diagnostics/backfill_submitted_by.py

@@ -79,16 +79,12 @@ def main():
                     fm = sfm
                     break
 
-        # `submitted_by` is stored as a canonical handle (lowercase, no @, no
-        # "(self-directed)" / "(reweave)" suffix). Read consumers normalize via
-        # attribution.normalize_handle, so writing decorated strings produces
-        # downstream 404s on /contributors/{handle} (livingip-web timeline).
         if fm:
             proposed_by = fm.get("proposed_by")
             intake_tier = fm.get("intake_tier")
 
             if proposed_by:
-                contributor = proposed_by.strip().strip('"').strip("'").lower().lstrip("@")
+                contributor = proposed_by.strip().strip('"').strip("'")
             elif intake_tier == "research-task":
                 # Derive agent from branch prefix
                 prefix = branch.split("/", 1)[0] if "/" in branch else "unknown"
@@ -98,12 +94,13 @@ def main():
                     "clay": "clay", "astra": "astra", "leo": "leo",
                     "reweave": "pipeline",
                 }
-                contributor = agent_map.get(prefix, prefix)
+                agent = agent_map.get(prefix, prefix)
+                contributor = f"{agent} (self-directed)"
             elif intake_tier == "directed":
-                contributor = "m3taversal"
+                contributor = "@m3taversal"
             else:
-                # Default: if source exists but no proposed_by, operator submitted it.
+                # Default: if source exists but no proposed_by, it was Cory's submission
-                contributor = "m3taversal"
+                contributor = "@m3taversal"
 
             if contributor:
                 conn.execute(
@@ -117,19 +114,19 @@ def main():
                 agent = branch.split("/", 1)[0]
                 conn.execute(
                     "UPDATE prs SET submitted_by = ? WHERE number = ?",
-                    (agent, pr["number"]),
+                    (f"{agent} (self-directed)", pr["number"]),
                 )
                 updated += 1
             elif branch.startswith("reweave/"):
                 conn.execute(
-                    "UPDATE prs SET submitted_by = 'pipeline' WHERE number = ?",
+                    "UPDATE prs SET submitted_by = 'pipeline (reweave)' WHERE number = ?",
                     (pr["number"],),
                 )
                 updated += 1
             else:
-                # Everything else (extract/, ingestion/, unknown) → operator directed it
+                # Everything else (extract/, ingestion/, unknown) → Cory directed it
                 conn.execute(
-                    "UPDATE prs SET submitted_by = 'm3taversal' WHERE number = ?",
+                    "UPDATE prs SET submitted_by = '@m3taversal' WHERE number = ?",
                     (pr["number"],),
                 )
                 updated += 1

lib/extract.py

@@ -32,7 +32,6 @@ from datetime import date
 from pathlib import Path
 
 from . import config
-from .attribution import normalize_handle
 from .costs import record_usage
 from .db import classify_source_channel
 from .domains import agent_for_domain
@@ -684,25 +683,16 @@ async def _extract_one_source(
         logger.info("PR #%d created for %s (%d claims, %d entities)", pr_num, source_file, len(claim_files), len(entity_files))
 
         # Store contributor attribution: who submitted this source?
-        # Priority: proposed_by field → intake_tier inference → operator default.
+        # Priority: proposed_by field → intake_tier inference → "unknown"
-        # NB: `submitted_by` is a CANONICAL HANDLE — lowercase, no @, no
-        # trailing "(self-directed)" decorator. The "self-directed" signal is
-        # already carried by intake_tier == "research-task" + the prs.agent
-        # column; persisting it here as a string suffix produced decorated
-        # values like "Vida (self-directed)" that broke /contributors/{handle}
-        # lookups downstream (livingip-web timeline → 404). Read consumers
-        # (lib/contributor.insert_contribution_event, scripts/scoring_digest,
-        # diagnostics/activity_feed_api) all normalize via attribution.normalize_handle
-        # anyway, so writing the canonical form is the source-of-truth fix.
         if proposed_by:
-            contributor = normalize_handle(proposed_by, conn=conn)
+            contributor = proposed_by.strip().strip('"').strip("'")
         elif intake_tier == "research-task":
-            contributor = normalize_handle(agent_name, conn=conn)
+            contributor = f"{agent_name} (self-directed)"
         elif intake_tier == "directed":
-            contributor = "m3taversal"
+            contributor = "@m3taversal"
         else:
-            # Default: if no proposed_by and not a research task, operator submitted it.
+            # Default: if no proposed_by and not a research task, Cory submitted it
-            contributor = "m3taversal"
+            contributor = "@m3taversal"
 
         # Build pipe-separated claim titles for the description field
         claim_titles = " | ".join(

scripts/backfill-research-session-attribution.py

@@ -17,9 +17,7 @@ Identification:
   PR whose `source_path` matches one of those filenames is research-derived.
 
