scripts/sync_hf.py

#!/usr/bin/env python3
"""
OpenClaw HF Spaces Persistence — Full Directory Sync
=====================================================

Simplified persistence: upload/download the entire ~/.openclaw directory
as-is to/from a Hugging Face Dataset repo.

- Startup:  snapshot_download  →  ~/.openclaw
- Periodic: upload_folder      →  dataset openclaw_data/
- Shutdown: final upload_folder →  dataset openclaw_data/
"""

import os
import sys
import time
import threading
import subprocess
import signal
import json
import shutil
import tempfile
import traceback
import re
from pathlib import Path
from datetime import datetime
# Set timeout BEFORE importing huggingface_hub
os.environ.setdefault("HF_HUB_DOWNLOAD_TIMEOUT", "300")
os.environ.setdefault("HF_HUB_UPLOAD_TIMEOUT", "600")

from huggingface_hub import HfApi, snapshot_download

# ── Logging helper ──────────────────────────────────────────────────────────

class TeeLogger:
    """Duplicate output to stream and file."""
    def __init__(self, filename, stream):
        self.stream = stream
        self.file = open(filename, "a", encoding="utf-8")
    def write(self, message):
        self.stream.write(message)
        self.file.write(message)
        self.flush()
    def flush(self):
        self.stream.flush()
        self.file.flush()
    def fileno(self):
        return self.stream.fileno()

# ── Configuration ───────────────────────────────────────────────────────────

HF_REPO_ID = os.environ.get("OPENCLAW_DATASET_REPO", "tao-shen/openclaw")
HF_TOKEN   = os.environ.get("HF_TOKEN")
OPENCLAW_HOME = Path.home() / ".openclaw"
APP_DIR       = Path("/app/openclaw")

# Use ".openclaw" - directly read/write the .openclaw folder in dataset
DATASET_PATH = ".openclaw"

TELEGRAM_BOT_TOKEN  = os.environ.get("TELEGRAM_BOT_TOKEN", "")
TELEGRAM_BOT_NAME   = os.environ.get("TELEGRAM_BOT_NAME", "opentauronbot")
TELEGRAM_ALLOW_USER = os.environ.get("TELEGRAM_ALLOW_USER", "taoshen1")

SYNC_INTERVAL = int(os.environ.get("SYNC_INTERVAL", "120"))

# Setup logging
log_dir = OPENCLAW_HOME / "workspace"
log_dir.mkdir(parents=True, exist_ok=True)
sys.stdout = TeeLogger(log_dir / "sync.log", sys.stdout)
sys.stderr = sys.stdout

# ── Sync Manager ────────────────────────────────────────────────────────────

class OpenClawFullSync:
    """Upload/download the entire ~/.openclaw directory to HF Dataset."""

    def __init__(self):
        if not HF_TOKEN:
            print("[SYNC] WARNING: HF_TOKEN not set. Persistence disabled.")
            self.enabled = False
            return
        self.enabled = True
        self.api = HfApi(token=HF_TOKEN)
        self._ensure_repo()

    # ── Repo management ────────────────────────────────────────────────

    def _ensure_repo(self):
        try:
            self.api.repo_info(repo_id=HF_REPO_ID, repo_type="dataset")
            print(f"[SYNC] Dataset repo found: {HF_REPO_ID}")
        except Exception:
            print(f"[SYNC] Creating dataset repo: {HF_REPO_ID}")
            self.api.create_repo(repo_id=HF_REPO_ID, repo_type="dataset", private=True)

    # ── Restore (startup) ─────────────────────────────────────────────

    def load_from_repo(self):
        """Download from dataset → ~/.openclaw"""
        if not self.enabled:
            return
        print(f"[SYNC] ▶ Restoring ~/.openclaw from dataset {HF_REPO_ID} ...")
        OPENCLAW_HOME.mkdir(parents=True, exist_ok=True)

        try:
            files = self.api.list_repo_files(repo_id=HF_REPO_ID, repo_type="dataset")
            openclaw_files = [f for f in files if f.startswith(f"{DATASET_PATH}/")]
            if not openclaw_files:
                print(f"[SYNC] No {DATASET_PATH}/ folder in dataset. Starting fresh.")
                self._ensure_default_config()
                self._ensure_telegram_credentials()
                return

