Fix KeyError in S3BasedDocs when prefix path has no common prefixes

[mykola-pereyma]

Description

Fixes a KeyError: 'CommonPrefixes' crash in both S3DocDownloader.download() and S3ChunkDownloader.download() when the S3 prefix path is empty or has no sub-prefixes.

Problem

The S3 list_objects_v2 paginator omits the CommonPrefixes key entirely from the response when there are no common prefixes. Both downloaders accessed this key directly:

for source_doc_obj in source_doc_page['CommonPrefixes']  # KeyError if key absent

This crashes when:

• The S3 bucket/prefix is empty (no data uploaded yet)
• The collection path doesn't exist
• A paginator page has no sub-prefixes

Fix

Replace direct dict access with .get() and empty list default at both locations:

for source_doc_obj in source_doc_page.get('CommonPrefixes', [])

Tests

Added 4 unit tests covering:

• S3DocDownloader with empty S3 response (no CommonPrefixes key)
• S3ChunkDownloader with empty S3 response
• S3DocDownloader with mixed pages (some with prefixes, some without)
• S3ChunkDownloader with mixed pages

All 21 tests in test_s3_based_docs.py pass.

By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.

[oussamahansal]

lgtm