts.
kotaemon 0.9.2 requires umap-learn==0.5.5, but you have umap-learn 0.5.7 which is incompatible.
Successfully installed Mako-1.3.6 POT-0.9.5 alembic-1.14.0 anytree-2.12.1 asyncer-0.0.8 autograd-1.7.0 beartype-0.18.5 datasets-3.1.0 dill-0.3.8 dspy-2.5.41 dspy-ai-2.5.41 fsspec-2024.9.0 future-1.0.0 gensim-4.3.3 graspologic-3.4.1 graspologic-native-1.2.1 hnswlib-0.8.0 hyppo-0.4.0 json-repair-0.30.2 jsonschema-4.23.0 jsonschema-specifications-2024.10.1 litellm-1.51.0 magicattr-0.1.6 multiprocess-0.70.16 nano-graphrag-0.0.8.2 nano-vectordb-0.0.4.3 neo4j-5.27.0 optuna-4.1.0 patsy-1.0.1 referencing-0.35.1 rpds-py-0.21.0 scipy-1.12.0 seaborn-0.13.2 smart-open-7.0.5 statsmodels-0.14.4 umap-learn-0.5.7 xxhash-3.5.0
starting crash if I dont use your :: To quickly fix: pip uninstall hnswlib chroma-hnswlib && pip install chroma-hnswlib
(kotaemon) :~/kotaemon$ python app.py
[nltk_data] Downloading package punkt_tab to
[nltk_data] /home/dromeuf/anaconda3/envs/kotaemon/lib/python3.11/s
[nltk_data] ite-packages/llama_index/core/_static/nltk_cache...
[nltk_data] Package punkt_tab is already up-to-date!
GraphRAG dependencies not installed. Try `pip install graphrag future` to install. GraphRAG retriever pipeline will not work properly.
LightRAG dependencies not installed. Try `pip install git+https://github.com/HKUDS/LightRAG.git` to install. LighthRAG retriever pipeline will not work properly.
INFO:chromadb.telemetry.product.posthog:Anonymized telemetry enabled. See https://docs.trychroma.com/telemetry for more information.
Traceback (most recent call last):
File "/home/kot/kotaemon/app.py", line 15, in <module>
app = App()
^^^^^
... ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/kot/anaconda3/envs/kotaemon/lib/python3.11/site-packages/chromadb/segment/impl/vector/local_persistent_hnsw.py", line 527, in get_file_handle_count
hnswlib_count = hnswlib.Index.file_handle_count
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
AttributeError: type object 'hnswlib.Index' has no attribute 'file_handle_count'
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "/home/kota/kotaemon/app.py", line 15, in <module>
app = App()
^^^^^
File "/home/kota/kotaemon/libs/ktem/ktem/app.py", line 73, in __init__
... File "/home/kota/kotaemon/libs/ktem/ktem/index/manager.py", line 68, in build_index
raise ValueError(f'Cannot create index "{name}": {e}')
ValueError: Cannot create index "NanoGraphRAG": type object 'hnswlib.Index' has no attribute 'file_handle_count'
If I quick fix with your :: To quickly fix: pip uninstall hnswlib chroma-hnswlib && pip install chroma-hnswlib
I can import PDF documents in NanoGraphRAG collection file. I see entitys in log console...
But when I request prompt in the CHAT with nanoGraphRAG file select, I see Thinking but it displays error surrounded by red in the middle of the chat screen and the information panel on the right.
log
searching in doc_ids []
INFO:ktem.index.file.pipelines:Skip retrieval because of no selected files: DocumentRetrievalPipeline(
(vector_retrieval): <function Function._prepare_child.<locals>.exec at 0x7f2d9c31d9e0>
(embedding): <function Function._prepare_child.<locals>.exec at 0x7f2d9c31d9e0>
)
INFO:httpx:HTTP Request: POST http://localhost:11434/v1/embeddings "HTTP/1.1 200 OK"
GraphRAG embedding dim 5120
INFO:nano-graphrag:Load KV full_docs with 28 data
INFO:nano-graphrag:Load KV text_chunks with 33 data
INFO:nano-graphrag:Load KV llm_response_cache with 90 data
INFO:nano-graphrag:Load KV community_reports with 0 data
INFO:nano-graphrag:Loaded graph from /home/dromeuf/kotaemon/ktem_app_data/user_data/files/nano_graphrag/3266e868-e745-4b76-8d8b-4280c6eb6185/input/graph_chunk_entity_relation.graphml with 56 nodes, 28 edges
INFO:nano-vectordb:Load (52, 5120) data
INFO:nano-vectordb:Init {'embedding_dim': 5120, 'metric': 'cosine', 'storage_file': '/home/kota/kotaemon/ktem_app_data/user_data/files/nano_graphrag/3266e868-e745-4b76-8d8b-4280c6eb6185/input/vdb_entities.json'} 52 data
INFO:httpx:HTTP Request: POST http://localhost:11434/v1/embeddings "HTTP/1.1 200 OK"
Traceback (most recent call last):
File "/home/kota/anaconda3/envs/kotaemon/lib/python3.11/site-packages/gradio/queueing.py", line 575, in process_events
response = await route_utils.call_process_api(
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
... File "/home/kota/anaconda3/envs/kotaemon/lib/python3.11/site-packages/nano_graphrag/_op.py", line 702, in <lambda>
related_community_datas[k]["report_json"].get("rating", -1),
~~~~~~~~~~~~~~~~~~~~~~~^^^
KeyError: '1'
INFO:httpx:HTTP Request: POST http://localhost:11434/v1/chat/completions "HTTP/1.1 200 OK"
User-id: 1, can see public conversations: True
Idem for LightRAG with error dependency.
