fix: Add try-except and row counter to ReadFileBlobFunction for robust CSV parsing

[the-damian]

Our blob pipeline is cool for watching CSVs, but it bombs on malformed files (e.g., bad UTF-8 or missing headers)—real data uploads are messy! Wrap the reader in try-except to catch errors (csv.Error, UnicodeDecodeError), log them meaningfully, and count/log successful rows. Optionally, return a summary dict (e.g., {"processed": 5, "errors": 0}) via logging. This ensures our ETL starter doesn't halt on junk, perfect for gig-economy data flows.