"""Re-fetch the same 1,563 llms.txt files, six weeks after the August run.

This is a NEW dated observation, not part of the published deposit. It exists so
outreach to the directory maintainers carries current numbers rather than a
six-week-old snapshot.

It reuses crawl.py's own fetch and classification functions unchanged, so the
two runs are directly comparable. The frame is rebuilt from the August run's
own src field, which is the same 1,563 URLs.
"""
import concurrent.futures as cf
import json, sys, time

sys.path.insert(0, "/root/research/llms-txt/raw")
import crawl  # same UA, same timeout, same is_html and link-parsing rules

RAW = "/root/research/llms-txt/raw/phase_a.jsonl"
OUT = "/root/research/llms-txt/phase_a_2026-09-20.jsonl"

aug = [json.loads(l) for l in open(RAW)]
urls = [r["src"] for r in aug]
print(f"re-fetching {len(urls)} llms.txt files", flush=True)
print(f"user agent: {crawl.UA}\n", flush=True)

t0 = time.time()
rows = []
with cf.ThreadPoolExecutor(48) as ex:
    for i, r in enumerate(ex.map(crawl.phase_a, urls), 1):
        rows.append({k: v for k, v in r.items() if k != "links"})
        if i % 300 == 0:
            print(f"  {i}/{len(urls)}  ({time.time()-t0:.0f}s)", flush=True)

with open(OUT, "w") as f:
    for r in rows:
        f.write(json.dumps(r) + "\n")

def healthy(r):
    return r["status"] == 200 and not r["is_html"]

ha, hn = sum(map(healthy, aug)), sum(map(healthy, rows))
print(f"\ndone in {time.time()-t0:.0f}s -> {OUT}")
print(f"  2026-08-05  healthy {ha}/{len(aug)}  broken {len(aug)-ha}  ({100*(len(aug)-ha)/len(aug):.1f}%)")
print(f"  2026-09-20  healthy {hn}/{len(rows)}  broken {len(rows)-hn}  ({100*(len(rows)-hn)/len(rows):.1f}%)")