When I install MS GraphRAG this error dependency. But document integration using embedding qwen2.5:14b seems to work well (but obviously takes a long time) as do chat queries (LLM qwen2.5:14b too) with mindmap and entity graph display.
Attempting uninstall: lancedb
Found existing installation: lancedb 0.16.0
Uninstalling lancedb-0.16.0:
Successfully uninstalled lancedb-0.16.0
ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts.
lightrag-hku 1.0.2 requires hnswlib, which is not installed.
nano-graphrag 0.0.8.2 requires hnswlib, which is not installed.
langchain-community 0.2.11 requires tenacity!=8.4.0,<9.0.0,>=8.1.0, but you have tenacity 9.0.0 which is incompatible.
gradio 4.39.0 requires aiofiles<24.0,>=22.0, but you have aiofiles 24.1.0 which is incompatible.
langchain-core 0.2.43 requires tenacity!=8.4.0,<9.0.0,>=8.1.0, but you have tenacity 9.0.0 which is incompatible.
langchain 0.2.15 requires tenacity!=8.4.0,<9.0.0,>=8.1.0, but you have tenacity 9.0.0 which is incompatible.
llama-index-core 0.10.68.post1 requires tenacity!=8.4.0,<9.0.0,>=8.2.0, but you have tenacity 9.0.0 which is incompatible.
kotaemon 0.9.2 requires tenacity<8.3,>=8.2.3, but you have tenacity 9.0.0 which is incompatible.
kotaemon 0.9.2 requires umap-learn==0.5.5, but you have umap-learn 0.5.7 which is incompatible.
llama-index-legacy 0.9.48.post4 requires tenacity<9.0.0,>=8.2.0, but you have tenacity 9.0.0 which is incompatible.
Successfully installed PyJWT-2.10.1 aiofiles-24.1.0 aiolimiter-1.1.1 azure-common-1.1.28 azure-identity-1.19.0 azure-search-documents-11.5.2 azure-storage-blob-12.24.0 datashaper-0.0.49 devtools-0.12.2 environs-11.2.1 graphrag-0.3.6 json-repair-0.28.4 lancedb-0.13.0 msal-1.31.1 msal-extensions-1.2.0 py-1.11.0 pyaml-env-1.2.1 pyarrow-15.0.2 pylance-0.17.0 retry-0.9.2 tenacity-9.0.0 tiktoken-0.7.0
Reproduction steps
1. Go to '...'
2. Click on '....'
3. Scroll down to '....'
4. See error
Screenshots
Logs
No response
Browsers
Firefox
OS
Linux
Additional information
No response
And Docling error dependencys :
Installing collected packages: python-bidi, pyclipper, XlsxWriter, tifffile, Shapely, rtree, pypdfium2, opencv-python-headless, ninja, marko, lazy-loader, jsonref, jsonlines, imageio, docling-parse, scikit-image, python-pptx, typer, pydantic-settings, torchvision, docling-core, easyocr, docling-ibm-models, deepsearch-glm, docling
Attempting uninstall: typer
Found existing installation: typer 0.14.0
Uninstalling typer-0.14.0:
Successfully uninstalled typer-0.14.0
ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts.
gradio 4.39.0 requires aiofiles<24.0,>=22.0, but you have aiofiles 24.1.0 which is incompatible.
kotaemon 0.9.2 requires tenacity<8.3,>=8.2.3, but you have tenacity 9.0.0 which is incompatible.
kotaemon 0.9.2 requires umap-learn==0.5.5, but you have umap-learn 0.5.7 which is incompatible.
Successfully installed Shapely-2.0.6 XlsxWriter-3.2.0 deepsearch-glm-0.26.1 docling-2.8.1 docling-core-2.6.1 docling-ibm-models-2.0.7 docling-parse-2.1.2 easyocr-1.7.2 imageio-2.36.1 jsonlines-3.1.0 jsonref-1.1.0 lazy-loader-0.4 marko-2.1.2 ninja-1.11.1.2 opencv-python-headless-4.10.0.84 pyclipper-1.3.0.post6 pydantic-settings-2.6.1 pypdfium2-4.30.0 python-bidi-0.6.3 python-pptx-1.0.2 rtree-1.3.0 scikit-image-0.24.0 tifffile-2024.9.20 torchvision-0.20.1 typer-0.12.5
Similar problem... after running: pip install git+https://github.com/HKUDS/LightRAG.git Before that thewre was no problem after clean install only of: Installed conda, python 3.10. pip install -e "libs/kotaemon[all]" and pip install -e "libs/ktem"
It was running, but no ollama... so I wanted to install the needed LightRag... So end up having problem with hnswlib
pip show hnswlib Name: hnswlib Version: 0.8.0 Summary: hnswlib Home-page: https://github.com/yurymalkov/hnsw Location: /home/dragon/miniconda3/envs/kotaemon/lib/python3.10/site-packages Requires: numpy Required-by: lightrag-hku, nano-graphrag
Similar problem... after running: pip install git+https://github.com/HKUDS/LightRAG.git Before that thewre was no problem after clean install only of: Installed conda, python 3.10. pip install -e "libs/kotaemon[all]" and pip install -e "libs/ktem"
It was running, but no ollama... so I wanted to install the needed LightRag... So end up having problem with hnswlib
pip show hnswlib Name: hnswlib Version: 0.8.0 Summary: hnswlib Home-page: https://github.com/yurymalkov/hnsw Location: /home/dragon/miniconda3/envs/kotaemon/lib/python3.10/site-packages Requires: numpy Required-by: lightrag-hku, nano-graphrag
I tried again ... delete the environment and recreate again. But this time I installed LightRag first: so, I tried in reverse order : ... pip install git+https://github.com/HKUDS/LightRAG.git ... pip install -e "libs/kotaemon[all]" pip install -e "libs/ktem"
That worked (workaround) :) :)