diff --git a/language_packs/data/en_core_math_v1/lexicon.jsonl b/language_packs/data/en_core_math_v1/lexicon.jsonl new file mode 100644 index 00000000..978bd53f --- /dev/null +++ b/language_packs/data/en_core_math_v1/lexicon.jsonl @@ -0,0 +1,208 @@ +{"entry_id": "en-math-0001", "surface": "buy", "lemma": "buy", "language": "en", "pos": "VERB", "semantic_domains": ["math.accumulation_verb", "math.verb.additive"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0002", "surface": "get", "lemma": "get", "language": "en", "pos": "VERB", "semantic_domains": ["math.accumulation_verb", "math.verb.additive"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0003", "surface": "find", "lemma": "find", "language": "en", "pos": "VERB", "semantic_domains": ["math.accumulation_verb", "math.verb.additive"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0004", "surface": "receive", "lemma": "receive", "language": "en", "pos": "VERB", "semantic_domains": ["math.accumulation_verb", "math.verb.additive"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0005", "surface": "earn", "lemma": "earn", "language": "en", "pos": "VERB", "semantic_domains": ["math.accumulation_verb", "math.verb.additive"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0006", "surface": "add", "lemma": "add", "language": "en", "pos": "VERB", "semantic_domains": ["math.accumulation_verb", "math.verb.additive"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0007", "surface": "collect", "lemma": "collect", "language": "en", "pos": "VERB", "semantic_domains": ["math.accumulation_verb", "math.verb.additive"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0008", "surface": "gather", "lemma": "gather", "language": "en", "pos": "VERB", "semantic_domains": ["math.accumulation_verb", "math.verb.additive"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0009", "surface": "catch", "lemma": "catch", "language": "en", "pos": "VERB", "semantic_domains": ["math.accumulation_verb", "math.verb.additive"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0010", "surface": "save", "lemma": "save", "language": "en", "pos": "VERB", "semantic_domains": ["math.accumulation_verb", "math.verb.additive"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0011", "surface": "bake", "lemma": "bake", "language": "en", "pos": "VERB", "semantic_domains": ["math.accumulation_verb", "math.verb.additive"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0012", "surface": "cook", "lemma": "cook", "language": "en", "pos": "VERB", "semantic_domains": ["math.accumulation_verb", "math.verb.additive"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0013", "surface": "slice", "lemma": "slice", "language": "en", "pos": "VERB", "semantic_domains": ["math.accumulation_verb", "math.verb.additive"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0014", "surface": "build", "lemma": "build", "language": "en", "pos": "VERB", "semantic_domains": ["math.accumulation_verb", "math.verb.additive"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0015", "surface": "grow", "lemma": "grow", "language": "en", "pos": "VERB", "semantic_domains": ["math.accumulation_verb", "math.verb.additive"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0016", "surface": "gain", "lemma": "gain", "language": "en", "pos": "VERB", "semantic_domains": ["math.accumulation_verb", "math.verb.additive"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0017", "surface": "charge", "lemma": "charge", "language": "en", "pos": "VERB", "semantic_domains": ["math.accumulation_verb", "math.verb.additive"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0018", "surface": "shuck", "lemma": "shuck", "language": "en", "pos": "VERB", "semantic_domains": ["math.capacity_verb", "math.verb.productive"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0019", "surface": "pick", "lemma": "pick", "language": "en", "pos": "VERB", "semantic_domains": ["math.capacity_verb", "math.verb.productive"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0020", "surface": "pack", "lemma": "pack", "language": "en", "pos": "VERB", "semantic_domains": ["math.capacity_verb", "math.verb.productive"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0021", "surface": "make", "lemma": "make", "language": "en", "pos": "VERB", "semantic_domains": ["math.capacity_verb", "math.verb.productive"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0022", "surface": "produce", "lemma": "produce", "language": "en", "pos": "VERB", "semantic_domains": ["math.capacity_verb", "math.verb.productive"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0023", "surface": "type", "lemma": "type", "language": "en", "pos": "VERB", "semantic_domains": ["math.capacity_verb", "math.verb.productive"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0024", "surface": "read", "lemma": "read", "language": "en", "pos": "VERB", "semantic_domains": ["math.capacity_verb", "math.verb.productive"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0025", "surface": "write", "lemma": "write", "language": "en", "pos": "VERB", "semantic_domains": ["math.capacity_verb", "math.verb.productive"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0026", "surface": "paint", "lemma": "paint", "language": "en", "pos": "VERB", "semantic_domains": ["math.capacity_verb", "math.verb.productive"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0027", "surface": "run", "lemma": "run", "language": "en", "pos": "VERB", "semantic_domains": ["math.capacity_verb", "math.verb.productive"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0028", "surface": "score", "lemma": "score", "language": "en", "pos": "VERB", "semantic_domains": ["math.capacity_verb", "math.verb.productive"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0029", "surface": "answer", "lemma": "answer", "language": "en", "pos": "VERB", "semantic_domains": ["math.capacity_verb", "math.verb.productive"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0030", "surface": "complete", "lemma": "complete", "language": "en", "pos": "VERB", "semantic_domains": ["math.capacity_verb", "math.verb.productive"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0031", "surface": "money", "lemma": "money", "language": "en", "pos": "NOUN", "semantic_domains": ["math.currency_unit_noun", "math.noun.quantity"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0032", "surface": "profit", "lemma": "profit", "language": "en", "pos": "NOUN", "semantic_domains": ["math.currency_unit_noun", "math.noun.quantity"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0033", "surface": "interest", "lemma": "interest", "language": "en", "pos": "NOUN", "semantic_domains": ["math.currency_unit_noun", "math.noun.quantity"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0034", "surface": "income", "lemma": "income", "language": "en", "pos": "NOUN", "semantic_domains": ["math.currency_unit_noun", "math.noun.quantity"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0035", "surface": "savings", "lemma": "savings", "language": "en", "pos": "NOUN", "semantic_domains": ["math.currency_unit_noun", "math.noun.quantity"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0036", "surface": "cost", "lemma": "cost", "language": "en", "pos": "NOUN", "semantic_domains": ["math.currency_unit_noun", "math.noun.quantity"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0037", "surface": "amount", "lemma": "amount", "language": "en", "pos": "NOUN", "semantic_domains": ["math.currency_unit_noun", "math.noun.quantity"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0038", "surface": "total", "lemma": "total", "language": "en", "pos": "NOUN", "semantic_domains": ["math.currency_unit_noun", "math.noun.quantity"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0039", "surface": "eat", "lemma": "eat", "language": "en", "pos": "VERB", "semantic_domains": ["math.depletion_verb", "math.verb.subtractive"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0040", "surface": "lose", "lemma": "lose", "language": "en", "pos": "VERB", "semantic_domains": ["math.depletion_verb", "math.verb.subtractive"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0041", "surface": "sell", "lemma": "sell", "