            print(f"[SYNC] Found {len(openclaw_files)} files under {DATASET_PATH}/ in dataset")

            with tempfile.TemporaryDirectory() as tmpdir:
                snapshot_download(
                    repo_id=HF_REPO_ID,
                    repo_type="dataset",
                    allow_patterns=f"{DATASET_PATH}/**",
                    local_dir=tmpdir,
                    token=HF_TOKEN,
                )
                downloaded_root = Path(tmpdir) / DATASET_PATH
                if downloaded_root.exists():
                    for item in downloaded_root.rglob("*"):
                        if item.is_file():
                            rel = item.relative_to(downloaded_root)
                            dest = OPENCLAW_HOME / rel
                            dest.parent.mkdir(parents=True, exist_ok=True)
                            shutil.copy2(str(item), str(dest))
                    print("[SYNC] ✓ Restore completed.")
                else:
                    print("[SYNC] Downloaded snapshot but dir not found. Starting fresh.")

        except Exception as e:
            print(f"[SYNC] ✗ Restore failed: {e}")
            traceback.print_exc()

        # Patch config & telegram after restore
        self._patch_config()
        self._ensure_telegram_credentials()
        self._debug_list_files()

    # ── Save (periodic + shutdown) ─────────────────────────────────────

    def save_to_repo(self):
        """Upload entire ~/.openclaw directory → dataset (all files, no filtering)"""
        if not self.enabled:
            return
        if not OPENCLAW_HOME.exists():
            print("[SYNC] ~/.openclaw does not exist, nothing to save.")
            return

        print(f"[SYNC] ▶ Uploading ~/.openclaw → dataset {HF_REPO_ID}/{DATASET_PATH}/ ...")

        try:
            # Log what will be uploaded
            total_size = 0
            file_count = 0
            for root, dirs, fls in os.walk(OPENCLAW_HOME):
                for fn in fls:
                    fp = os.path.join(root, fn)
                    sz = os.path.getsize(fp)
                    total_size += sz
                    file_count += 1
                    rel = os.path.relpath(fp, OPENCLAW_HOME)
                    print(f"[SYNC]   uploading: {rel} ({sz} bytes)")
            print(f"[SYNC] Uploading: {file_count} files, {total_size} bytes total")

            if file_count == 0:
                print("[SYNC] Nothing to upload.")
                return

            # Upload directory, excluding large log files that trigger LFS errors
            self.api.upload_folder(
                folder_path=str(OPENCLAW_HOME),
                path_in_repo=DATASET_PATH,
                repo_id=HF_REPO_ID,
                repo_type="dataset",
                token=HF_TOKEN,
                commit_message=f"Sync .openclaw — {datetime.now().isoformat()}",
                ignore_patterns=["*.log", "*.log.*"],  # Exclude logs (sync.log is 60MB+, triggers LFS errors)
            )
            print(f"[SYNC] ✓ Upload completed at {datetime.now().isoformat()}")

            # Verify
            try:
                files = self.api.list_repo_files(repo_id=HF_REPO_ID, repo_type="dataset")
                oc_files = [f for f in files if f.startswith(f"{DATASET_PATH}/")]
                print(f"[SYNC] Dataset now has {len(oc_files)} files under {DATASET_PATH}/")
                for f in oc_files[:30]:
                    print(f"[SYNC]   {f}")
                if len(oc_files) > 30:
                    print(f"[SYNC]   ... and {len(oc_files) - 30} more")
            except Exception:
                pass

        except Exception as e:
            print(f"[SYNC] ✗ Upload failed: {e}")
            traceback.print_exc()

