Upload github_storage.py with huggingface_hub

github_storage.py

@@ -2,9 +2,11 @@ from __future__ import annotations
 
 import base64
 import csv
+import hashlib
 import io
 import json
 import logging
+import time
 from typing import Any
 
 import requests
@@ -22,6 +24,13 @@ FIELDNAMES = [
     "timestamp",
 ]
 
+MAX_RETRIES = 3
+RETRY_BACKOFF = 2.0
+
+
+def _safe_reviewer_id(reviewer_id: str) -> str:
+    return hashlib.sha256(reviewer_id.encode()).hexdigest()[:12]
+
 
 def _get_config() -> dict[str, str]:
     import os
@@ -59,7 +68,7 @@ def _get_file(repo: str, path: str, token: str, branch: str) -> tuple[str | None
     return content, data["sha"]
 
 
-def _put_file(
+def _put_file_with_retry(
     repo: str,
     path: str,
     content: str,
@@ -76,8 +85,25 @@ def _put_file(
     }
     if sha:
         body["sha"] = sha
-    resp = requests.put(url, headers=_headers(token), json=body, timeout=15)
-    resp.raise_for_status()
+
+    last_error: Exception | None = None
+    for attempt in range(MAX_RETRIES):
+        try:
+            resp = requests.put(url, headers=_headers(token), json=body, timeout=15)
+            if resp.status_code == 409:
+                _, fresh_sha = _get_file(repo, path, token, branch)
+                if fresh_sha:
+                    body["sha"] = fresh_sha
+                last_error = requests.HTTPError(response=resp)
+                time.sleep(RETRY_BACKOFF ** attempt)
+                continue
+            resp.raise_for_status()
+            return
+        except requests.RequestException as e:
+            last_error = e
+            if attempt < MAX_RETRIES - 1:
+                time.sleep(RETRY_BACKOFF ** attempt)
+    raise last_error  # type: ignore[misc]
 
 
 def _ensure_branch(repo: str, token: str, branch: str) -> None:
@@ -119,7 +145,8 @@ def load_pairs() -> list[dict]:
 
 def load_existing_reviews(reviewer_id: str) -> dict[int, dict]:
     cfg = _get_config()
-    path = f"{cfg['data_dir']}/reviews_{reviewer_id}.csv"
+    safe_id = _safe_reviewer_id(reviewer_id)
+    path = f"{cfg['data_dir']}/reviews_{safe_id}.csv"
     content, _ = _get_file(cfg["repo"], path, cfg["token"], cfg["branch"])
     if content is None:
         return {}
@@ -134,7 +161,8 @@ def save_review(reviewer_id: str, pair_index: int, review: dict) -> None:
     cfg = _get_config()
     _ensure_branch(cfg["repo"], cfg["token"], cfg["branch"])
 
-    path = f"{cfg['data_dir']}/reviews_{reviewer_id}.csv"
+    safe_id = _safe_reviewer_id(reviewer_id)
+    path = f"{cfg['data_dir']}/reviews_{safe_id}.csv"
     content, sha = _get_file(cfg["repo"], path, cfg["token"], cfg["branch"])
 
     existing: dict[int, dict] = {}
@@ -151,20 +179,18 @@ def save_review(reviewer_id: str, pair_index: int, review: dict) -> None:
     for idx in sorted(existing.keys()):
         writer.writerow(existing[idx])
 
-    _put_file(
+    _put_file_with_retry(
         cfg["repo"],
         path,
         output.getvalue(),
         cfg["token"],
         cfg["branch"],
         sha,
-        f"Review by {reviewer_id}: pair {pair_index}",
+        f"Review by {safe_id}: pair {pair_index}",
     )
 
 
 def append_audit_log(reviewer_id: str, pair_index: int, action: str) -> None:
-    import time
-
     cfg = _get_config()
     _ensure_branch(cfg["repo"], cfg["token"], cfg["branch"])
 
@@ -175,15 +201,16 @@ def append_audit_log(reviewer_id: str, pair_index: int, action: str) -> None:
         content = "timestamp,reviewer_id,pair_index,action\n"
         sha = None
 
+    safe_id = _safe_reviewer_id(reviewer_id)
     timestamp = time.strftime("%Y-%m-%d %H:%M:%S UTC", time.gmtime())
-    content += f"{timestamp},{reviewer_id},{pair_index},{action}\n"
+    content += f"{timestamp},{safe_id},{pair_index},{action}\n"
 
-    _put_file(
+    _put_file_with_retry(
         cfg["repo"],
         path,
         content,
         cfg["token"],
         cfg["branch"],
         sha,
-        f"Audit: {action} by {reviewer_id}",
+        f"Audit: {action} by {safe_id}",
     )