language": "en", "pos": "VERB", "semantic_domains": ["math.depletion_verb", "math.verb.subtractive"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0042", "surface": "donate", "lemma": "donate", "language": "en", "pos": "VERB", "semantic_domains": ["math.depletion_verb", "math.verb.subtractive"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0043", "surface": "use", "lemma": "use", "language": "en", "pos": "VERB", "semantic_domains": ["math.depletion_verb", "math.verb.subtractive"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0044", "surface": "spend", "lemma": "spend", "language": "en", "pos": "VERB", "semantic_domains": ["math.depletion_verb", "math.verb.subtractive"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0045", "surface": "drop", "lemma": "drop", "language": "en", "pos": "VERB", "semantic_domains": ["math.depletion_verb", "math.verb.subtractive"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0046", "surface": "remove", "lemma": "remove", "language": "en", "pos": "VERB", "semantic_domains": ["math.depletion_verb", "math.verb.subtractive"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0047", "surface": "break", "lemma": "break", "language": "en", "pos": "VERB", "semantic_domains": ["math.depletion_verb", "math.verb.subtractive"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0048", "surface": "destroy", "lemma": "destroy", "language": "en", "pos": "VERB", "semantic_domains": ["math.depletion_verb", "math.verb.subtractive"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0049", "surface": "throw", "lemma": "throw", "language": "en", "pos": "VERB", "semantic_domains": ["math.depletion_verb", "math.verb.subtractive"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0050", "surface": "discard", "lemma": "discard", "language": "en", "pos": "VERB", "semantic_domains": ["math.depletion_verb", "math.verb.subtractive"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0051", "surface": "consume", "lemma": "consume", "language": "en", "pos": "VERB", "semantic_domains": ["math.depletion_verb", "math.verb.subtractive"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0052", "surface": "take", "lemma": "take", "language": "en", "pos": "VERB", "semantic_domains": ["math.depletion_verb", "math.verb.subtractive"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0053", "surface": "pay", "lemma": "pay", "language": "en", "pos": "VERB", "semantic_domains": ["math.depletion_verb", "math.verb.subtractive"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0054", "surface": "she", "lemma": "she", "language": "en", "pos": "PRON", "semantic_domains": ["math.entity_pronoun", "math.entity.pronominal"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0055", "surface": "he", "lemma": "he", "language": "en", "pos": "PRON", "semantic_domains": ["math.entity_pronoun", "math.entity.pronominal"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0056", "surface": "they", "lemma": "they", "language": "en", "pos": "PRON", "semantic_domains": ["math.entity_pronoun", "math.entity.pronominal"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0057", "surface": "it", "lemma": "it", "language": "en", "pos": "PRON", "semantic_domains": ["math.entity_pronoun", "math.entity.pronominal"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0058", "surface": "have", "lemma": "have", "language": "en", "pos": "VERB", "semantic_domains": ["math.possession_verb", "math.verb.stative"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0059", "surface": "alexa", "lemma": "alexa", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0060", "surface": "alice", "lemma": "alice", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0061", "surface": "amy", "lemma": "amy", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0062", "surface": "ann", "lemma": "ann", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0063", "surface": "anna", "lemma": "anna", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0064", "surface": "barbara", "lemma": "barbara", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0065", "surface": "betty", "lemma": "betty", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0066", "surface": "carol", "lemma": "carol", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0067", "surface": "carolyn", "lemma": "carolyn", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0068", "surface": "christine", "lemma": "christine", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0069", "surface": "cindy", "lemma": "cindy", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0070", "surface": "claire", "lemma": "claire", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0071", "surface": "cynthia", "lemma": "cynthia", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0072", "surface": "deborah", "lemma": "deborah", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0073", "surface": "diana", "lemma": "diana", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0074", "surface": "donna", "lemma": "donna", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0075", "surface": "dorothy", "lemma": "dorothy", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0076", "surface": "elizabeth", "lemma": "elizabeth", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0077", "surface": "ella", "lemma": "ella", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0078", "surface": "emily", "lemma": "emily", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0079", "surface": "emma", "lemma": "emma", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0080", "surface": "erica", "lemma": "erica", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0081", "surface": "francine", "lemma": "francine", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0082", "surface": "helen", "lemma": "helen", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0083", "surface": "jane", "lemma": "jane", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0084", "surface": "janet", "lemma": "janet", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0085", "surface": "jen", "lemma": "jen", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0086", "surface": "jennifer", "lemma": "jennifer", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0087", "surface": "jessica", "lemma": "jessica", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0088", "surface": "joyce", "lemma": "joyce", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0089", "surface": "judith", "lemma": "judith", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0090", "surface": "julie", "lemma": "julie", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0091", "surface": "karen", "lemma": "karen", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0092", "surface": "kate", "lemma": "kate", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0093", "surface": "kathleen", "lemma": "kathleen", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0094", "surface": "kelly", "lemma": "kelly", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0095", "surface": "laura", "lemma": "laura", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0096", "surface": "linda", "lemma": "linda", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0097", "surface": "lisa", "lemma": "lisa", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0098", "surface": "lilibeth", "lemma": "lilibeth", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0099", "surface": "lori", "lemma": "lori", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0100", "surface": "mandy", "lemma": "mandy", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0101", "surface": "marie", "lemma": "marie", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0102", "surface": "martha", "lemma": "martha", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0103", "surface": "marnie", "lemma": "marnie", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0104", "surface": "mary", "lemma": "mary", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0105", "surface": "melissa", "lemma": "melissa", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0106", "surface": "nancy", "lemma": "nancy", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0107