Add shared HTTP retry adapter for sub-indexers

[YoshihitoAso]

📌 Description

Improve resilience when fetching public information lists from S3 URLs that temporarily return HTTP 503 responses.

✅ Related Issues

None

🔄 Changes

• Added a shared HTTP retry adapter for public information fetches.
• Retry HTTP 500, 502, 503, and 504 responses for GET requests.
• Retry configuration:
  • total=3: retry up to three times after the initial request, for a maximum of four requests in total.
  • backoff_factor=1.0: use exponential backoff between retries, resulting in approximately 0, 2, and 4 seconds of delay.
  • backoff_jitter=0.5: add a small random delay to avoid simultaneous retries from multiple processes.
  • respect_retry_after_header=True: honor the wait time specified by the S3 response.
  • raise_on_status=False: return the final HTTP response after retries are exhausted so the indexer can handle it consistently.
  • status_forcelist=(500, 502, 503, 504): retry temporary server-side errors.
  • allowed_methods=("GET",): limit retries to idempotent GET requests.
• Applied the retry policy to Company List, Token List, and Public Account List indexers.
• Preserve ERROR-level exception logging when all retry attempts fail.
• Added test coverage for the retry configuration and failed request logging.

📌 Checklist

• I have added tests where necessary.
• I have updated the documentation where necessary.

[github-actions]

Coverage Report •

File	Stmts	Miss	Branch	BrPart	Cover	Missing
batch/lib
http.py	4	0	0	0	100%
batch/sub_indexers
indexer_Company_List.py	72	7	12	2	89%	52->53, 53–54, 95->105, 105–106, 109–111
indexer_PublicInfo_PublicAccountList.py	75	4	12	2	93%	52->53, 53–54, 104->105, 105, 108
indexer_PublicInfo_TokenList.py	67	5	8	1	92%	53->54, 54–55, 99–101
tests/batch/sub_indexers
indexer_Company_List_test.py	558	0	0	0	100%
TOTAL	37460	1470	3524	647	95%

Tests	Skipped	Failures	Errors	Time
1281	0 💤	0 ❌	0 🔥	9m 30s ⏱️