    # ── Config helpers ─────────────────────────────────────────────────

    def _ensure_default_config(self):
        config_path = OPENCLAW_HOME / "openclaw.json"
        if config_path.exists():
            return
        default_src = Path(__file__).parent / "openclaw.json.default"
        if default_src.exists():
            shutil.copy2(str(default_src), str(config_path))
            print("[SYNC] Created openclaw.json from default template")
        else:
            with open(config_path, "w") as f:
                json.dump({
                    "gateway": {
                        "mode": "local", "bind": "lan", "port": 7860,
                        "trustedProxies": ["0.0.0.0/0"],
                        "controlUi": {
                            "allowInsecureAuth": True,
                            "allowedOrigins": [
                                "https://tao-shen-openclaw-ai.hf.space",
                                "https://huggingface.co"
                            ]
                        }
                    },
                    "session": {"scope": "global"},
                    "models": {"mode": "merge", "providers": {}},
                    "agents": {"defaults": {"workspace": "~/.openclaw/workspace"}}
                }, f)
            print("[SYNC] Created minimal openclaw.json")

    def _patch_config(self):
        """Ensure critical settings after restore."""
        config_path = OPENCLAW_HOME / "openclaw.json"
        if not config_path.exists():
            self._ensure_default_config()
            return

        print("[SYNC] Patching configuration...")
        try:
            with open(config_path, "r") as f:
                data = json.load(f)
            print("[SYNC] Config parsed OK.")
        except (json.JSONDecodeError, Exception) as e:
            # Config is corrupt — back up and start fresh
            print(f"[SYNC] Config JSON is corrupt: {e}")
            backup = config_path.with_suffix(f".corrupt_{int(time.time())}")
            try:
                import shutil
                shutil.copy2(config_path, backup)
                print(f"[SYNC] Backed up corrupt config to {backup.name}")
            except Exception:
                pass
            data = {}
            print("[SYNC] Starting from clean config.")

        try:
            # Remove /dev/null from plugins.locations
            if "plugins" in data and isinstance(data.get("plugins"), dict):
                locs = data["plugins"].get("locations", [])
                if isinstance(locs, list) and "/dev/null" in locs:
                    data["plugins"]["locations"] = [l for l in locs if l != "/dev/null"]

            # Force full gateway config for HF Spaces
            # Note: Dockerfile injects "openclaw-space-default" token into Control UI,
            # so we MUST set it here to match what the browser sends.
            data["gateway"] = {
                "mode": "local",
                "bind": "lan",
                "port": 7860,
                "auth": {"token": "openclaw-space-default"},
                "trustedProxies": ["0.0.0.0/0"],
                "controlUi": {
                    "allowInsecureAuth": True,
                    "allowedOrigins": [
                        "https://tao-shen-openclaw-ai.hf.space",
                        "https://huggingface.co"
                    ]
                }
            }
            print("[SYNC] Set gateway config (auth=default, trustedProxies=all)")

            # Ensure agents defaults
            data.setdefault("agents", {}).setdefault("defaults", {}).setdefault("model", {})
            data.setdefault("session", {})["scope"] = "global"

            # Force OpenRouter provider
            data.setdefault("models", {}).setdefault("providers", {})
            data["models"]["providers"]["openrouter"] = {
                "baseUrl": "https://openrouter.ai/api/v1",
                "apiKey": "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",
                "api": "openai-completions",
                "models": [
                    {"id": "stepfun/step-3.5-flash:free", "name": "Step-3.5-Flash (Free)"},
                    {"id": "deepseek/deepseek-chat:free", "name": "DeepSeek V3 (Free)"}
                ]
            }
            # Remove old gemini provider if present
            data["models"]["providers"].pop("gemini", None)
            data["agents"]["defaults"]["model"]["primary"] = "openrouter/stepfun/step-3.5-flash:free"

            # Telegram plugin
            data.setdefault("plugins", {}).setdefault("entries", {})
            if "telegram" not in data["plugins"]["entries"]:
                data["plugins"]["entries"]["telegram"] = {"enabled": True}
            elif isinstance(data["plugins"]["entries"]["telegram"], dict):
                data["plugins"]["entries"]["telegram"]["enabled"] = True

            with open(config_path, "w") as f:
                json.dump(data, f, indent=2)
            print("[SYNC] Config patched and saved.")