", "surface": "nicole", "lemma": "nicole", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0108", "surface": "pamela", "lemma": "pamela", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0109", "surface": "patricia", "lemma": "patricia", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0110", "surface": "rachel", "lemma": "rachel", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0111", "surface": "rebecca", "lemma": "rebecca", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0112", "surface": "ruth", "lemma": "ruth", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0113", "surface": "sandra", "lemma": "sandra", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0114", "surface": "sarah", "lemma": "sarah", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0115", "surface": "sharon", "lemma": "sharon", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0116", "surface": "shirley", "lemma": "shirley", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0117", "surface": "stephanie", "lemma": "stephanie", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0118", "surface": "susan", "lemma": "susan", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0119", "surface": "tina", "lemma": "tina", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0120", "surface": "virginia", "lemma": "virginia", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_female", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0121", "surface": "aaron", "lemma": "aaron", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0122", "surface": "adam", "lemma": "adam", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0123", "surface": "alan", "lemma": "alan", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0124", "surface": "albert", "lemma": "albert", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0125", "surface": "andrew", "lemma": "andrew", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0126", "surface": "anthony", "lemma": "anthony", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0127", "surface": "arthur", "lemma": "arthur", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0128", "surface": "benjamin", "lemma": "benjamin", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0129", "surface": "bob", "lemma": "bob", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0130", "surface": "brian", "lemma": "brian", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0131", "surface": "bruce", "lemma": "bruce", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0132", "surface": "carl", "lemma": "carl", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0133", "surface": "carson", "lemma": "carson", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0134", "surface": "charles", "lemma": "charles", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0135", "surface": "christopher", "lemma": "christopher", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0136", "surface": "daniel", "lemma": "daniel", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0137", "surface": "david", "lemma": "david", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0138", "surface": "dennis", "lemma": "dennis", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0139", "surface": "donald", "lemma": "donald", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0140", "surface": "douglas", "lemma": "douglas", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0141", "surface": "edward", "lemma": "edward", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0142", "surface": "eric", "lemma": "eric", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0143", "surface": "ethan", "lemma": "ethan", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0144", "surface": "eugene", "lemma": "eugene", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0145", "surface": "fabian", "lemma": "fabian", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0146", "surface": "frank", "lemma": "frank", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0147", "surface": "gary", "lemma": "gary", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0148", "surface": "george", "lemma": "george", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0149", "surface": "gerald", "lemma": "gerald", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0150", "surface": "gregory", "lemma": "gregory", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0151", "surface": "harold", "lemma": "harold", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0152", "surface": "harry", "lemma": "harry", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0153", "surface": "henry", "lemma": "henry", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0154", "surface": "jack", "lemma": "jack", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0155", "surface": "james", "lemma": "james", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0156", "surface": "jason", "lemma": "jason", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0157", "surface": "jeffrey", "lemma": "jeffrey", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0158", "surface": "jeremy", "lemma": "jeremy", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0159", "surface": "jerry", "lemma": "jerry", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0160", "surface": "jesse", "lemma": "jesse", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0161", "surface": "john", "lemma": "john", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0162", "surface": "jonathan", "lemma": "jonathan", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0163", "surface": "joseph", "lemma": "joseph", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0164", "surface": "joshua", "lemma": "joshua", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0165", "surface": "keith", "lemma": "keith", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0166", "surface": "kenneth", "lemma": "kenneth", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0167", "surface": "kevin", "lemma": "kevin", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0168", "surface": "larry", "lemma": "larry", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0169", "surface": "lawrence", "lemma": "lawrence", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0170", "surface": "mark", "lemma": "mark", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0171", "surface": "matthew", "lemma": "matthew", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0172", "surface": "michael", "lemma": "michael", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0173", "surface": "nathan", "lemma": "nathan", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0174", "surface": "nicholas", "lemma": "nicholas", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0175", "surface": "patrick", "lemma": "patrick", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0176", "surface": "paul", "lemma": "paul", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0177", "surface": "peter", "lemma": "peter", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0178", "surface": "philip", "lemma": "philip", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0179", "surface": "ralph", "lemma": "ralph", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0180", "surface": "raymond", "lemma": "raymond", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0181", "surface": "richard", "lemma": "richard", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0182", "surface": "robert", "lemma": "robert", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0183", "surface": "roger", "lemma": "roger", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0184", "surface": "ronald", "lemma": "ronald", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0185", "surface": "ryan", "lemma": "ryan", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0186", "surface": "samuel", "lemma": "samuel", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0187", "surface": "scott", "lemma": "scott", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0188", "surface": "sean", "lemma": "sean", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0189", "surface": "stephen", "lemma": "stephen", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0190", "surface": "steven", "lemma": "steven", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0191", "surface": "thomas", "lemma": "thomas", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0192", "surface": "timothy", "lemma": "timothy", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0193", "surface": "tom", "lemma": "tom", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0194", "surface": "walter", "lemma": "walter", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0195", "surface": "wayne", "lemma": "wayne", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0196", "surface": "william", "lemma": "william", "language": "en", "pos": "PROPN", "semantic_domains": ["math.proper_noun_entity_male", "math.entity.proper"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0197", "surface": "how", "lemma": "how", "language": "en", "pos": "ADV", "semantic_domains": ["math.question_open", "math.interrogative"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0198", "surface": "what", "lemma": "what", "language": "en", "pos": "ADV", "semantic_domains": ["math.question_open", "math.interrogative"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0199", "surface": "left", "lemma": "left", "language": "en", "pos": "ADJ", "semantic_domains": ["math.residual_modifier", "math.modifier.terminal"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0200", "surface": "remaining", "lemma": "remaining", "language": "en", "pos": "ADJ", "semantic_domains": ["math.residual_modifier", "math.modifier.terminal"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0201", "surface": "after", "lemma": "after", "language": "en", "pos": "ADJ", "semantic_domains": ["math.residual_modifier", "math.modifier.terminal"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0202", "surface": "give", "lemma": "give", "language": "en", "pos": "VERB", "semantic_domains": ["math.transfer_verb", "math.verb.directional"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0203", "surface": "send", "lemma": "send", "language": "en", "pos": "VERB", "semantic_domains": ["math.transfer_verb", "math.verb.directional"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0204", "surface": "hand", "lemma": "hand", "language": "en", "pos": "VERB", "semantic_domains": ["math.transfer_verb", "math.verb.directional"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0205", "surface": "pass", "lemma": "pass", "language": "en", "pos": "VERB", "semantic_domains": ["math.transfer_verb", "math.verb.directional"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0206", "surface": "mail", "lemma": "mail", "language": "en", "pos": "VERB", "semantic_domains": ["math.transfer_verb", "math.verb.directional"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0207", "surface": "deliver", "lemma": "deliver", "language": "en", "pos": "VERB", "semantic_domains": ["math.transfer_verb", "math.verb.directional"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} +{"entry_id": "en-math-0208", "surface": "return", "lemma": "return", "language": "en", "pos": "VERB", "semantic_domains": ["math.transfer_verb", "math.verb.directional"], "provenance_ids": ["ported_from_math_candidate_parser_2026-05-26"], "epistemic_status": "coherent"} diff --git a/language_packs/data/en_core_math_v1/lexicon/accumulation_verb.jsonl b/language_packs/data/en_core_math_v1/lexicon/accumulation_verb.jsonl new file mode 100644 index 00000000..bb9e63a5 --- /dev/null +++ b/language_packs/data/en_core_math_v1/lexicon/accumulation_verb.jsonl @@ -0,0 +1,17 @@ +{"lemma": "buy", "category": "accumulation_verb", "aliases": ["buys", "bought"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "get", "category": "accumulation_verb", "aliases": ["gets", "got"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "find", "category": "accumulation_verb", "aliases": ["finds", "found"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "receive", "category": "accumulation_verb", "aliases": ["receives", "received"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "earn", "category": "accumulation_verb", "aliases": ["earns", "earned"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "add", "category": "accumulation_verb", "aliases": ["adds", "added"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "collect", "category": "accumulation_verb", "aliases": ["collects", "collected"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "gather", "category": "accumulation_verb", "aliases": ["gathers", "gathered"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "catch", "category": "accumulation_verb", "aliases": ["catches", "caught"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "save", "category": "accumulation_verb", "aliases": ["saves", "saved"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "bake", "category": "accumulation_verb", "aliases": ["bakes", "baked"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "cook", "category": "accumulation_verb", "aliases": ["cooks", "cooked"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "slice", "category": "accumulation_verb", "aliases": ["slices", "sliced"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "build", "category": "accumulation_verb", "aliases": ["builds", "built"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "grow", "category": "accumulation_verb", "aliases": ["grows", "grew"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "gain", "category": "accumulation_verb", "aliases": ["gains", "gained"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "charge", "category": "accumulation_verb", "aliases": ["charges", "charged"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} diff --git a/language_packs/data/en_core_math_v1/lexicon/capacity_verb.jsonl b/language_packs/data/en_core_math_v1/lexicon/capacity_verb.jsonl new file mode 100644 index 00000000..9c76eb7c --- /dev/null +++ b/language_packs/data/en_core_math_v1/lexicon/capacity_verb.jsonl @@ -0,0 +1,13 @@ +{"lemma": "shuck", "category": "capacity_verb", "aliases": ["shucks"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "pick", "category": "capacity_verb", "aliases": ["picks"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "pack", "category": "capacity_verb", "aliases": ["packs"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "make", "category": "capacity_verb", "aliases": ["makes", "made"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "produce", "category": "capacity_verb", "aliases": ["produces", "produced"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "type", "category": "capacity_verb", "aliases": ["types", "typed"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "read", "category": "capacity_verb", "aliases": ["reads"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "write", "category": "capacity_verb", "aliases": ["writes", "wrote"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "paint", "category": "capacity_verb", "aliases": ["paints", "painted"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "run", "category": "capacity_verb", "aliases": ["runs", "ran"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "score", "category": "capacity_verb", "aliases": ["scores", "scored"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "answer", "category": "capacity_verb", "aliases": ["answers", "answered"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "complete", "category": "capacity_verb", "aliases": ["completes", "completed"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} diff --git a/language_packs/data/en_core_math_v1/lexicon/currency_unit_noun.jsonl b/language_packs/data/en_core_math_v1/lexicon/currency_unit_noun.jsonl new file mode 100644 index 00000000..ef888b46 --- /dev/null +++ b/language_packs/data/en_core_math_v1/lexicon/currency_unit_noun.jsonl @@ -0,0 +1,8 @@ +{"lemma": "money", "category": "currency_unit_noun", "aliases": [], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "profit", "category": "currency_unit_noun", "aliases": [], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "interest", "category": "currency_unit_noun", "aliases": [], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "income", "category": "currency_unit_noun", "aliases": [], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "savings", "category": "currency_unit_noun", "aliases": [], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "cost", "category": "currency_unit_noun", "aliases": ["costs"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "amount", "category": "currency_unit_noun", "aliases": ["amounts"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "total", "category": "currency_unit_noun", "aliases": ["totals"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} diff --git a/language_packs/data/en_core_math_v1/lexicon/depletion_verb.jsonl b/language_packs/data/en_core_math_v1/lexicon/depletion_verb.jsonl new file mode 100644 index 00000000..9570eec2 --- /dev/null +++ b/language_packs/data/en_core_math_v1/lexicon/depletion_verb.jsonl @@ -0,0 +1,15 @@ +{"lemma": "eat", "category": "depletion_verb", "aliases": ["eats", "ate"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "lose", "category": "depletion_verb", "aliases": ["loses", "lost"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "sell", "category": "depletion_verb", "aliases": ["sells", "sold"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "donate", "category": "depletion_verb", "aliases": ["donates", "donated"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "use", "category": "depletion_verb", "aliases": ["uses", "used"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "spend", "category": "depletion_verb", "aliases": ["spends", "spent"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "drop", "category": "depletion_verb", "aliases": ["drops", "dropped"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "remove", "category": "depletion_verb", "aliases": ["removes", "removed"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "break", "category": "depletion_verb", "aliases": ["breaks", "broke"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "destroy", "category": "depletion_verb", "aliases": ["destroys", "destroyed"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "throw", "category": "depletion_verb", "aliases": ["throws", "threw"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "discard", "category": "depletion_verb", "aliases": ["discards", "discarded"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "consume", "category": "depletion_verb", "aliases": ["consumes", "consumed"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "take", "category": "depletion_verb", "aliases": ["takes", "took"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "pay", "category": "depletion_verb", "aliases": ["pays", "paid"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} diff --git a/language_packs/data/en_core_math_v1/lexicon/entity_pronoun.jsonl b/language_packs/data/en_core_math_v1/lexicon/entity_pronoun.jsonl new file mode 100644 index 00000000..5cd235a5 --- /dev/null +++ b/language_packs/data/en_core_math_v1/lexicon/entity_pronoun.jsonl @@ -0,0 +1,4 @@ +{"lemma": "she", "category": "entity_pronoun", "aliases": ["her"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "he", "category": "entity_pronoun", "aliases": ["him", "his"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "they", "category": "entity_pronoun", "aliases": ["them", "their"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "it", "category": "entity_pronoun", "aliases": ["its"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} diff --git a/language_packs/data/en_core_math_v1/lexicon/possession_verb.jsonl b/language_packs/data/en_core_math_v1/lexicon/possession_verb.jsonl new file mode 100644 index 00000000..c58f038c --- /dev/null +++ b/language_packs/data/en_core_math_v1/lexicon/possession_verb.jsonl @@ -0,0 +1 @@ +{"lemma": "have", "category": "possession_verb", "aliases": ["has", "had"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} diff --git a/language_packs/data/en_core_math_v1/lexicon/proper_noun_entity_female.jsonl b/language_packs/data/en_core_math_v1/lexicon/proper_noun_entity_female.jsonl new file mode 100644 index 00000000..2ce787ec --- /dev/null +++ b/language_packs/data/en_core_math_v1/lexicon/proper_noun_entity_female.jsonl @@ -0,0 +1,62 @@ +{"lemma": "alexa", "category": "proper_noun_entity_female", "aliases": ["Alexa"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "alice", "category": "proper_noun_entity_female", "aliases": ["Alice"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "amy", "category": "proper_noun_entity_female", "aliases": ["Amy"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "ann", "category": "proper_noun_entity_female", "aliases": ["Ann"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "anna", "category": "proper_noun_entity_female", "aliases": ["Anna"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "barbara", "category": "proper_noun_entity_female", "aliases": ["Barbara"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "betty", "category": "proper_noun_entity_female", "aliases": ["Betty"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "carol", "category": "proper_noun_entity_female", "aliases": ["Carol"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "carolyn", "category": "proper_noun_entity_female", "aliases": ["Carolyn"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "christine", "category": "proper_noun_entity_female", "aliases": ["Christine"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "cindy", "category": "proper_noun_entity_female", "aliases": ["Cindy"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "claire", "category": "proper_noun_entity_female", "aliases": ["Claire"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "cynthia", "category": "proper_noun_entity_female", "aliases": ["Cynthia"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "deborah", "category": "proper_noun_entity_female", "aliases": ["Deborah"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "diana", "category": "proper_noun_entity_female", "aliases": ["Diana"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "donna", "category": "proper_noun_entity_female", "aliases": ["Donna"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "dorothy", "category": "proper_noun_entity_female", "aliases": ["Dorothy"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "elizabeth", "category": "proper_noun_entity_female", "aliases": ["Elizabeth"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "ella", "category": "proper_noun_entity_female", "aliases": ["Ella"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "emily", "category": "proper_noun_entity_female", "aliases": ["Emily"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "emma", "category": "proper_noun_entity_female", "aliases": ["Emma"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "erica", "category": "proper_noun_entity_female", "aliases": ["Erica"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "francine", "category": "proper_noun_entity_female", "aliases": ["Francine"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "helen", "category": "proper_noun_entity_female", "aliases": ["Helen"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "jane", "category": "proper_noun_entity_female", "aliases": ["Jane"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "janet", "category": "proper_noun_entity_female", "aliases": ["Janet"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "jen", "category": "proper_noun_entity_female", "aliases": ["Jen"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "jennifer", "category": "proper_noun_entity_female", "aliases": ["Jennifer"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "jessica", "category": "proper_noun_entity_female", "aliases": ["Jessica"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "joyce", "category": "proper_noun_entity_female", "aliases": ["Joyce"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "judith", "category": "proper_noun_entity_female", "aliases": ["Judith"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "julie", "category": "proper_noun_entity_female", "aliases": ["Julie"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "karen", "category": "proper_noun_entity_female", "aliases": ["Karen"