"""KIN07DAI 型多文件照合程序的前缀检测与前置序列构造测试。

复现根因：_coordinate_multi_file_keys 从 01 记录名推导前缀（R01INNREC→'R01'），
但 COPY REPLACING 把字段前缀改为 SW/SR/SL，导致键对匹配不到、跨文件键协调失效。
修复后应生成前置匹配序列，运行时 merge-join 应覆盖全匹配/单明细匹配等组合。
"""

import sys
import os

sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.dirname(os.path.abspath(__file__)))))

import pytest

from cobol_testgen.__init__ import (
    _coordinate_multi_file_keys,
    _coordinate_tertiary_fd,
    _find_key_pairs_inner,
)


def _group(name):
    return {'name': name, 'level': 1, 'pic': None,
            'pic_info': {'type': 'group', 'digits': 0, 'decimal': 0, 'length': 80, 'signed': False},
            'section': 'FILE', 'is_filler': False, 'redefines': None, 'usage': 'DISPLAY',
            'occurs': 0, 'occurs_depending': None, 'value': None, 'values': None}


def _field(name, level, pic, ftype, length):
    return {'name': name, 'level': level, 'pic': pic,
            'pic_info': {'type': ftype, 'digits': 0, 'decimal': 0, 'length': length, 'signed': False},
            'section': 'FILE', 'is_filler': False, 'redefines': None, 'usage': 'DISPLAY',
            'occurs': 0, 'occurs_depending': None, 'value': None, 'values': None}


def _make_data_fields():
    # 字段顺序与源文件一致：01 记录紧跟其子字段
    return [
        _group('R01INNREC'),
        _field('SWEMP-ID', 3, 'X(008)', 'alphanumeric', 8),
        _field('SWDATE', 3, '9(008)', 'numeric', 8),
        _group('R02INNREC'),
        _field('SREMP-ID', 3, 'X(008)', 'alphanumeric', 8),
        _field('SRWORK-DATE', 3, '9(008)', 'numeric', 8),
        _field('SRLEAVE-TYPE', 3, 'X(002)', 'alphanumeric', 2),
        _group('R03INNREC'),
        _field('SLEMP-ID', 3, 'X(008)', 'alphanumeric', 8),
        _field('SLDATE', 3, '9(008)', 'numeric', 8),
        _field('SLLEAVE-TYPE', 3, 'X(002)', 'alphanumeric', 2),
        _field('SLSTART-TIME', 3, '9(004)', 'numeric', 4),
        _field('SLEND-TIME', 3, '9(004)', 'numeric', 4),
    ]


def _make_records(n=12):
    recs = []
    for i in range(n):
        recs.append({
            'SWEMP-ID': f'R{i:07d}', 'SWDATE': f'{20000101 + i:08d}',
            'SREMP-ID': f'C{i:07d}', 'SRWORK-DATE': f'{20000101 + i:08d}',
            'SRLEAVE-TYPE': '99',
            'SLEMP-ID': f'I{i:07d}', 'SLDATE': f'{20000101 + i:08d}',
            'SLLEAVE-TYPE': '99',
            'SLSTART-TIME': '0822', 'SLEND-TIME': '0922',
        })
    return recs


FILE_SEC = {
    'R01INNFIL': ['R01INNREC'],
    'R02INNFIL': ['R02INNREC'],
    'R03INNFIL': ['R03INNREC'],
    'W01OUTFIL': ['W01OUTREC'],
}
OPEN_DIR = {
    'R01INNFIL': 'INPUT',
    'R02INNFIL': 'INPUT',
    'R03INNFIL': 'INPUT',
    'W01OUTFIL': 'OUTPUT',
}


def test_find_key_pairs_uses_field_prefix_not_record_prefix():
    """字段前缀 SW/SR/SL 应能配对。"""
    fields = _make_data_fields()
    assert _find_key_pairs_inner('SW', 'SR', fields), 'SW-SR 应配对'
    assert _find_key_pairs_inner('SW', 'SL', fields), 'SW-SL 应配对'


def test_coordinate_multi_file_keys_builds_front_sequence():
    """KIN07DAI 形态（记录名 R01INNREC + 字段 SW/SR/SL）下协调应生效：
    记录0 三文件键一致、R03 同键重复块 + 休暇种别注入、键单调递增。"""
    fields = _make_data_fields()
    records = _make_records(12)
    _coordinate_multi_file_keys(
        records, [], fields, {}, FILE_SEC,
        term_types=['normal'] * len(records), open_dir=OPEN_DIR,
    )

    r0 = records[0]
    assert r0['SWEMP-ID'] == r0['SREMP-ID'] == r0['SLEMP-ID'], '记录0 三文件键应一致'
    assert r0['SWDATE'] == r0['SRWORK-DATE'] == r0['SLDATE'], '记录0 三文件日期应一致'

