#!/usr/bin/env python3 """Rebuild project files from the Codex Desktop sessions by replaying apply_patch calls and heredoc writes. usage: replay_codex_sessions.py OUT_DIR Streams ~/.codex/sessions and ~/.codex/archived_sessions (never copies them), orders the patches by time and writes the reconstructed files that are NOT on disk now (plus a report of failed hunks) to OUT_DIR. """ import glob import json import os import re import sys ROOT = '/Users/chudy/Library/Application Support/Kodi/addons/plugin.video.milionar/' CODEX = os.path.expanduser('~/.codex/') MARK = 'Kodi/addons/plugin.video.milionar' def sessions(): files = glob.glob(CODEX + 'sessions/*/*/*/*.jsonl') + glob.glob(CODEX + 'archived_sessions/**/*.jsonl', recursive=True) found = [] for path in files: with open(path, 'rb') as handle: hit = False for chunk in iter(lambda: handle.read(1 << 24), b''): if MARK.encode() in chunk: hit = True break if hit: found.append(path) return found def js_strings_after(text, marker): """Yield JS string literals that are the first argument of `marker(`.""" pos = 0 while True: index = text.find(marker, pos) if index < 0: return i = index + len(marker) while i < len(text) and text[i] in ' \n\t': i += 1 if i >= len(text) or text[i] not in '"\'`': pos = i continue quote = text[i] j = i + 1 while j < len(text): if text[j] == '\\': j += 2 continue if text[j] == quote: break j += 1 raw = text[i + 1:j] pos = j + 1 try: if quote == '"': yield json.loads('"' + raw + '"') else: yield raw.encode('utf-8').decode('unicode_escape') if '\\' in raw else raw except Exception: continue def parse_patch(text): """-> list of ops (kind, path, extra) from an apply_patch body.""" ops, lines, i = [], text.split('\n'), 0 current = None while i < len(lines): line = lines[i] if line.startswith('*** Add File: '): current = ['add', line[14:].strip(), []] ops.append(current) elif line.startswith('*** Update File: '): current = ['update', line[17:].strip(), [], None] ops.append(current) elif line.startswith('*** Delete File: '): ops.append(['delete', line[17:].strip(), []]) current = None elif line.startswith('*** Move to: ') and current and current[0] == 'update': current[3] = line[13:].strip() elif line.startswith('*** End Patch') or line.startswith('*** Begin Patch'): current = None elif current is not None: current[2].append(line) i += 1 return ops def rel(path): path = path.strip() return path[len(ROOT):] if path.startswith(ROOT) else path.lstrip('/') if path.startswith('/') else path def apply_update(content, body): hunks, hunk = [], [] for line in body: if line.startswith('@@'): if hunk: hunks.append(hunk) hunk = [] else: hunk.append(line) if hunk: hunks.append(hunk) lines = content.split('\n') cursor = 0 for hunk in hunks: old = [l[1:] for l in hunk if l[:1] in (' ', '-')] new = [l[1:] for l in hunk if l[:1] in (' ', '+')] if not old: # pure insertion at the end lines[len(lines):] = new continue found = -1 for start in list(range(cursor, len(lines) - len(old) + 1)) + list(range(0, cursor)): if all(lines[start + k].rstrip() == old[k].rstrip() for k in range(len(old))): found = start break if found < 0: return None lines[found:found + len(old)] = new cursor = found + len(new) return '\n'.join(lines) def main(out): events = [] for path in sessions(): with open(path, encoding='utf-8', errors='ignore') as handle: for number, line in enumerate(handle): if 'apply_patch' not in line and 'EOF' not in line and "<<'" not in line: continue try: record = json.loads(line) except ValueError: continue payload = record.get('payload') or {} text = payload.get('input') if isinstance(payload.get('input'), str) else payload.get('arguments') if not isinstance(text, str): continue events.append((record.get('timestamp') or '', os.path.basename(path), number, text)) events.sort() vfs, failed, log = {}, [], [] for stamp, name, number, text in events: patches = list(js_strings_after(text, 'apply_patch(')) + [m for m in re.findall(r'\*\*\* Begin Patch.*?\*\*\* End Patch', text, re.S) if '\\n' not in m] for body in patches: if '*** Begin Patch' not in body: continue for op in parse_patch(body): kind, path = op[0], rel(op[1]) if kind == 'add': vfs[path] = '\n'.join(l[1:] if l.startswith('+') else l for l in op[2]).rstrip('\n') + '\n' elif kind == 'delete': vfs.pop(path, None) elif kind == 'update': if path not in vfs: failed.append((stamp, name, path, 'no base')) continue result = apply_update(vfs[path], op[2]) if result is None: failed.append((stamp, name, path, 'hunk mismatch')) continue vfs[path] = result if op[3]: vfs[rel(op[3])] = vfs.pop(path) for match in re.finditer(r"cat > ?(\S+) <<'?(\w+)'?\n(.*?)\n\2\b", text.replace('\\n', '\n'), re.S): vfs[rel(match.group(1).strip('"\''))] = match.group(3) + '\n' os.makedirs(out, exist_ok=True) report = open(os.path.join(out, 'REPORT.txt'), 'w') kept = 0 for path, content in sorted(vfs.items()): if os.path.exists(ROOT + path): report.write('on-disk-now\t' + path + '\n') continue target = os.path.join(out, 'files', path) os.makedirs(os.path.dirname(target), exist_ok=True) with open(target, 'w', encoding='utf-8') as handle: handle.write(content) report.write('recovered\t{}\t{}\n'.format(path, len(content))) kept += 1 for item in failed: report.write('FAILED\t' + '\t'.join(map(str, item)) + '\n') report.close() print('patch events', len(events), 'files in vfs', len(vfs), 'recovered (not on disk)', kept, 'failed hunks', len(failed)) if __name__ == '__main__': main(sys.argv[1])