Spaces:
Running
Running
| """Self-check for the public widget. | |
| The important one is test_unpublished_content_never_leaks: this endpoint is | |
| unauthenticated, so the state/access filter is a trust boundary, not a detail. | |
| Run: python3 test_answer.py | |
| """ | |
| import answer | |
| import mesh | |
| TERMS = ["lunch", "price"] | |
| def test_keywords_drop_question_noise(): | |
| assert answer.keywords("How much is the student lunch?") == ["student", "lunch"] | |
| assert answer.keywords("Who is the superintendent") == ["superintendent"] | |
| # A question made entirely of stopwords still yields something to search. | |
| assert answer.keywords("who is it") == ["who", "is", "it"] | |
| def test_intent_detection(): | |
| for q, want in [("who is the principal", "who"), | |
| ("when is spring break", "when"), | |
| ("how much does lunch cost", "cost"), | |
| ("where is the school", "where"), | |
| ("what is the phone number", "contact"), | |
| ("is there a lunch menu", "menu"), | |
| ("tell me about robotics", "what")]: | |
| got, _ = answer.intent_of(q) | |
| assert got == want, f"{q!r} → {got}, expected {want}" | |
| def test_markdown_never_reaches_a_visitor(): | |
| assert answer.clean("**Student Lunch:** $2.85") == "Student Lunch: $2.85" | |
| assert "**" not in answer.clean("a **b** c **d**") | |
| assert answer.clean("spaced out\n\ntext") == "spaced out text" | |
| def test_bare_headings_lose_to_real_sentences(): | |
| """'Superintendent' repeats the question; it must not win as an answer.""" | |
| heading = answer.score_sentence("Superintendent", ["superintendent"], "who") | |
| real = answer.score_sentence( | |
| "The superintendent leads the district and reports to the school board.", | |
| ["superintendent"], "who") | |
| assert real > heading | |
| def test_fragments_lose_to_complete_sentences(): | |
| frag = answer.score_sentence("For school-specific contacts, visit the", TERMS, "contact") | |
| whole = answer.score_sentence("For school-specific contacts, visit the office page.", | |
| TERMS, "contact") | |
| assert whole > frag | |
| def test_intent_boosts_matching_shapes(): | |
| money = answer.score_sentence("Student lunch costs $2.85 each day.", TERMS, "cost") | |
| plain = answer.score_sentence("Student lunch is served every day here.", TERMS, "cost") | |
| assert money > plain | |
| dated = answer.score_sentence("Spring break runs March 24 through March 28.", ["spring"], "when") | |
| vague = answer.score_sentence("Spring break is a scheduled district holiday.", ["spring"], "when") | |
| assert dated > vague | |
| def test_unknown_question_says_so_instead_of_guessing(): | |
| r = answer.ask("do you sell unicorns and dragons here") | |
| assert r["ok"] is False | |
| assert "could not find" in r["answer"].lower() | |
| assert r["sources"] == [] | |
| def test_real_question_quotes_the_site(): | |
| r = answer.ask("how much is student lunch") | |
| assert r["ok"] is True | |
| assert "$" in r["answer"], f"a pricing question should quote a price: {r['answer']!r}" | |
| assert r["source"]["article_id"] > 0 | |
| assert "**" not in r["answer"] | |
| def test_unpublished_content_never_leaks(): | |
| """Insert an unpublished node, confirm the public search refuses it.""" | |
| db = mesh.grid() | |
| secret = "zzqx-unpublished-marker" | |
| rows = [("test.unpub", 999901, "Draft Page", "0", "text", f"{secret} draft", "", 1, 0, 1), | |
| ("test.private", 999902, "Private Page", "0", "text", f"{secret} private", "", 1, 1, 3), | |
| ("test.public", 999903, "Public Page", "0", "text", f"{secret} public", "", 1, 1, 1)] | |
| try: | |
| for r in rows: | |
| db.execute("INSERT OR REPLACE INTO nodes VALUES (?,?,?,?,?,?,?,?,?,?)", r) | |
| db.execute("INSERT INTO nodes_fts(coord,article,type,text,media) VALUES (?,?,?,?,?)", | |
| (r[0], r[2], r[4], r[5], r[6])) | |
| db.commit() | |
| titles = {h["article"] for h in mesh.public_search(secret)} | |
| assert "Public Page" in titles, "published public content must be findable" | |
| assert "Draft Page" not in titles, "UNPUBLISHED CONTENT LEAKED" | |
| assert "Private Page" not in titles, "ACCESS-RESTRICTED CONTENT LEAKED" | |
| served = answer.ask(f"{secret} information") | |
| for bad in ("draft", "private"): | |
| assert bad not in served["answer"].lower(), f"{bad} content reached the answer" | |
| finally: | |
| for r in rows: | |
| db.execute("DELETE FROM nodes WHERE coord = ?", (r[0],)) | |
| db.execute("DELETE FROM nodes_fts WHERE coord = ?", (r[0],)) | |
| db.commit() | |
| if __name__ == "__main__": | |
| tests = [v for k, v in sorted(globals().items()) if k.startswith("test_")] | |
| for t in tests: | |
| t() | |
| print(f" ok {t.__name__}") | |
| print(f"\n{len(tests)} checks passed") | |