slavingia / askmybook Goto Github PK

Should make this significantly cheaper and faster

https://beta.openai.com/docs/guides/embeddings/

[OKAY] Loaded ENV .env File as KEY=VALUE Format
2:24:44 PM web.1 | [2023-01-16 14:24:44 -0700] [226789] [INFO] Starting gunicorn 20.1.0
2:24:44 PM web.1 | [2023-01-16 14:24:44 -0700] [226789] [INFO] Listening at: http://0.0.0.0:33555 (226789)
2:24:44 PM web.1 | [2023-01-16 14:24:44 -0700] [226789] [INFO] Using worker: sync
2:24:44 PM web.1 | [2023-01-16 14:24:44 -0700] [226790] [INFO] Booting worker with pid: 226790
2:24:48 PM web.1 | Internal Server Error: /ask
2:24:48 PM web.1 | Traceback (most recent call last):
2:24:48 PM web.1 | File "/home/loganjones/.local/lib/python3.11/site-packages/django/core/handlers/exception.py", line 55, in inner
2:24:48 PM web.1 | response = get_response(request)
2:24:48 PM web.1 | ^^^^^^^^^^^^^^^^^^^^^
2:24:48 PM web.1 | File "/home/loganjones/.local/lib/python3.11/site-packages/django/core/handlers/base.py", line 197, in _get_response
2:24:48 PM web.1 | response = wrapped_callback(request, *callback_args, **callback_kwargs)
2:24:48 PM web.1 | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2:24:48 PM web.1 | File "/home/loganjones/.local/lib/python3.11/site-packages/django/views/decorators/csrf.py", line 54, in wrapped_view
2:24:48 PM web.1 | return view_func(*args, **kwargs)
2:24:48 PM web.1 | ^^^^^^^^^^^^^^^^^^^^^^^^^^
2:24:48 PM web.1 | File "/home/loganjones/CMPUT/QueryMyBook/hello/views.py", line 167, in ask
2:24:48 PM web.1 | document_embeddings = load_embeddings('book.pdf.embeddings.csv')
2:24:48 PM web.1 | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2:24:48 PM web.1 | File "/home/loganjones/CMPUT/QueryMyBook/hello/views.py", line 84, in load_embeddings
2:24:48 PM web.1 | max_dim = max([int(c) for c in df.columns if c != "title"])
2:24:48 PM web.1 | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
2:24:48 PM web.1 | File "/home/loganjones/CMPUT/QueryMyBook/hello/views.py", line 84, in
2:24:48 PM web.1 | max_dim = max([int(c) for c in df.columns if c != "title"])
2:24:48 PM web.1 | ^^^^^^
2:24:48 PM web.1 | ValueError: invalid literal for int() with base 10: 'content'

Hiya,

Not my usual area of extpertise but the demo was so impressive I had to give it a go.

I got everything up and running however I can't get any answers out of it, when I sumbit the request I get a 500 error and visiting the /ask page does the same.

Any suggestions, or where I can look for app based debug logs?

Hi, this is a great project! I noticed OpenAI has recently published embedding model v2 ada-002 which is "cheaper, faster, better and simpler." https://openai.com/blog/new-and-improved-embedding-model/. One suggestion update the project to switch to using this newest one.