 Touch list (per matched PR):
-  1. UPDATE prs SET submitted_by = '<agent>'   (canonical handle, lowercase, no
+  1. UPDATE prs SET submitted_by = '<agent> (self-directed)'
-       trailing "(self-directed)" suffix — see lib/extract.py and
-       diagnostics/activity_feed_api.py for why decorators leak into 404s)
   2. DELETE FROM contribution_events
        WHERE handle='m3taversal' AND role='author' AND pr_number=?
   3. INSERT OR IGNORE INTO contribution_events with handle=<agent>,
@@ -171,7 +169,7 @@ def apply_backfill(conn: sqlite3.Connection, matches: list[dict], dry_run: bool)
 
             if dry_run:
                 logger.info(
-                    "would update pr=%d submitted_by '%s' → '%s' "
+                    "would update pr=%d submitted_by '%s' → '%s (self-directed)' "
                     "[m3ta_event=%s, agent_event=%s]",
                     pr, m["current_submitted_by"], agent,
                     old_event_exists, new_event_exists,
@@ -183,10 +181,10 @@ def apply_backfill(conn: sqlite3.Connection, matches: list[dict], dry_run: bool)
                     counters["events_to_insert"] += 1
                 continue
 
-            # 1. UPDATE prs.submitted_by — canonical handle only.
+            # 1. UPDATE prs.submitted_by
             conn.execute(
                 "UPDATE prs SET submitted_by = ? WHERE number = ?",
-                (agent, pr),
+                (f"{agent} (self-directed)", pr),
             )
             counters["prs"] += 1
 

scripts/normalize-submitted-by.py

@@ -1,114 +0,0 @@
-#!/usr/bin/env python3
-"""One-time backfill: canonicalize prs.submitted_by and sources.submitted_by.
-
-Strips legacy decorators ("(self-directed)", "(reweave)"), lowercases, drops
-the @ prefix. After this runs, every value matches the contract documented
-on diagnostics/activity_feed_api.py::_normalize_contributor — and the
-companion read-side fix becomes redundant defense-in-depth instead of
-load-bearing.
-
-Defaults to --dry-run. Pass --apply to commit.
-
-Usage:
-    python3 normalize-submitted-by.py --dry-run
-    python3 normalize-submitted-by.py --apply
-"""
-
-import argparse
-import os
-import re
-import sqlite3
-import sys
-from collections import Counter
-
-DEFAULT_DB = os.environ.get("PIPELINE_DB", "/opt/teleo-eval/pipeline/pipeline.db")
-
-# Valid handle: lowercase alphanum + _-, 1-39 chars (matches GitHub rules,
-# same as pipeline/lib/attribution._HANDLE_RE). Anything with parens, spaces,
-# or uppercase needs canonicalization.
-_TRAILING_PAREN_RE = re.compile(r"\s*\([^)]*\)\s*$")
-_HANDLE_RE = re.compile(r"^[a-z0-9][a-z0-9_-]{0,38}$")
-
-
-def canonicalize(raw):
-    if raw is None:
-        return None
-    h = raw.strip().lower().lstrip("@")
-    h = _TRAILING_PAREN_RE.sub("", h).strip()
-    return h or None
-
-
-def normalize_table(conn, table, dry_run):
-    cur = conn.execute(
-        f"SELECT rowid, submitted_by FROM {table} WHERE submitted_by IS NOT NULL"
-    )
-    changes = []
-    for row in cur.fetchall():
-        old = row[1]
-        new = canonicalize(old)
-        if new != old:
-            changes.append((row[0], old, new))
-
-    print(f"\n{table}: {len(changes)} rows need normalization")
-    if not changes:
-        return 0
-
-    # Distribution preview
-    from_to = Counter((old, new) for _, old, new in changes)
-    for (old, new), count in from_to.most_common(15):
-        print(f"  {count:>5}  {old!r:40} -> {new!r}")
-    if len(from_to) > 15:
-        print(f"  ... ({len(from_to) - 15} more distinct mappings)")
-
-    # Sanity: every result is a valid handle (no garbage falls through).
-    invalid = [(rowid, old, new) for rowid, old, new in changes
-               if new is not None and not _HANDLE_RE.match(new)]
-    if invalid:
-        print(f"\n  WARNING: {len(invalid)} rows would normalize to invalid handles:")
-        for rowid, old, new in invalid[:10]:
-            print(f"    rowid={rowid} {old!r} -> {new!r}")
-        print("  These rows will be SKIPPED (left as-is). Inspect manually.")
-
-    valid_changes = [(rowid, old, new) for rowid, old, new in changes
-                     if new is None or _HANDLE_RE.match(new)]
-
-    if dry_run:
-        print(f"  [dry-run] would update {len(valid_changes)} rows in {table}")
-        return len(valid_changes)
-
-    for rowid, _, new in valid_changes:
-        conn.execute(
-            f"UPDATE {table} SET submitted_by = ? WHERE rowid = ?",
-            (new, rowid),
-        )
-    print(f"  updated {len(valid_changes)} rows in {table}")
-    return len(valid_changes)
-
-
-def main():
-    ap = argparse.ArgumentParser()
-    ap.add_argument("--db", default=DEFAULT_DB)
-    ap.add_argument("--apply", action="store_true", help="Commit changes (default is dry-run)")
-    ap.add_argument("--dry-run", action="store_true", help="Preview only (default)")
-    args = ap.parse_args()
-
-    dry_run = not args.apply
-    print(f"DB: {args.db}")
-    print(f"Mode: {'DRY-RUN' if dry_run else 'APPLY'}")
-
-    conn = sqlite3.connect(args.db, timeout=30)
-    try:
-        total = 0
-        total += normalize_table(conn, "prs", dry_run)
-        total += normalize_table(conn, "sources", dry_run)
-        if not dry_run:
-            conn.commit()
-            print(f"\nCommitted. Total rows updated: {total}")
-        else:
-            print(f"\nDry-run complete. Run with --apply to commit ({total} rows pending).")
-    finally:
-        conn.close()
-
-
-if __name__ == "__main__":
-    main()