#!/usr/bin/env python3
"""Build course/atlas.json from newOS4 source. Ground truth for every lesson claim."""
import json, os, re, glob, hashlib
from datetime import datetime, timezone

ROOT = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
OUT = os.path.join(ROOT, 'course', 'atlas.json')
CORE_DIR, EXT_DIR = os.path.join(ROOT, 'modules', 'core'), os.path.join(ROOT, 'modules', 'ext')
SKIP = {'core-bundle.js', 'critical-ext-bundle.js', 'sandbox-worker.js'}

CATEGORY = [
    ('module-fs', 'filesystem'), ('virtual-fs', 'filesystem'), ('module-aiassist', 'ai'), ('module-ollama', 'ai'),
    ('module-memory-graph', 'ai'), ('module-smartctx', 'ai'), ('module-neural-vfs', 'ai'), ('module-world-model', 'ai'),
    ('module-sandbox', 'ai'), ('module-searxng', 'web'), ('module-fetcher', 'web'), ('webdav', 'cloud'),
    ('editor', 'editor'), ('filemgr', 'filesystem'), ('finder', 'filesystem'), ('diff', 'editor'),
    ('cron', 'automation'), ('kanban', 'automation'), ('shell', 'shell'), ('busybox', 'shell'),
    ('theme', 'ui'), ('floating', 'ui'), ('voice', 'ui'), ('three', 'ui'), ('console', 'dev'), ('debug-panel', 'dev'),
    ('wpm', 'system'), ('tutorial', 'learning'), ('skill-pack-core', 'ai'), ('cryptogzip', 'system'),
    ('module-agent-registry', 'agents'), ('module-orchestrator', 'agents'), ('module-leases', 'agents'),
    ('module-chatstudio', 'ai'), ('module-servermon', 'system'), ('module-config', 'system'), ('module-system', 'system'),
    ('course', 'learning'),
]

def category_for(filename, cmd):
    for frag, cat in CATEGORY:
        if frag in filename: return cat
    if cmd.startswith('webdav'): return 'cloud'
    if cmd.startswith('cron') or cmd.startswith('kanban'): return 'automation'
    if cmd.startswith('ai-') or cmd in ('ai', 'ask'): return 'ai'
    return 'general'

def tier_lists():
    rebuild = open(os.path.join(ROOT, 'scripts', 'rebuild-bundles.py')).read()
    def listed(name):
        m = re.search(name + r"\s*=\s*\[([^\]]+)\]", rebuild, re.S)
        return set(re.findall(r"([a-z0-9-]+\.js)", m.group(1))) if m else set()
    loader = open(os.path.join(ROOT, 'modules', 'core', 'module-loader.js')).read()
    m = re.search(r'extKeys\s*=\s*\[([^\]]+)\]', loader)
    ext = set(re.findall(r"'([a-z0-9-]+)'", m.group(1))) if m else set()
    return listed('CORE_FILES'), listed('CRITICAL_EXT_FILES'), ext

CORE_FILES, CRITICAL_EXT_FILES, EXT_KEYS = tier_lists()

def module_tier(filename):
    base = os.path.basename(filename)
    if base in CRITICAL_EXT_FILES: return 'critical-ext-bundle'
    if base in CORE_FILES: return 'core-bundle'
    if base[:-3] in EXT_KEYS: return 'ext-auto'
    return 'ext-manual'

def desc_map(text):
    out = {}
    for m in re.finditer(r"'([a-z0-9][a-z0-9_-]{1,40})'\s*:\s*'((?:[^'\\]|\\.){8,300})'", text):
        cmd, desc = m.group(1), m.group(2)
        if len(desc) < 10 or ' ' not in desc or desc.endswith('.js'):
            continue
        out.setdefault(cmd, desc.replace("\\'", "'"))
    return out

def main():
    commands, modules = {}, {}
    files = sorted(glob.glob(os.path.join(CORE_DIR, '*.js')) + glob.glob(os.path.join(EXT_DIR, '*.js')))
    for path in files:
        base = os.path.basename(path)
        if base in SKIP: continue
        text = open(path, encoding='utf-8', errors='ignore').read()
        cmds = sorted(set(re.findall(r"commands\['([A-Za-z0-9_-]+)'\]\s*=", text)))
        if not cmds: continue
        modname = base[:-3]
        modules[modname] = {'file': os.path.relpath(path, ROOT), 'tier': module_tier(path),
                            'commands': cmds, 'lines': text.count('\n') + 1}
        dm = desc_map(text)
        for c in cmds:
            commands[c] = {'module': modname, 'file': os.path.relpath(path, ROOT), 'tier': module_tier(path),
                           'category': category_for(base, c), 'desc': dm.get(c, ''),
                           'builtin': False, 'example': c}
    term = open(os.path.join(ROOT, 'modules', 'core', 'terminal.js'), encoding='utf-8').read()
    for c in sorted(set(re.findall(r"cmd === '([a-z0-9-]+)'", term))):
        if c in commands: continue
        commands[c] = {'module': 'terminal', 'file': 'modules/core/terminal.js', 'tier': 'core-bundle',
                       'category': 'system', 'desc': '', 'builtin': True, 'example': c}
    idx_path = os.path.join(ROOT, 'modules-index.json')
    mod_by_cmd = {}
    if os.path.exists(idx_path):
        idx = json.load(open(idx_path))
        for m in idx.get('modules', {}).values():
            for c in m.get('commands', []):
                mod_by_cmd.setdefault(c, m)
    for c, e in commands.items():
        if not e['desc']:
            m = mod_by_cmd.get(c) or {}
            e['desc'] = (m.get('description') or f'{c} — command from {e["module"]}')[:300]
    payload = {
        'version': datetime.now(timezone.utc).strftime('%Y-%m-%dT%H:%M:%SZ'),
        'source_hash': hashlib.sha256(''.join(sorted(commands)).encode()).hexdigest()[:16],
        'commands': dict(sorted(commands.items())),
        'modules': dict(sorted(modules.items())),
        'coverage': {'commands': len(commands),
                     'described': sum(1 for e in commands.values() if e['desc']),
                     'modules': len(modules),
                     'tiers': {t: sum(1 for e in commands.values() if e['tier'] == t)
                               for t in ('core-bundle', 'critical-ext-bundle', 'ext-auto', 'ext-manual')}},
    }
    os.makedirs(os.path.dirname(OUT), exist_ok=True)
    tmp = OUT + '.tmp'
    with open(tmp, 'w', encoding='utf-8') as f:
        json.dump(payload, f, ensure_ascii=False, indent=1)
    os.replace(tmp, OUT)
    os.chmod(OUT, 0o644)
    print(json.dumps(payload['coverage'], ensure_ascii=False))

if __name__ == '__main__':
    main()