Minor dependency issues and a mismatch in the GraphRAG Complete notebook

[MoktarEls]

Hello,

First of all, thank you very much for this project, the work you've done is truly impressive and very well polished. We really appreciate the effort put into the documentation and the cookbook; it helps a lot.

During our usage, we encountered two issues that we would like to report :

1. Repository cloning and dependency issues

We ran into several dependency problems when trying to use the PyPI package (notably circular dependencies between some classes). We managed to work around this by installing directly from the GitHub repository:

pip install git+https://github.com/Hawksight-AI/semantica.git@main

However, even with this approach, we still faced two issues:

• protobuf and grpcio version incompatibility
  We resolved this by explicitly installing compatible versions:

pip install protobuf==4.25.3 grpcio==1.67.1

• Missing dependencies
  GitPython and chardet packages were not installed by default, though this was easy to fix manually

pip install GitPython
pip install chardet

2. GraphRAG Complete notebook : File Ingestion phase

The notebook is very useful and comprehensive overall. However, in Phase 1 (File Ingestion), the code no longer seems to match the current version.

Specifically, FileObject no longer has a text attribute, which causes an issue during compilation:

for file_path in local_files:
    try:
        if os.path.exists(file_path):
            print(f"  Processing: {file_path}")
            file_data = file_ingestor.ingest_file(file_path)
            # ingest_file returns a single FileObject, not a list
            if hasattr(file_data, 'text') and file_data.text:
                all_content.append(file_data.text)
            elif hasattr(file_data, 'content') and file_data.content:
                # If content is bytes, decode it
...

We worked around this by removing the first if condition, but we’re not sure whether:

• the notebook was written for an older version of the library, or
• we made a mistake during setup.

Any clarification would be greatly appreciated.
Once again, congratulations on the project, the work is excellent, and we fully support it.

[KaifAhmad1]

Hello @MoktarEls ,

Thank you for the kind words and for reporting these issues! We’ve addressed them in the latest update:

• Dependencies : We have pinned protobuf==4.25.8 (addressing a security vulnerability) and grpcio==1.67.1 in pyproject.toml . We also added GitPython and chardet as default dependencies.
• GraphRAG Alignment : The FileObject in file_ingestor.py has been verified to include the text property (which handles decoding). The notebook code is now fully aligned with the library.
• Orchestrator : Improved lazy initialization and configuration normalization for better stability.
  You can test these fixes by installing from the utils branch: pip install git+https://github.com/Hawksight-AI/semantica.git@main

We appreciate your support