Conversation
…l_dataframe.ipynb
|
Check out this pull request on See visual diffs & provide feedback on Jupyter Notebooks. Powered by ReviewNB |
product-auto-label
bot
added
size: xl
| @@ -2,7 +2,7 @@ | |||
| "cells": [ | |||
There was a problem hiding this comment.
Line #7. # df = bpd.read_gbq_object_table("", name="blob_col")
this can remove too.
Reply via ReviewNB
| @@ -2,7 +2,7 @@ | |||
| "cells": [ | |||
There was a problem hiding this comment.
Line #2. display_blob(df_image[df_image["author"] == "alice"]["image"])
We still keep preview in _repr_html_, don't need this.
Reply via ReviewNB
| @@ -2,7 +2,7 @@ | |||
| "cells": [ | |||
There was a problem hiding this comment.
Line #22. try:
Now the functions only serve the purpose as examples. We probably don't want to make them too complex. We can remove those try-catch and retry logics, only keep the core part.
Also we can remove other images functions, maybe just keep one is enough.
Reply via ReviewNB
There was a problem hiding this comment.
Simplified all transformation functions (image_blur, pdf_extract, etc.) by removing retry logic and try-except blocks, while keeping requests.Session() for connection stability. Removed image_resize and image_normalize to focus on the core example
product-auto-label
bot
added
size: l
| @@ -83,7 +83,7 @@ | |||
| }, | |||
There was a problem hiding this comment.
Line #13. src_rt: str, dst_rt: str, ext: str, kx: int, ky: int, verbose: bool
lets remove the ext and verbose to keep the function simple. Just use png or jpeg whichever is fine.
Reply via ReviewNB
| @@ -83,7 +83,7 @@ | |||
| }, | |||
There was a problem hiding this comment.
Line #22. session = requests.Session()
session is for retry and timeout, we can remove too to keep the function simple
Reply via ReviewNB
2 participants
This PR updates notebooks/multimodal/multimodal_dataframe.ipynb to demonstrate image modifications using custom BigQuery Python UDFs with the opencv library.
verified at: screen/BfesAowVQWEGsCf
Fixes #<478952827> 🦕