"""Shared computation for the Sol-lane profile-state triage.

Defines ONE classifier and ONE walk over /root/.hermes/profiles/sol.
Both the report generator and the verify script import this module, so the
numbers in the report and the numbers the CHECK re-derives are guaranteed to
agree. Classification is by entry NAME; the walk is os.walk (no symlink
following), so "files" counts every name os.walk reports as a file,
including broken symlinks.
"""

import os
import datetime
from collections import Counter

ROOT = '/root/.hermes/profiles/sol'

# Name -> category. Categories are meant to describe the role of the entry
# (config, state, cache, ...), not to be exhaustive.
_CONFIG_NAMES = {
    'config.yaml', '.env', '.env.bak-20261014T165715Z', 'context_length_cache.yaml',
    'auth.json', 'SOUL.md', '_check_repo_json.py', '_gen_daily_ledger_json.py',
    '_gen_daily_ledger_report.py', '_gen_daily_triage.py', '_gen_daily_triage_fresh.py',
    '_verify_daily_triage.py',
}
_STATE_NAMES = {'state.db', 'state.db-shm', 'state.db-wal', 'state.db.auto-maintenance.lock', 'state.db.quarantine.lock'}
_CACHE_NAMES = {'cache', 'image_cache', 'audio_cache'}
_LOGS_NAMES = {'logs'}
_SKILLS_NAMES = {'skills'}
_PLUGINS_NAMES = {'plugins', 'plugin-update-checks'}
_RUNTIME_NAMES = {'runtime', 'source-checks'}
_CRON_NAMES = {'cron'}
_BACKUP_NAMES = {'backups'}
_KEYS_NAMES = {'keys'}
_VAULT_NAMES = {'vault'}
_HOME_NAMES = {'home'}
_HOOKS_NAMES = {'hooks'}
_LSP_NAMES = {'lsp'}
_MEMORY_NAMES = {'memories'}
_PAIRING_NAMES = {'pairing'}
_PLANS_NAMES = {'plans'}
_PLATFORMS_NAMES = {'platforms'}
_SANDBOXES_NAMES = {'sandboxes'}
_SESSION_NAMES = {'sessions'}
_SKINS_NAMES = {'skins'}
_FORGE_NAMES = {'forge'}
_BIN_NAMES = {'bin', 'forge'}
_WORKSPACE_NAMES = {'workspace'}


def classify(name):
    if name in _CONFIG_NAMES:
        return 'config'
    if name in _STATE_NAMES:
        return 'state'
    if name in _CACHE_NAMES:
        return 'cache'
    if name in _LOGS_NAMES:
        return 'logs'
    if name in _SKILLS_NAMES:
        return 'skills'
    if name in _PLUGINS_NAMES:
        return 'plugins'
    if name in _RUNTIME_NAMES:
        return 'runtime'
    if name in _CRON_NAMES:
        return 'cron'
    if name in _BACKUP_NAMES:
        return 'backups'
    if name in _KEYS_NAMES:
        return 'keys'
    if name in _VAULT_NAMES:
        return 'vault'
    if name in _HOME_NAMES:
        return 'home'
    if name in _HOOKS_NAMES:
        return 'hooks'
    if name in _LSP_NAMES:
        return 'lsp'
    if name in _MEMORY_NAMES:
        return 'memories'
    if name in _PAIRING_NAMES:
        return 'pairing'
    if name in _PLANS_NAMES:
        return 'plans'
    if name in _PLATFORMS_NAMES:
        return 'platforms'
    if name in _SANDBOXES_NAMES:
        return 'sandboxes'
    if name in _SESSION_NAMES:
        return 'sessions'
    if name in _SKINS_NAMES:
        return 'skins'
    if name in _FORGE_NAMES:
        return 'forge'
    if name in _BIN_NAMES:
        return 'bin'
    if name in _WORKSPACE_NAMES:
        return 'workspace'
    return 'other'


def compute(root=ROOT):
    now = datetime.datetime.now(datetime.timezone.utc)
    entries = []
    for dirpath, dirnames, filenames in os.walk(root):
        for d in dirnames:
            full = os.path.join(dirpath, d)
            st = os.lstat(full)
            entries.append({
                'path': os.path.relpath(full, root),
                'type': 'dir',
                'category': classify(d),
                'mtime': datetime.datetime.fromtimestamp(st.st_mtime, datetime.timezone.utc).isoformat(),
                'age_days': (now - datetime.datetime.fromtimestamp(st.st_mtime, datetime.timezone.utc)).total_seconds() / 86400.0,
            })
        for f in filenames:
            full = os.path.join(dirpath, f)
            st = os.lstat(full)
            entries.append({
                'path': os.path.relpath(full, root),
                'type': 'file',
                'category': classify(f),
                'mtime': datetime.datetime.fromtimestamp(st.st_mtime, datetime.timezone.utc).isoformat(),
                'age_days': (now - datetime.datetime.fromtimestamp(st.st_mtime, datetime.timezone.utc)).total_seconds() / 86400.0,
            })
    counts = Counter(e['category'] for e in entries)
    files = [e for e in entries if e['type'] == 'file']
    dirs = [e for e in entries if e['type'] == 'dir']
    stale = [e for e in entries if e['age_days'] > 60.0]
    return {
        'now': now.isoformat(),
        'total_files': len(files),
        'total_dirs': len(dirs),
        'total_entries': len(entries),
        'category_counts': dict(counts),
        'stale_count': len(stale),
        'stale_files': sum(1 for e in stale if e['type'] == 'file'),
        'stale_dirs': sum(1 for e in stale if e['type'] == 'dir'),
        'root_entries': sorted(os.listdir(root)),
    }


if __name__ == '__main__':
    import json
    print(json.dumps(compute(), indent=2, default=str))