# -*- coding: utf-8 -*-
"""gen_parity_es.py — emit an El parity program that checks the .el Spanish
morphology against the validated Python (UniMorph-backed) realizer.

Python is the ORACLE. For every lemma in the held-out inventory we embed the
Python-produced form, call the corresponding .el function, and the El program
prints PASS/FAIL per category. Aggregation is done in bash (grep -c), so no
El-side mutable counters are needed.

Categories:
  Vpres  verb present-ind-3sg      es_conjugate(v,present,third,singular)
  Vpret  verb preterite-3sg        es_conjugate(v,past,third,singular)
  Vpart  past participle           es_participle(v)
  Npl    noun plural               es_pluralize(n)
  Gheur  noun gender HEURISTIC     es_gender(n)                  [exposes the bug]
  Aheur  def article via heuristic es_agree_article(n,true,sg)   [inherits the bug]
  Avocab def article via REAL g    es_article_for_gender(realg,...) [the fix]
  Ampl   adj masc-plural           es_inflect_adj(a,m,plural)
  Afsg   adj fem-singular          es_inflect_adj(a,f,singular)
  Ctr    contraction del/al        es_contract(prep, np)
"""
import morphology_es_full as M
import realizer_es as R
from gen_elp_es import _collect_from_specs, _CORE_VERBS, _CORE_NOUNS, _CORE_ADJS


def _esc(s):
    return str(s).replace('\\', '\\\\').replace('"', '\\"')


def _check(cat, lemma, el_call, expected):
    return (f'  es_check("{cat}", "{_esc(lemma)}", {el_call}, "{_esc(expected)}")')


def main(out_path):
    v_specs, n_specs, a_specs = _collect_from_specs()
    verbs = sorted(set(_CORE_VERBS) | v_specs)
    nouns = sorted(set(_CORE_NOUNS) | n_specs)
    adjs = sorted(set(_CORE_ADJS) | a_specs)

    lines = []
    lines.append("// gen'd parity checks — Python oracle embedded, El functions called.")
    lines.append("fn es_check(cat: String, lemma: String, got: String, exp: String) {")
    lines.append('    if str_eq(got, exp) {')
    lines.append('        println("PASS " + cat)')
    lines.append('    } else {')
    lines.append('        println("FAIL " + cat + " " + lemma + " got=" + got + " exp=" + exp)')
    lines.append('    }')
    lines.append("}")
    lines.append("")
    lines.append("fn es_parity() {")

    # verbs
    for v in verbs:
        p0 = M.conjugate(v, "ind", "present", "third", "singular")[0]
        p1 = M.conjugate(v, "ind", "preterite", "third", "singular")[0]
        pp = M.participle(v)[0]
        lines.append(_check("Vpres", v, f'es_conjugate("{_esc(v)}", "present", "third", "singular")', p0))
        lines.append(_check("Vpret", v, f'es_conjugate("{_esc(v)}", "past", "third", "singular")', p1))
        lines.append(_check("Vpart", v, f'es_participle("{_esc(v)}")', pp))

    # nouns
    for n in nouns:
        pl = M.inflect_noun(n, "plural")[0]
        g = M.noun_gender(n)                    # REAL lexicon gender
        art = R._article(g, "singular", "def", n)  # oracle article from real gender
        lines.append(_check("Npl", n, f'es_pluralize("{_esc(n)}")', pl))
        lines.append(_check("Gheur", n, f'es_gender("{_esc(n)}")', g))
        lines.append(_check("Aheur", n, f'es_agree_article("{_esc(n)}", "true", "singular")', art))
        lines.append(_check("Avocab", n, f'es_article_for_gender("{_esc(g)}", "{_esc(n)}", "true", "singular")', art))

    # adjectives
    for a in adjs:
        mpl = M.inflect_adj(a, "m", "plural")[0]
        fsg = M.inflect_adj(a, "f", "singular")[0]
        lines.append(_check("Ampl", a, f'es_inflect_adj("{_esc(a)}", "m", "plural")', mpl))
        lines.append(_check("Afsg", a, f'es_inflect_adj("{_esc(a)}", "f", "singular")', fsg))

    # contractions (mandatory)
    ctr_cases = [("de", "el día"), ("a", "el hombre"), ("de", "el mundo"),
                 ("a", "el país"), ("en", "el mar"), ("de", "el año"),
                 ("a", "la casa"), ("de", "la ciudad")]
    for prep, np in ctr_cases:
        exp = R._contract(prep, np)
        lines.append(_check("Ctr", prep + "+" + np, f'es_contract("{_esc(prep)}", "{_esc(np)}")', exp))

    lines.append("}")
    lines.append("")
    lines.append("fn main() {")
    lines.append("    es_parity()")
    lines.append("}")

    with open(out_path, "w", encoding="utf-8") as fh:
        fh.write("\n".join(lines) + "\n")
    print(f"wrote {out_path}  ({len(verbs)} verbs, {len(nouns)} nouns, {len(adjs)} adjs)")


if __name__ == "__main__":
    import sys
    main(sys.argv[1] if len(sys.argv) > 1 else "parity_es_checks.el")