"""Cross-reference fresh findings against April baseline (master_worklist.csv).""" import csv import re from collections import defaultdict, Counter from urllib.parse import urlparse def load_csv(path): with open(path, encoding="utf-8") as f: return list(csv.DictReader(f)) def page_path(url): """Strip protocol/host, return path only (no trailing slash).""" p = urlparse(url).path if url.startswith("http") else url return p.rstrip("/") or "/" def norm_href(h): """Strip query/fragment for matching.""" return h.split("?", 1)[0].split("#", 1)[0].rstrip("/") or "/" def main(): master = load_csv("../master_worklist.csv") t016 = load_csv("t016_findings_2026_05_08.csv") t063 = load_csv("t063_findings_2026_05_08.csv") # Index April baseline by (page_path, link_href). Page_url in master may be path-only. april_index = defaultdict(list) for r in master: page = page_path(r["page_url"]) href = norm_href(r["current_href"]) april_index[(page, href)].append(r) print(f"April master_worklist rows: {len(master)}") print(f"April unique (page,href) keys: {len(april_index)}") print() # April status by task print("April status breakdown:") by_status = Counter((r["task"], r["execution_status"]) for r in master) for k, v in sorted(by_status.items()): print(f" {k}: {v}") print() # For each fresh finding, classify against April def classify(page, href): pagep = page_path(page) hrefn = norm_href(href) if (pagep, hrefn) in april_index: statuses = [r["execution_status"] for r in april_index[(pagep, hrefn)]] tasks = [r["task"] for r in april_index[(pagep, hrefn)]] return "carry-over", statuses, tasks return "new", [], [] # T016 classification t016_main = [r for r in t016 if r["in_main"] == "1" and r["in_chrome"] == "0"] t016_chrome = [r for r in t016 if r["in_chrome"] == "1"] t016_carry_main = [] t016_new_main = [] for r in t016_main: c, s, t = classify(r["page_url"], r["link_href"]) if c == "carry-over": t016_carry_main.append({**r, "april_status": ",".join(set(s)), "april_task": ",".join(set(t))}) else: t016_new_main.append(r) print(f"=== T016 (link-via-redirect) ===") print(f" Total findings (in_main): {len(t016_main)}") print(f" Carried over from April baseline: {len(t016_carry_main)}") print(f" New since April: {len(t016_new_main)}") print(f" Total findings (in_chrome / theme-level): {len(t016_chrome)}") print() # Status of carried-over if t016_carry_main: carry_status = Counter(r["april_status"] for r in t016_carry_main) print(f" Carried-over T016 by April status: {dict(carry_status)}") print() # T063 classification t063_main = [r for r in t063 if r["in_main"] == "1"] t063_main_fixable = [r for r in t063_main if r["fixable"] == "1"] t063_main_broken = [r for r in t063_main if r["fixable"] == "0"] t063_carry_fixable = [] t063_new_fixable = [] for r in t063_main_fixable: c, s, t = classify(r["page_url"], r["link_href"]) (t063_carry_fixable if c == "carry-over" else t063_new_fixable).append({**r, "april_status": ",".join(set(s))}) t063_carry_broken = [] t063_new_broken = [] for r in t063_main_broken: c, s, t = classify(r["page_url"], r["link_href"]) (t063_carry_broken if c == "carry-over" else t063_new_broken).append({**r, "april_status": ",".join(set(s))}) print(f"=== T063 (legacy /node/N) ===") print(f" Fixable (has redirect target):") print(f" Total: {len(t063_main_fixable)}") print(f" Carried over: {len(t063_carry_fixable)}") print(f" New: {len(t063_new_fixable)}") print(f" Broken (no redirect, deleted target):") print(f" Total: {len(t063_main_broken)}") print(f" Carried over: {len(t063_carry_broken)}") print(f" New: {len(t063_new_broken)}") print() if t063_carry_fixable: cs = Counter(r["april_status"] for r in t063_carry_fixable) print(f" Carried-over T063 fixable by April status: {dict(cs)}") if t063_carry_broken: cs = Counter(r["april_status"] for r in t063_carry_broken) print(f" Carried-over T063 broken by April status: {dict(cs)}") print() # Save annotated CSVs def save(rows, path, extra_cols=None): if not rows: with open(path, "w", encoding="utf-8") as f: f.write("(empty)\n") return cols = list(rows[0].keys()) with open(path, "w", encoding="utf-8", newline="") as f: w = csv.DictWriter(f, fieldnames=cols) w.writeheader() for r in rows: w.writerow(r) save(t016_carry_main, "t016_carry_over_main.csv") save(t016_new_main, "t016_new_since_april_main.csv") save(t016_chrome, "t016_theme_level.csv") save(t063_carry_fixable, "t063_carry_over_fixable.csv") save(t063_new_fixable, "t063_new_since_april_fixable.csv") save(t063_carry_broken, "t063_carry_over_broken.csv") save(t063_new_broken, "t063_new_since_april_broken.csv") # Final scorecard print("=" * 60) print("SCORECARD: April baseline vs May 8 reaudit") print("=" * 60) # T015 apr_t015_in_scope = sum(1 for r in master if r["task"] == "T015") print(f"\nT015 (redirect chains over 2 hops)") print(f" April: 15 in scope (12 to fix + 3 SKIPPED lang variants)") print(f" May 8: 3 chains remain — all 3 are the SKIPPED IT/ES/FR variants") print(f" Net: 12/12 fixable items resolved (100%). 0 new chains.") # T016 apr_t016 = [r for r in master if r["task"] == "T016"] apr_t016_done = [r for r in apr_t016 if r["execution_status"] == "DONE"] apr_t016_skipped = [r for r in apr_t016 if r["execution_status"] in ("SKIPPED",)] apr_t016_pending = [r for r in apr_t016 if r["execution_status"] in ("", "MANUAL-REQUIRED", "FAILED", "NOT-FOUND")] print(f"\nT016 (links via redirect)") print(f" April: 183 in scope") print(f" DONE in April: {len(apr_t016_done)}") print(f" Pending/Failed/NotFound: {len(apr_t016_pending)}") print(f" May 8: {len(t016_main)} body-content findings ({len(t016_carry_main)} carry-over + {len(t016_new_main)} new)") print(f" Plus: {len(t016_chrome)} theme-level (header/footer menu) — single fix in theme config") # T063 apr_t063 = [r for r in master if r["task"] == "T063"] apr_t063_done = [r for r in apr_t063 if r["execution_status"] == "DONE"] apr_brokenlink = [r for r in master if r["task"] == "BROKEN-LINK"] apr_brokenlink_done = [r for r in apr_brokenlink if r["execution_status"] == "DONE"] print(f"\nT063 (legacy /node/N links)") print(f" April: 175 fixable + 39 broken (master_worklist sum incl ones discovered during exec)") print(f" DONE in April (T063): {len(apr_t063_done)}") print(f" DONE in April (BROKEN-LINK): {len(apr_brokenlink_done)}") print(f" May 8: {len(t063_main_fixable)} fixable + {len(t063_main_broken)} broken = {len(t063_main)} total") print(f" Carry-over: {len(t063_carry_fixable)} fixable + {len(t063_carry_broken)} broken") print(f" New since April: {len(t063_new_fixable)} fixable + {len(t063_new_broken)} broken") if __name__ == "__main__": main()