    # R03 同键重复块（休暇种别 01-04 + WHEN OTHER '99'）
    r03_keys = [records[i]['SLEMP-ID'] for i in range(5)]
    assert all(k == r03_keys[0] for k in r03_keys), 'R03 前5条记录键应一致'
    lts = [records[i]['SLLEAVE-TYPE'] for i in range(5)]
    assert lts == ['01', '02', '03', '04', '99'], "休暇种别应注入 01-04+'99', 实际 {lts}".format(lts=lts)

    # 键单调递增（避免 merge-join 指针跳过）
    k0, k1 = records[0]['SWEMP-ID'], records[1]['SWEMP-ID']
    assert k0 < k1, f'前置序列主键应递增: {k0} < {k1}'


def _simulate_merge(R01, R02, R03):
    """模拟 KIN07DAI 的 3 文件 merge-join，返回 (pattern_counts, match_flags)。

    pattern 集合: A=全匹配, B=仅R02, C=仅R03, D=无匹配
    match_flags: (r02_match, r03_match, leave_calc, lt_seen)
    """
    def key(e, d):
        return (str(e).strip(), str(d))

    n1, n2, n3 = len(R01), len(R02), len(R03)
    r1 = r2 = r3 = 0
    e1 = e2 = e3 = False

    def rdR1():
        nonlocal r1, e1
        if r1 < n1:
            r1 += 1
        else:
            e1 = True

    def rdR2():
        nonlocal r2, e2
        if r2 < n2:
            r2 += 1
        else:
            e2 = True

    def rdR3():
        nonlocal r3, e3
        if r3 < n3:
            r3 += 1
        else:
            e3 = True

    rdR1(); rdR2(); rdR3()
    pat = {'A': 0, 'B': 0, 'C': 0, 'D': 0}
    r02m = r03m = calcs = 0
    lt_seen = set()
    while not e1:
        r01k = key(R01[r1 - 1][0], R01[r1 - 1][1])
        ptn = 4
        while not e2:
            if key(R02[r2 - 1][0], R02[r2 - 1][1]) >= r01k:
                break
            rdR2()
        if not e2 and key(R02[r2 - 1][0], R02[r2 - 1][1]) == r01k:
            ptn = 2
            r02m += 1
            rdR2()
        while not e3:
            if key(R03[r3 - 1][0], R03[r3 - 1][1]) >= r01k:
                break
            rdR3()
        if not e3 and key(R03[r3 - 1][0], R03[r3 - 1][1]) == r01k:
            r03m += 1
            while True:
                calcs += 1
                lt_seen.add(R03[r3 - 1][3])
                rdR3()
                if e3:
                    break
                if key(R03[r3 - 1][0], R03[r3 - 1][1]) != r01k:
                    break
            ptn = 1 if ptn == 2 else 3
        pat[['A', 'B', 'C', 'D'][ptn - 1]] += 1
        rdR1()
    return pat, (r02m, r03m, calcs, lt_seen)


def test_coordination_reaches_all_match_combinations_at_runtime():
    """协调后运行时 merge-join 应覆盖全匹配/仅R02/仅R03 三种组合。"""
    fields = _make_data_fields()
    records = _make_records(12)
    _coordinate_multi_file_keys(
        records, [], fields, {}, FILE_SEC,
        term_types=['normal'] * len(records), open_dir=OPEN_DIR,
    )

    R01 = [(r['SWEMP-ID'], r['SWDATE']) for r in records]
    R02 = [(r['SREMP-ID'], r['SRWORK-DATE']) for r in records]
    R03 = [(r['SLEMP-ID'], r['SLDATE'], r['SLLEAVE-TYPE'], r['SLLEAVE-TYPE']) for r in records]
    pat, (r02m, r03m, calcs, lt_seen) = _simulate_merge(R01, R02, R03)

    assert r02m >= 1, 'R02 照合应至少发生 1 次'
    assert r03m >= 1, 'R03 照合应至少发生 1 次'
    assert calcs >= 4, '休暇計算应至少 4 次（EVALUATE 全 WHEN）'
    assert lt_seen >= {'01', '02', '03', '04'}, f'休暇種別应全覆盖, 实际 {lt_seen}'
    assert pat['A'] >= 1, f'全匹配(PATTERN-A)应覆盖, 实际 {pat}'
    assert pat['B'] >= 1, f'仅R02匹配(PATTERN-B)应覆盖, 实际 {pat}'
    assert pat['C'] >= 1, f'仅R03匹配(PATTERN-C)应覆盖, 实际 {pat}'


def test_coordinate_tertiary_fd_long_leave_for_cap():
    """PATTERN-C 用判别块应注入长休暇时长（覆盖 8h 上限 T 分支）。"""
    fields = _make_data_fields()
    records = _make_records(12)
    _coordinate_tertiary_fd(records, ['SW', 'SR', 'SL'], fields)

    # 块2（索引 max(3,5)=5..9）：长时长 0600-2200
    for i in range(5, 10):
        assert records[i]['SLSTART-TIME'] == '0600', f'rec{i} 长休暇开始时间'
        assert records[i]['SLEND-TIME'] == '2200', f'rec{i} 长休暇结束时间'