search

mneedham / LearnDataWithMark

Code and scripts behind the @LearnDataWithMark YouTube channel
https://learndatawithmark.com
108 stars 32 forks source link

Issue in the code for Qdrant embedding for RAG - metadata #61

Open arkamaldeen opened 5 months ago

arkamaldeen commented 5 months ago

I was going through your video and referring to the GitHub repository at the same time. The video was very helpful.

I noticed that you managed to get 1225 embedded documents with only 51 JSON documents, which was quite confusing!

I suspect that the line for line in doc['body'] was looping through the characters, causing this discrepancy.

This is my first time trying to create an issue on GitHub, so please pardon me for any mistakes. By the way, the video was very useful. Keep up the good work!

mneedham commented 5 months ago

@arkamaldeen can you point me to which one you were watching? I've done 2 or 3 videos about Qdrant, so trying to figure out which one is wrong!

arkamaldeen commented 5 months ago

The below notebook: metadata/notebooks/User-Provided-Metadata-Tutorial.ipynb

Yt link: https://www.youtube.com/watch?v=qcn7YAJfCeE

mneedham commented 3 months ago

@arkamaldeen sorry, I didn't see you'd replied. I did one document per line in the JSON document, that's how we got so many documents.