            # Verify write
            with open(config_path, "r") as f:
                verify_data = json.load(f)
                gw = verify_data.get("gateway", {})
                providers = list(verify_data.get("models", {}).get("providers", {}).keys())
                primary = verify_data.get("agents", {}).get("defaults", {}).get("model", {}).get("primary")
                print(f"[SYNC] VERIFY: gateway.port={gw.get('port')}, providers={providers}, primary={primary}")

        except Exception as e:
            print(f"[SYNC] Failed to patch config: {e}")
            traceback.print_exc()

    def _ensure_telegram_credentials(self):
        """Configure Telegram bot token and allowed users."""
        creds_dir = OPENCLAW_HOME / "credentials"
        creds_dir.mkdir(parents=True, exist_ok=True)

        if TELEGRAM_BOT_TOKEN:
            bot_file = creds_dir / "telegram-bot-token.json"
            with open(bot_file, "w") as f:
                json.dump({"token": TELEGRAM_BOT_TOKEN, "bot": TELEGRAM_BOT_NAME}, f, indent=2)
            print(f"[SYNC] Telegram bot configured: {TELEGRAM_BOT_NAME}")

        allow_file = creds_dir / "telegram-allowFrom.json"
        if not allow_file.exists():
            with open(allow_file, "w") as f:
                json.dump([TELEGRAM_ALLOW_USER], f, indent=2)
            print(f"[SYNC] Created telegram-allowFrom.json for {TELEGRAM_ALLOW_USER}")
        else:
            try:
                with open(allow_file, "r") as f:
                    data = json.load(f)
                if not isinstance(data, list):
                    data = [TELEGRAM_ALLOW_USER]
                elif TELEGRAM_ALLOW_USER not in data:
                    data.append(TELEGRAM_ALLOW_USER)
                with open(allow_file, "w") as f:
                    json.dump(data, f, indent=2)
            except Exception:
                with open(allow_file, "w") as f:
                    json.dump([TELEGRAM_ALLOW_USER], f, indent=2)

    def _debug_list_files(self):
        print(f"[SYNC] Local ~/.openclaw tree:")
        try:
            count = 0
            for root, dirs, files in os.walk(OPENCLAW_HOME):
                dirs[:] = [d for d in dirs if d not in {".cache", "node_modules", "__pycache__"}]
                for name in sorted(files):
                    rel = os.path.relpath(os.path.join(root, name), OPENCLAW_HOME)
                    print(f"[SYNC]   {rel}")
                    count += 1
                    if count > 50:
                        print("[SYNC]   ... (truncated)")
                        return
        except Exception as e:
            print(f"[SYNC] listing failed: {e}")

    # ── Background sync loop ──────────────────────────────────────────

    def background_sync_loop(self, stop_event):
        print(f"[SYNC] Background sync started (interval={SYNC_INTERVAL}s)")
        while not stop_event.is_set():
            if stop_event.wait(timeout=SYNC_INTERVAL):
                break
            print(f"[SYNC] ── Periodic sync triggered at {datetime.now().isoformat()} ──")
            self.save_to_repo()

    # ── Application runner ─────────────────────────────────────────────

    def run_openclaw(self):
        log_file = OPENCLAW_HOME / "workspace" / "startup.log"
        log_file.parent.mkdir(parents=True, exist_ok=True)
        cmd = f"node dist/entry.js gateway 2>&1 | tee -a {log_file}"
        print(f"[SYNC] Launching: {cmd}")
        return subprocess.Popen(cmd, shell=True, cwd=str(APP_DIR),
                                stdout=sys.stdout, stderr=sys.stderr)

# ── Main ────────────────────────────────────────────────────────────────────

def main():
    sync = OpenClawFullSync()

    # 1. Restore
    sync.load_from_repo()

    # 2. Background sync
    stop_event = threading.Event()
    t = threading.Thread(target=sync.background_sync_loop, args=(stop_event,), daemon=True)
    t.start()

    # 3. Start application
    process = sync.run_openclaw()

    # Signal handler
    def handle_signal(sig, frame):
        print(f"\n[SYNC] Signal {sig} received. Shutting down...")
        stop_event.set()
        if process:
            process.terminate()
            try:
                process.wait(timeout=5)
            except subprocess.TimeoutExpired:
                process.kill()
        print("[SYNC] Final sync...")
        sync.save_to_repo()
        sys.exit(0)

    signal.signal(signal.SIGINT, handle_signal)
    signal.signal(signal.SIGTERM, handle_signal)

    # Wait
    exit_code = process.wait()
    print(f"[SYNC] OpenClaw exited with code {exit_code}")
    stop_event.set()
    print("[SYNC] Final sync...")
    sync.save_to_repo()
    sys.exit(exit_code)


if __name__ == "__main__":
    main()