Feat/raae 9/repo upgrades #23
Conversation
| // Use IntelliSense to learn about possible attributes. | ||
| // Hover to view descriptions of existing attributes. | ||
| // For more information, visit: https://go.microsoft.com/fwlink/?linkid=830387 | ||
| "version": "0.2.0", |
There was a problem hiding this comment.
This is the setup for vscode debugger. I like to check these in in case people want it for development purposes since it's a reference architecture.
| RUN npm run build | ||
|
|
||
| FROM python:3.8-slim-buster AS ApiImage | ||
| FROM python:3.11-slim-buster AS ApiImage |
There was a problem hiding this comment.
@tylerhutcherson slim worked fine for this one so kept the smaller sizes
| vector_distance: float | ||
| similarity_score: float | ||
|
|
||
| def __init__(self, *args, **kwargs): |
There was a problem hiding this comment.
similarity score is derived from vector_distance so we can let pydantic handle the creation as part of making the object
| product_vectors = json.load(f) | ||
| except FileNotFoundError: | ||
| print("File not found, reading from S3") | ||
| product_vectors = read_from_s3() |
There was a problem hiding this comment.
@tylerhutcherson I think I'd like to simplify this flow. Either load from local volume or pull from S3. If you pull from S3 I don't think we need to write that data because it's only needed to populate the db and causes bloat in the docker containers.
I could add an additional script to pull this data into local for those developing locally but I think this is cleaner
tylerhutcherson
left a comment
tylerhutcherson
left a comment
There was a problem hiding this comment.
Looking great. left a few cleanup comments, but pretty awesome
justin-cechmanek
left a comment
justin-cechmanek
left a comment
There was a problem hiding this comment.
Just the one comment. Lots of great work here
Primary changes: