"""Diagnostic for the keyword-mode misses (not part of the main table, which keeps defaults).
Runs the six questions in keyword mode with the default drop_tokens_threshold (1) and with 50.
Prints `found` (matching chunks) and the rank of the expected page among the first 5 pages."""
from common import hit_rank, pages, questions
from ts_query import search, urls

print("keyword mode, per_page 50 | found = matching chunks | rank = expected page among first 5 pages (- = miss)")
for thr in (1, 50):
    print(f"\ndrop_tokens_threshold {thr}" + (" (default)" if thr == 1 else ""))
    ok = 0
    for q in questions():
        res = search(q["question"], extra={"drop_tokens_threshold": thr})
        r = hit_rank(urls(res), q["expected_url_contains"])
        ok += r is not None
        print(f"  {q['question'][:62]:<64} found {res['found']:>3} rank {r or '-'}")
    print(f"  expected page in top 5: {ok}/6")

print("\ntypo case: the correct spelling vs the misspelling, default threshold 1")
for text in ("cancel crawl job", "cancle crawl job"):
    res = search(text)
    print(f"  '{text}': found {res['found']} | first pages {pages(urls(res))[:3]}")