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "kate", "category": "proper_noun_entity_female", "aliases": ["Kate"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "kathleen", "category": "proper_noun_entity_female", "aliases": ["Kathleen"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "kelly", "category": "proper_noun_entity_female", "aliases": ["Kelly"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "laura", "category": "proper_noun_entity_female", "aliases": ["Laura"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "linda", "category": "proper_noun_entity_female", "aliases": ["Linda"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "lisa", "category": "proper_noun_entity_female", "aliases": ["Lisa"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "lilibeth", "category": "proper_noun_entity_female", "aliases": ["Lilibeth"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "lori", "category": "proper_noun_entity_female", "aliases": ["Lori"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "mandy", "category": "proper_noun_entity_female", "aliases": ["Mandy"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "marie", "category": "proper_noun_entity_female", "aliases": ["Marie"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "martha", "category": "proper_noun_entity_female", "aliases": ["Martha"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "marnie", "category": "proper_noun_entity_female", "aliases": ["Marnie"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "mary", "category": "proper_noun_entity_female", "aliases": ["Mary"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "melissa", "category": "proper_noun_entity_female", "aliases": ["Melissa"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "nancy", "category": "proper_noun_entity_female", "aliases": ["Nancy"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "nicole", "category": "proper_noun_entity_female", "aliases": ["Nicole"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "pamela", "category": "proper_noun_entity_female", "aliases": ["Pamela"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "patricia", "category": "proper_noun_entity_female", "aliases": ["Patricia"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "rachel", "category": "proper_noun_entity_female", "aliases": ["Rachel"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "rebecca", "category": "proper_noun_entity_female", "aliases": ["Rebecca"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "ruth", "category": "proper_noun_entity_female", "aliases": ["Ruth"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "sandra", "category": "proper_noun_entity_female", "aliases": ["Sandra"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "sarah", "category": "proper_noun_entity_female", "aliases": ["Sarah"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "sharon", "category": "proper_noun_entity_female", "aliases": ["Sharon"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "shirley", "category": "proper_noun_entity_female", "aliases": ["Shirley"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "stephanie", "category": "proper_noun_entity_female", "aliases": ["Stephanie"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "susan", "category": "proper_noun_entity_female", "aliases": ["Susan"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "tina", "category": "proper_noun_entity_female", "aliases": ["Tina"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "virginia", "category": "proper_noun_entity_female", "aliases": ["Virginia"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} diff --git a/language_packs/data/en_core_math_v1/lexicon/proper_noun_entity_male.jsonl b/language_packs/data/en_core_math_v1/lexicon/proper_noun_entity_male.jsonl new file mode 100644 index 00000000..85716763 --- /dev/null +++ b/language_packs/data/en_core_math_v1/lexicon/proper_noun_entity_male.jsonl @@ -0,0 +1,76 @@ +{"lemma": "aaron", "category": "proper_noun_entity_male", "aliases": ["Aaron"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "adam", "category": "proper_noun_entity_male", "aliases": ["Adam"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "alan", "category": "proper_noun_entity_male", "aliases": ["Alan"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "albert", "category": "proper_noun_entity_male", "aliases": ["Albert"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "andrew", "category": "proper_noun_entity_male", "aliases": ["Andrew"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "anthony", "category": "proper_noun_entity_male", "aliases": ["Anthony"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "arthur", "category": "proper_noun_entity_male", "aliases": ["Arthur"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "benjamin", "category": "proper_noun_entity_male", "aliases": ["Benjamin"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "bob", "category": "proper_noun_entity_male", "aliases": ["Bob"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "brian", "category": "proper_noun_entity_male", "aliases": ["Brian"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "bruce", "category": "proper_noun_entity_male", "aliases": ["Bruce"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "carl", "category": "proper_noun_entity_male", "aliases": ["Carl"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "carson", "category": "proper_noun_entity_male", "aliases": ["Carson"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "charles", "category": "proper_noun_entity_male", "aliases": ["Charles"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "christopher", "category": "proper_noun_entity_male", "aliases": ["Christopher"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "daniel", "category": "proper_noun_entity_male", "aliases": ["Daniel"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "david", "category": "proper_noun_entity_male", "aliases": ["David"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "dennis", "category": "proper_noun_entity_male", "aliases": ["Dennis"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "donald", "category": "proper_noun_entity_male", "aliases": ["Donald"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "douglas", "category": "proper_noun_entity_male", "aliases": ["Douglas"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "edward", "category": "proper_noun_entity_male", "aliases": ["Edward"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "eric", "category": "proper_noun_entity_male", "aliases": ["Eric"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "ethan", "category": "proper_noun_entity_male", "aliases": ["Ethan"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "eugene", "category": "proper_noun_entity_male", "aliases": ["Eugene"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "fabian", "category": "proper_noun_entity_male", "aliases": ["Fabian"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "frank", "category": "proper_noun_entity_male", "aliases": ["Frank"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "gary", "category": "proper_noun_entity_male", "aliases": ["Gary"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "george", "category": "proper_noun_entity_male", "aliases": ["George"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "gerald", "category": "proper_noun_entity_male", "aliases": ["Gerald"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "gregory", "category": "proper_noun_entity_male", "aliases": ["Gregory"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "harold", "category": "proper_noun_entity_male", "aliases": ["Harold"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "harry", "category": "proper_noun_entity_male", "aliases": ["Harry"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "henry", "category": "proper_noun_entity_male", "aliases": ["Henry"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "jack", "category": "proper_noun_entity_male", "aliases": ["Jack"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "james", "category": "proper_noun_entity_male", "aliases": ["James"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "jason", "category