error: no module named "injector"

[tvogt]

trying to get privategpt to run and running into this error no matter what I try:

Traceback (most recent call last):
  File "<frozen runpy>", line 198, in _run_module_as_main
  File "<frozen runpy>", line 88, in _run_code
  File "[redacted]/private-gpt/private_gpt/__main__.py", line 5, in <module>
    from private_gpt.main import app
  File "[redacted]/private-gpt/private_gpt/main.py", line 3, in <module>
    from private_gpt.di import global_injector
  File "[redacted]/private-gpt/private_gpt/di.py", line 1, in <module>
    from injector import Injector
ModuleNotFoundError: No module named 'injector'
make: *** [run] Error 1

I've been following the install instructions. I'm on MacOS 14.4 using Python 3.11.8. My precise command sequence was:

• ollama pull mistral
• ollama pull nomic-embed-text
• ollama serve
• git clone https://github.com/zylon-ai/private-gpt.git
• cd private-gpt
• python3.11 -m venv venv
• . ./venv/bin/activate
• pip3.11 install poetry
• poetry install --extras "ui llms-ollama embeddings-ollama vector-stores-qdrant"
• PGPT_PROFILES=ollama make run

the result is the above error. All steps prior to the last one complete without errors, and ollama runs locally just fine, the model is loaded (I can chat with it), etc.

and

[tvogt]

After some more searching and fiddling, I came across issue 1378 - https://github.com/zylon-ai/private-gpt/issues/1378

Checking back, I noticed that the poetry install step indeed throws an error with ffmpy, though it then continues and completes without showing errors at the bottom.

Applying the fix from issue 1378's last comment:

pip install -U build==0.10.0
pip install -U ffmpy==0.3.1

before calling poetry install works and I now have privateGPT running.

Someone more familiar with pip and poetry should check this dependency issue. Apparently ffmpy is now 0.3.2 but the poetry.lock is set to 0.3.1 which poetry can't find?