from pathlib import Path
import hashlib
import json
import shutil

ROOT = Path(__file__).resolve().parent
REPO = ROOT.parent.parent


def write_source(relative, text):
    path = ROOT / relative
    assert not path.exists(), path
    path.parent.mkdir(parents=True, exist_ok=True)
    path.write_bytes(text.replace('\r\n', '\n').replace('\n', '\r\n').encode('utf-8'))


files = {
    'project.arcadia': '''arcadia 7;
project navigation_trial "Navigation — 救援" id "9e8f0000-0000-4000-8000-000000000001" {
    description "Prose @assess must not be a lexical use.";
    include "救援/assess.arcadia";
    include "救援/confirm.arcadia";
    include "links.arcadia";
}
''',
    '救援/assess.arcadia': '''// Comment @assess must not be a lexical use.
oa_activity assess "Assess 救援 request" id "9e8f0000-0000-4000-8000-000000000002" {
    description "Keep @assess as prose, not a use.";
}
''',
    '救援/confirm.arcadia': '''oa_activity confirm "Confirm 救援 response" id "9e8f0000-0000-4000-8000-000000000003" {}
''',
    'links.arcadia': '''interaction handoff "Assessment handoff" id "9e8f0000-0000-4000-8000-000000000004" {
    description "救援 @assess is prose"; source @assess; target @confirm;
    // Another comment @assess is excluded.
}
''',
}
for name, text in files.items():
    write_source('valid-baseline/' + name, text)
shutil.copytree(ROOT / 'valid-baseline', ROOT / 'valid-work')
for case in ['ambiguous-open', 'missing-include-open']:
    for name, text in files.items():
        if case == 'ambiguous-open' and name == '救援/confirm.arcadia':
            text += 'oa_activity assess "Duplicate draft assessment" id "9e8f0000-0000-4000-8000-000000000005" {}\n'
        if case == 'missing-include-open' and name == 'project.arcadia':
            text = text.replace('    include "links.arcadia";', '    include "links.arcadia";\n    include "missing/not-loaded.arcadia";')
        write_source(case + '/' + name, text)

# A realistic existing teaching model, copied without changing source bytes.
shutil.copytree(REPO / 'examples/release1', ROOT / 'release1-baseline')
shutil.copytree(ROOT / 'release1-baseline', ROOT / 'release1-work')


def position(raw, offset):
    line_start = raw.rfind(b'\n', 0, offset) + 1
    return {'byte_offset': offset, 'line_1based': raw.count(b'\n', 0, offset) + 1,
            'utf8_byte_column_1based': offset - line_start + 1}


locations = []
for folder in ['valid-baseline', 'ambiguous-open', 'missing-include-open']:
    for path in sorted((ROOT / folder).rglob('*.arcadia')):
        raw = path.read_bytes()
        text = raw.decode('utf-8')
        assert b'\n' not in raw.replace(b'\r\n', b'')
        for alias in ['assess', 'confirm']:
            token = alias.encode('ascii')
            start = 0
            while (found := raw.find(token, start)) != -1:
                start = found + len(token)
                line_start = raw.rfind(b'\n', 0, found) + 1
                prefix = raw[line_start:found]
                # Explicit fixture expectations, independent of the application lexer.
                if prefix == b'oa_activity ':
                    kind = 'declaration'
                elif prefix.endswith(b'source @') or prefix.endswith(b'target @'):
                    kind = 'lexical-use'
                elif found > 0 and raw[found-1:found] == b'@':
                    kind = 'excluded-comment-or-string'
                else:
                    continue
                locations.append({'fixture': folder, 'path': path.relative_to(ROOT / folder).as_posix(),
                                  'alias': alias, 'kind': kind, 'span': [found, start],
                                  **position(raw, found)})
expected = {'encoding': 'UTF-8 without BOM; every authored fixture newline CRLF',
            'valid_totals': {'assess': {'declarations': 1, 'lexical_uses': 1},
                             'confirm': {'declarations': 1, 'lexical_uses': 1}},
            'ambiguous_totals': {'assess': {'declarations': 2, 'lexical_uses': 1}},
            'locations': locations}
(ROOT / 'expected-locations.json').write_text(json.dumps(expected, ensure_ascii=False, indent=2) + '\n', encoding='utf-8')
manifest = [{'path': p.relative_to(ROOT).as_posix(), 'bytes': p.stat().st_size,
             'sha256': hashlib.sha256(p.read_bytes()).hexdigest()}
            for folder in ['valid-baseline', 'valid-work', 'ambiguous-open', 'missing-include-open', 'release1-baseline', 'release1-work']
            for p in sorted((ROOT / folder).rglob('*')) if p.is_file()]
(ROOT / 'fixture-manifest.json').write_text(json.dumps(manifest, ensure_ascii=False, indent=2) + '\n', encoding='utf-8')
print(json.dumps({'files': len(manifest), 'bytes': sum(p['bytes'] for p in manifest),
                  'expected_locations': len(locations)}))