fix(github): reject garbage in _parse_date; consolidate date parsing

github.py _parse_date used naive string slicing (return iso_str[:10])
which accepted any 10+ character string as a "date." For input
"hello world" it returned "hello worl". Now delegates to
dates.parse_date() which validates the format and returns None for
non-dates.

Also migrated reddit.py and threads.py _parse_date to the shared
dates.parse_date(). Both previously reimplemented ISO-with-trailing-
offset handling (the .replace("Z", "+00:00") dance) and reddit.py
also had its own Unix timestamp branch. dates.parse_date() already
handles all of this, including the +0000 no-colon variant Reddit emits.

Preserved reddit.py's original falsy-check so 0 still returns None
(epoch 0 would otherwise parse as "1970-01-01", breaking an existing
test and changing long-standing behavior).

Added 4 new github tests for garbage rejection and offset variants.
All 1026 existing tests pass (15 pre-existing failures unchanged).

scripts/lib/github.py

@@ -17,7 +17,7 @@ import urllib.request
 from concurrent.futures import ThreadPoolExecutor, as_completed
 from typing import Any, Dict, List, Optional
 
-from . import log
+from . import dates, log
 from .query import extract_core_subject
 from .relevance import token_overlap_relevance
 
@@ -106,13 +106,14 @@ def _parse_repo_from_url(html_url: str) -> str:
 
 
 def _parse_date(iso_str: Optional[str]) -> Optional[str]:
-    """Extract YYYY-MM-DD from ISO 8601 datetime string."""
-    if not iso_str:
-        return None
-    try:
-        return iso_str[:10]
-    except (IndexError, TypeError):
-        return None
+    """Parse a GitHub ISO 8601 datetime string and return YYYY-MM-DD.
+
+    Returns None for non-date input. GitHub's API always emits ISO 8601
+    (e.g. "2026-02-26T16:00:00Z"), but we defer to dates.parse_date() so
+    garbage input gets rejected instead of silently sliced.
+    """
+    dt = dates.parse_date(iso_str)
+    return dt.strftime("%Y-%m-%d") if dt else None
 
 
 def _compute_relevance(

scripts/lib/reddit.py

@@ -12,7 +12,6 @@ import sys
 import time
 from collections import Counter
 from concurrent.futures import ThreadPoolExecutor, as_completed, wait as futures_wait
-from datetime import datetime, timezone
 from typing import Any, Dict, List, Optional, Set
 
 try:
@@ -28,7 +27,7 @@ def _first_of(*values, default=None):
             return v
     return default
 
-from . import http, log
+from . import dates, http, log
 
 SCRAPECREATORS_BASE = "https://api.scrapecreators.com/v1/reddit"
 
@@ -212,27 +211,16 @@ def _parse_date(value) -> Optional[str]:
 
     Global search returns ``created_at`` as an ISO string
     (e.g. "2018-05-03T01:09:17.620000+0000"); subreddit search returns
-    ``created_utc`` as a Unix timestamp.  Handle both.
+    ``created_utc`` as a Unix timestamp. dates.parse_date() handles both,
+    plus edge cases like Z suffix and +0000 (no colon) offset.
+
+    Falsy inputs (None, "", 0) return None, matching the original behavior
+    where a Unix timestamp of 0 meant "no date" rather than epoch 0.
     """
     if not value:
         return None
-    # ISO-8601 string (contains 'T' or '-')
-    if isinstance(value, str) and ("T" in value or "-" in value):
-        try:
-            # Strip trailing offset variations (+0000, Z) for fromisoformat
-            clean = value.replace("Z", "+00:00")
-            if clean.endswith("+0000"):
-                clean = clean[:-5] + "+00:00"
-            dt = datetime.fromisoformat(clean)
-            return dt.strftime("%Y-%m-%d")
-        except (ValueError, TypeError):
-            pass
-    # Unix timestamp (int or float or numeric string)
-    try:
-        dt = datetime.fromtimestamp(float(value), tz=timezone.utc)
-        return dt.strftime("%Y-%m-%d")
-    except (ValueError, TypeError, OSError):
-        return None
+    dt = dates.parse_date(str(value))
+    return dt.strftime("%Y-%m-%d") if dt else None
 
 
 def _extract_subreddit_name(value: Any) -> str:

scripts/lib/threads.py

@@ -9,10 +9,9 @@ API docs: https://scrapecreators.com/docs
 
 import math
 import re
-from datetime import datetime, timezone
 from typing import Any, Dict, List, Optional
 
-from . import http, log
+from . import dates, http, log
 from .relevance import token_overlap_relevance as _compute_relevance
 
 SCRAPECREATORS_BASE = "https://api.scrapecreators.com/v1/threads"
@@ -52,29 +51,17 @@ def _extract_core_subject(topic: str) -> str:
 def _parse_date(item: Dict[str, Any]) -> Optional[str]:
     """Parse date from Threads item to YYYY-MM-DD.
 
-    Tries common timestamp fields: taken_at (unix), created_at (ISO),
-    and falls back to any date-like string field.
+    Tries common timestamp fields in order: taken_at and create_time
+    (unix timestamps in Meta APIs), then created_at, published_at, and
+    date (ISO 8601 strings). dates.parse_date() handles both.
     """
-    # Unix timestamp (taken_at is common in Meta APIs)
-    for key in ("taken_at", "create_time"):
-        ts = item.get(key)
-        if ts:
-            try:
-                from . import dates
-                return dates.timestamp_to_date(int(ts))
-            except (ValueError, TypeError):
-                pass
-
-    # ISO 8601 string
-    for key in ("created_at", "published_at", "date"):
+    for key in ("taken_at", "create_time", "created_at", "published_at", "date"):
         val = item.get(key)
-        if val and isinstance(val, str):
-            try:
-                dt = datetime.fromisoformat(val.replace("Z", "+00:00"))
-                return dt.strftime("%Y-%m-%d")
-            except (ValueError, TypeError):
-                pass
-
+        if val is None:
+            continue
+        dt = dates.parse_date(str(val))
+        if dt:
+            return dt.strftime("%Y-%m-%d")
     return None