"""Tests for the answer-eval citation parser (eval/run_answers.py)."""

import unittest

from eval.run_answers import cited_ids, is_refusal, lang_ok


class TestCitedIds(unittest.TestCase):
    def test_quran_urls(self):
        answer = ("[[1](https://surahquran.com/aya-255-sora-2.html)] ﴿...﴾ "
                  "[[1](https://quran.ksu.edu.sa/tafseer/qortobi/sura24-aya35.html)]")
        self.assertEqual(cited_ids(answer, {}), {"2:255", "24:35"})
        self.assertEqual(cited_ids("[[1](https://quran.com/2/278)]", {}), {"2:278"})

    def test_english_hadith_refs(self):
        answer = ("Recorded in Sahih al-Bukhari 6116. Also Sunan Abu Dawood No. 4800 "
                  "and Jami' at-Tirmidhi #2516, Sunan Ibn Majah, hadith 224.")
        self.assertEqual(cited_ids(answer, {}),
                         {"bukhari_6116", "abudawud_4800", "tirmidhi_2516", "ibnmajah_224"})

    def test_arabic_hadith_refs(self):
        answer = "رواه صحيح البخاري، رقم 13. وفي صحيح مسلم، رقم 45. وفي موطأ مالك، رقم 1."
        self.assertEqual(cited_ids(answer, {}), {"bukhari_13", "muslim_45", "malik_1"})
        answer = ("الترمذي، رقم الحديث 2024. وصحيح مسلم برقم 440. وسنن ابن ماجه، الرقم 4303. "
                  "والبخاري، كتاب الإيمان، حديث رقم 1. ورواه مسلم في صحيحه (6382).")
        self.assertEqual(cited_ids(answer, {}),
                         {"tirmidhi_2024", "muslim_440", "ibnmajah_4303", "bukhari_1", "muslim_6382"})

    def test_hadith_ref_needs_adjacent_number(self):
        # The number belongs to the collection right before it, not one earlier in the sentence
        self.assertEqual(cited_ids("Sahih al-Bukhari and Sahih Muslim 1907", {}), {"muslim_1907"})
        # "مسلم" as an ordinary word ("every Muslim") is not a citation
        self.assertEqual(cited_ids("طلب العلم فريضة على كل مسلم 3 مرات", {}), set())

    def test_fatwa_urls(self):
        url_to_id = {"https://www.islamweb.net/ar/fatwa/857": "islamweb_857"}
        answer = "مركز الفتوى [[1](https://www.islamweb.net/ar/fatwa/857/)]."
        self.assertEqual(cited_ids(answer, url_to_id), {"islamweb_857"})


class TestAnswerChecks(unittest.TestCase):
    def test_refusal_and_language(self):
        not_found = {"arabic": "لم أجد", "english": "not found"}
        self.assertTrue(is_refusal("The dataset does not contain sufficient information.", "en", not_found))
        self.assertFalse(is_refusal("Yes, it is obligatory. And Allah knows best.", "en", not_found))
        self.assertTrue(lang_ok("نعم، هذا واجب. والله أعلم.", "ar"))
        self.assertFalse(lang_ok("Yes, it is obligatory.", "ar"))


if __name__ == "__main__":
    unittest.main()


class TestRequiredLabels(unittest.TestCase):
    def test_hit_without_required_source_is_incomplete(self):
        from eval.run_eval import score_case
        case = {"id": "x", "category": "quran_topic", "query": "Is alcohol forbidden?",
                "relevant": ["5:90", "2:219"], "required": ["5:90"]}
        out = {"results": [{"id": "29:49"}, {"id": "2:219"}], "intent": "general", "top_score": 1.0,
               "gated": False, "confidence": 0.9, "confidence_source": "reranker"}
        row = score_case(case, out, 8, {})
        self.assertEqual(row["hit@k"], 1)          # still a hit, as before
        self.assertEqual(row["complete@k"], 0)     # but the decisive verse is missing
        self.assertEqual(row["missing_required"], ["5:90"])

    def test_required_must_be_subset_of_relevant(self):
        import tempfile
        from pathlib import Path
        from eval.run_eval import load_golden
        with tempfile.NamedTemporaryFile("w", suffix=".jsonl", delete=False, encoding="utf-8") as f:
            f.write('{"id": "x", "category": "c", "query": "q", "relevant": ["2:219"], "required": ["5:90"]}\n')
        with self.assertRaises(ValueError):
            load_golden(Path(f.name), None)