": "proper_noun_entity_male", "aliases": ["Jason"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "jeffrey", "category": "proper_noun_entity_male", "aliases": ["Jeffrey"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "jeremy", "category": "proper_noun_entity_male", "aliases": ["Jeremy"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "jerry", "category": "proper_noun_entity_male", "aliases": ["Jerry"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "jesse", "category": "proper_noun_entity_male", "aliases": ["Jesse"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "john", "category": "proper_noun_entity_male", "aliases": ["John"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "jonathan", "category": "proper_noun_entity_male", "aliases": ["Jonathan"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "joseph", "category": "proper_noun_entity_male", "aliases": ["Joseph"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "joshua", "category": "proper_noun_entity_male", "aliases": ["Joshua"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "keith", "category": "proper_noun_entity_male", "aliases": ["Keith"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "kenneth", "category": "proper_noun_entity_male", "aliases": ["Kenneth"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "kevin", "category": "proper_noun_entity_male", "aliases": ["Kevin"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "larry", "category": "proper_noun_entity_male", "aliases": ["Larry"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "lawrence", "category": "proper_noun_entity_male", "aliases": ["Lawrence"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "mark", "category": "proper_noun_entity_male", "aliases": ["Mark"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "matthew", "category": "proper_noun_entity_male", "aliases": ["Matthew"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "michael", "category": "proper_noun_entity_male", "aliases": ["Michael"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "nathan", "category": "proper_noun_entity_male", "aliases": ["Nathan"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "nicholas", "category": "proper_noun_entity_male", "aliases": ["Nicholas"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "patrick", "category": "proper_noun_entity_male", "aliases": ["Patrick"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "paul", "category": "proper_noun_entity_male", "aliases": ["Paul"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "peter", "category": "proper_noun_entity_male", "aliases": ["Peter"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "philip", "category": "proper_noun_entity_male", "aliases": ["Philip"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "ralph", "category": "proper_noun_entity_male", "aliases": ["Ralph"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "raymond", "category": "proper_noun_entity_male", "aliases": ["Raymond"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "richard", "category": "proper_noun_entity_male", "aliases": ["Richard"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "robert", "category": "proper_noun_entity_male", "aliases": ["Robert"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "roger", "category": "proper_noun_entity_male", "aliases": ["Roger"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "ronald", "category": "proper_noun_entity_male", "aliases": ["Ronald"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "ryan", "category": "proper_noun_entity_male", "aliases": ["Ryan"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "samuel", "category": "proper_noun_entity_male", "aliases": ["Samuel"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "scott", "category": "proper_noun_entity_male", "aliases": ["Scott"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "sean", "category": "proper_noun_entity_male", "aliases": ["Sean"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "stephen", "category": "proper_noun_entity_male", "aliases": ["Stephen"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "steven", "category": "proper_noun_entity_male", "aliases": ["Steven"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "thomas", "category": "proper_noun_entity_male", "aliases": ["Thomas"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "timothy", "category": "proper_noun_entity_male", "aliases": ["Timothy"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "tom", "category": "proper_noun_entity_male", "aliases": ["Tom"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "walter", "category": "proper_noun_entity_male", "aliases": ["Walter"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "wayne", "category": "proper_noun_entity_male", "aliases": ["Wayne"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "william", "category": "proper_noun_entity_male", "aliases": ["William"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} diff --git a/language_packs/data/en_core_math_v1/lexicon/question_open.jsonl b/language_packs/data/en_core_math_v1/lexicon/question_open.jsonl new file mode 100644 index 00000000..a48979ca --- /dev/null +++ b/language_packs/data/en_core_math_v1/lexicon/question_open.jsonl @@ -0,0 +1,2 @@ +{"lemma": "how", "category": "question_open", "aliases": [], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "what", "category": "question_open", "aliases": [], "provenance": "ported_from_math_candidate_parser_2026-05-26"} diff --git a/language_packs/data/en_core_math_v1/lexicon/residual_modifier.jsonl b/language_packs/data/en_core_math_v1/lexicon/residual_modifier.jsonl new file mode 100644 index 00000000..e176cdd1 --- /dev/null +++ b/language_packs/data/en_core_math_v1/lexicon/residual_modifier.jsonl @@ -0,0 +1,3 @@ +{"lemma": "left", "category": "residual_modifier", "aliases": [], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "remaining", "category": "residual_modifier", "aliases": [], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "after", "category": "residual_modifier", "aliases": [], "provenance": "ported_from_math_candidate_parser_2026-05-26"} diff --git a/language_packs/data/en_core_math_v1/lexicon/transfer_verb.jsonl b/language_packs/data/en_core_math_v1/lexicon/transfer_verb.jsonl new file mode 100644 index 00000000..cd937053 --- /dev/null +++ b/language_packs/data/en_core_math_v1/lexicon/transfer_verb.jsonl @@ -0,0 +1,7 @@ +{"lemma": "give", "category": "transfer_verb", "aliases": ["gives", "gave"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "send", "category": "transfer_verb", "aliases": ["sends", "sent"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "hand", "category": "transfer_verb", "aliases": ["hands", "handed"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "pass", "category": "transfer_verb", "aliases": ["passes", "passed"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "mail", "category": "transfer_verb", "aliases": ["mails", "mailed"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "deliver", "category": "transfer_verb", "aliases": ["delivers", "delivered"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} +{"lemma": "return", "category": "transfer_verb", "aliases": ["returns", "returned"], "provenance": "ported_from_math_candidate_parser_2026-05-26"} diff --git a/language_packs/data/en_core_math_v1/manifest.json b/language_packs/data/en_core_math_v1/manifest.json new file mode 100644 index 00000000..4d42c95b --- /dev/null +++ b/language_packs/data/en_core_math_v1/manifest.json @@ -0,0 +1,26 @@ +{ + "pack_id": "en_core_math_v1", + "language": "en", + "role": "domain_seed", + "script": "Latin", + "normalization_policy": "unitize_versor", + "source_manifest": "en_core_math_v1.lexicon.jsonl", + "determinism_class": "D0", + "checksum": "4e492514cdd293d1868054fb3c1bb8111e82c27a4c5c5ab0c1371ada1bed6536", + "version": "1.0.0", + "gate_engaged": false, + "oov_policy": "tagged_fallback", + "definitional_layer": false, + "provenance": "adr-0164:seed:ported_from_math_candidate_parser_2026-05-26", + "deferred": [ + "give_away: multi-word phrase from _INIT_MUT_SUBTRACT_VERBS; needs phrase-level lexicon support", + "pick_up: multi-word phrase from _INIT_MUT_ADD_VERBS; needs phrase-level lexicon support", + "aggregate_modifier category (in_all, altogether, combined): multi-word forms; not in single-token source whitelist", + "question_continuous_qty (much, long, far, old): not attested as isolated constants in math_candidate_parser.py; requires reader Phase 2 ratification", + "question_discrete_qty (many): not a standalone constant; deferred until reader integration", + "question_comparative (more, less, longer, fewer): from ADR-0164 table but not extracted as closed set in parser source", + "distributive_modifier (each, per): from ADR-0164 table; not in parser source as closed set", + "count_unit_noun: dynamic open class (apples, books, kids); not a fixed whitelist; deferred by design", + "time_unit_noun: available in _TIME_UNIT_SET but not in brief's required categories; deferred to Phase 3" + ] +} diff --git a/tests/test_en_core_math_v1_pack.py b/tests/test_en_core_math_v1_pack.py new file mode 100644 index 00000000..e2861e85 --- /dev/null +++ b/tests/test_en_core_math_v1_pack.py @@ -0,0 +1,197 @@ +"""Contract tests for the ``en_core_math_v1`` math seed lexicon pack. + +Pins: + - Pack loads via ``load_pack("en_core_math_v1")`` with passing checksum. + - Lemma counts per category match the ported source whitelists (regression). + - Zero overlap between mutually-exclusive categories. + - Every lemma entry carries the expected provenance tag. + - Per-category JSONL files in lexicon/ are structurally valid. +""" + +from __future__ import annotations + +import json +from pathlib import Path + +import pytest + +from language_packs.compiler import load_pack, load_pack_entries + +PACK_ID = "en_core_math_v1" + +_PACK_DIR = ( + Path(__file__).parent.parent / "language_packs" / "data" / PACK_ID +) +_LEXICON_DIR = _PACK_DIR / "lexicon" + +PROVENANCE_TAG = "ported_from_math_candidate_parser_2026-05-26" + +# Expected lemma counts sourced directly from the ported whitelist constants. +# Change only when the source constant changes AND an ADR ratifies the delta. +EXPECTED_CATEGORY_COUNTS: dict[str, int] = { + "accumulation_verb": 17, + "depletion_verb": 15, + "transfer_verb": 7, + "currency_unit_noun": 8, + "entity_pronoun": 4, + "proper_noun_entity_female": 62, + "proper_noun_entity_male": 76, + "possession_verb": 1, + "capacity_verb": 13, + "question_open": 2, + "residual_modifier": 3, +} + +EXPECTED_TOTAL = sum(EXPECTED_CATEGORY_COUNTS.values()) # 208 + + +def _read_category(cat: str) -> list[dict]: + path = _LEXICON_DIR / f"{cat}.jsonl" + return [ + json.loads(line) + for line in path.read_text(encoding="utf-8").splitlines() + if line.strip() + ] + + +# --------------------------------------------------------------------------- +# Pack-load + checksum +# --------------------------------------------------------------------------- + +def test_pack_loads_with_matching_checksum() -> None: + """``load_pack`` raises ``ValueError`` on checksum mismatch. + + A clean load proves manifest.checksum equals SHA-256 of bytes on disk — + per CLAUDE.md §Semantic Pack Discipline. + """ + manifest, _ = load_pack(PACK_ID) + assert manifest.pack_id == PACK_ID + assert len(manifest.checksum) == 64 + assert all(c in "0123456789abcdef" for c in manifest.checksum) + + +def test_total_lemma_count() -> None: + """Total entries in lexicon.jsonl must equal sum of per-category counts.""" + entries = load_pack_entries(PACK_ID) + assert len(entries) == EXPECTED_TOTAL, ( + f"Expected {EXPECTED_TOTAL} entries, got {len(entries)}. " + "A source whitelist changed without an ADR update." + ) + + +# --------------------------------------------------------------------------- +# Per-category file structure +# --------------------------------------------------------------------------- + +@pytest.mark.parametrize("cat", sorted(EXPECTED_CATEGORY_COUNTS)) +def test_category_file_exists(cat: str) -> None: + path = _LEXICON_DIR / f"{cat}.jsonl" + assert path.exists(), f"lexicon/{cat}.jsonl missing" + + +@pytest.mark.parametrize("cat,expected", sorted(EXPECTED_CATEGORY_COUNTS.items())) +def test_category_lemma_count(cat: str, expected: int) -> None: + rows = _read_category(cat) + assert len(rows) == expected, ( + f"{cat}: expected {expected} lemmas, got {len(rows)}. " + "Source whitelist changed — update EXPECTED_CATEGORY_COUNTS and the ADR." + ) + + +@pytest.mark.parametrize("cat", sorted(EXPECTED_CATEGORY_COUNTS)) +def test_category_provenance(cat: str) -> None: + """Every lemma in every per-category file must carry the provenance tag.""" + for row in _read_category(cat): + assert row.get("provenance") == PROVENANCE_TAG, ( + f"{cat}: lemma {row.get('lemma')!r} has wrong provenance " + f"{row.get('provenance')!r}; expected {PROVENANCE_TAG!r}" + ) + + +@pytest.mark.parametrize("cat", sorted(EXPECTED_CATEGORY_COUNTS)) +def test_category_field_is_set(cat: str) -> None: + for row in _read_category(cat): + assert row.get("category") == cat, ( + f"lemma {row.get('lemma')!r} in {cat}.jsonl has " + f"category={row.get('category')!r}" + ) + + +# --------------------------------------------------------------------------- +# Mutual exclusivity +# --------------------------------------------------------------------------- + +def test_accumulation_depletion_disjoint() -> None: + """accumulation_verb ∩ depletion_verb must be empty.""" + acc = {r["lemma"] for r in _read_category("accumulation_verb")} + dep = {r["lemma"] for r in _read_category("depletion_verb")} + overlap = acc & dep + assert not overlap, ( + f"Verbs in both accumulation_verb and depletion_verb: {sorted(overlap)}. " + "Each verb must have a unique polarity assignment." + ) + + +def test_accumulation_transfer_disjoint() -> None: + """accumulation_verb ∩ transfer_verb must be empty.""" + acc = {r["lemma"] for r in _read_category("accumulation_verb")} + trn = {r["lemma"] for r in _read_category("transfer_verb")} + overlap = acc & trn + assert not overlap, ( + f"Verbs in both accumulation_verb and transfer_verb: {sorted(overlap)}" + ) + + +def test_depletion_transfer_disjoint() -> None: + """depletion_verb ∩ transfer_verb must be empty. + + Transfer verbs (give, send, …) were explicitly removed from + depletion_verb to resolve the dual-entry in SUBTRACT_VERBS/TRANSFER_VERBS. + """ + dep = {r["lemma"] for r in _read_category("depletion_verb")} + trn = {r["lemma"] for r in _read_category("transfer_verb")} + overlap = dep & trn + assert not overlap, ( + f"Verbs in both depletion_verb and transfer_verb: {sorted(overlap)}" + ) + + +def test_capacity_verb_disjoint_from_accumulation() -> None: + """capacity_verb ∩ accumulation_verb must be empty (make/pick/pack placed in capacity).""" + cap = {r["lemma"] for r in _read_category("capacity_verb")} + acc = {r["lemma"] for r in _read_category("accumulation_verb")} + overlap = cap & acc + assert not overlap, ( + f"Verbs in both capacity_verb and accumulation_verb: {sorted(overlap)}. " + "make/pick/pack belong exclusively to capacity_verb." + ) + + +def test_female_male_names_disjoint() -> None: + """Female and male name lists must not share any lemma.""" + female = {r["lemma"] for r in _read_category("proper_noun_entity_female")} + male = {r["lemma"] for r in _read_category("proper_noun_entity_male")} + overlap = female & male + assert not overlap, ( + f"Names in both female and male lists: {sorted(overlap)}" + ) + + +# --------------------------------------------------------------------------- +# Compiled entry provenance +# --------------------------------------------------------------------------- + +def test_every_compiled_entry_has_provenance() -> None: + """Every entry in lexicon.jsonl carries the provenance tag in provenance_ids.""" + entries = load_pack_entries(PACK_ID) + missing = [e.lemma for e in entries if PROVENANCE_TAG not in (e.provenance_ids or ())] + assert not missing, ( + f"Entries missing provenance tag: {missing[:10]}" + ) + + +def test_every_compiled_entry_has_two_semantic_domains() -> None: + """Each entry must carry ≥2 semantic domains for surface composition headroom.""" + entries = load_pack_entries(PACK_ID) + thin = [e.lemma for e in entries if len(e.semantic_domains) < 2] + assert not thin, f"Entries with <2 semantic domains: {thin[:10]}"