sync-client.py

#!/usr/bin/env python3
"""
sync-client.py — Push config changes to your MC server's HuggingFace dataset
from your local machine.

The server's background watcher will detect the changes and apply them
within 60 seconds.

Usage:
    # Push a single config file
    python sync-client.py push server.properties

    # Push multiple files
    python sync-client.py push server.properties ops.json

    # Push an entire directory (e.g. plugin configs)
    python sync-client.py push plugins/

    # Pull all server files to local directory
    python sync-client.py pull [output_dir]

    # List what's in the dataset
    python sync-client.py list

    # Watch for changes locally and auto-push
    python sync-client.py watch [directory]

Environment variables:
    HF_DATASET_ID — Your HuggingFace dataset ID
    HF_TOKEN      — Your HuggingFace write token
"""

import os
import sys
import shutil
import tempfile
import subprocess
import time
import json
from pathlib import Path

DATASET_ID = os.environ.get("HF_DATASET_ID", "applepie69/MCP-Server")
HF_TOKEN = os.environ.get("HF_TOKEN", "")
BRANCH = os.environ.get("SYNC_BRANCH", "main")
REPO_URL = f"https://x-access-token:{HF_TOKEN}@huggingface.co/datasets/{DATASET_ID}"

# The files that are considered "config" and safe to push
CONFIG_FILES = {
    "server.properties", "ops.json", "whitelist.json",
    "banned-players.json", "banned-ips.json", "bukkit.yml",
    "spigot.yml", "paper-global.yml", "paper-world-defaults.yml",
    "commands.yml", "help.yml",
}

CONFIG_DIRS = {"plugins"}


def run(cmd, **kwargs):
    """Run a command and return the result."""
    return subprocess.run(cmd, shell=True, capture_output=True, text=True, **kwargs)


def clone_repo(dest):
    """Clone the dataset repo to dest directory."""
    if os.path.exists(dest):
        shutil.rmtree(dest)
    os.makedirs(dest, exist_ok=True)

    result = run(f'git clone --depth 1 --branch {BRANCH} "{REPO_URL}" "{dest}"')
    if result.returncode != 0:
        # Try without branch
        result = run(f'git clone --depth 1 "{REPO_URL}" "{dest}"')
        if result.returncode != 0:
            print(f"ERROR: Failed to clone dataset. Check HF_DATASET_ID and HF_TOKEN.")
            print(f"  stderr: {result.stderr}")
            sys.exit(1)

    # Configure git
    run(f'cd "{dest}" && git config user.email "mc-sync@local" && git config user.name "MC Sync Client"')


def push_to_repo(repo_dir, message="Config update from local"):
    """Commit and push changes in repo_dir."""
    run(f'cd "{repo_dir}" && git add -A')

    result = run(f'cd "{repo_dir}" && git diff --cached --quiet')
    if result.returncode == 0:
        print("No changes to push.")
        return

    run(f'cd "{repo_dir}" && git commit -m "{message}"')
    result = run(f'cd "{repo_dir}" && git push origin HEAD:{BRANCH}')
    if result.returncode != 0:
        print(f"ERROR: Push failed: {result.stderr}")
        sys.exit(1)
    print(f"Pushed: {message}")


def cmd_push(files):
    """Push local files to the dataset."""
    if not DATASET_ID or not HF_TOKEN:
        print("ERROR: Set HF_DATASET_ID and HF_TOKEN environment variables.")
        sys.exit(1)

    with tempfile.TemporaryDirectory() as tmp:
        print(f"Cloning dataset {DATASET_ID}...")
        clone_repo(tmp)

        for f in files:
            src = Path(f)
            if not src.exists():
                print(f"WARNING: {f} not found, skipping.")
                continue

            dst = Path(tmp) / f

            if src.is_dir():
                # Copy directory contents
                if dst.exists():
                    shutil.rmtree(dst)
                shutil.copytree(src, dst)
                print(f"  Staged directory: {f}/")
            else:
                dst.parent.mkdir(parents=True, exist_ok=True)
                shutil.copy2(src, dst)
                print(f"  Staged file: {f}")

        push_to_repo(tmp, f"Config push — {', '.join(files)}")
    print("Done! Server will pick up changes within 60 seconds.")


def cmd_pull(output_dir=None):
    """Pull all files from the dataset to a local directory."""
    if not DATASET_ID or not HF_TOKEN:
        print("ERROR: Set HF_DATASET_ID and HF_TOKEN environment variables.")
        sys.exit(1)

    out = Path(output_dir or "./mc-server-files")
    out.mkdir(parents=True, exist_ok=True)

    with tempfile.TemporaryDirectory() as tmp:
        print(f"Cloning dataset {DATASET_ID}...")
        clone_repo(tmp)

        # Copy everything except .git
        for item in Path(tmp).iterdir():
            if item.name == ".git":
                continue
            dst = out / item.name
            if item.is_dir():
                if dst.exists():
                    shutil.rmtree(dst)
                shutil.copytree(item, dst)
            else:
                shutil.copy2(item, dst)

    print(f"Pulled all files to {out}/")


def cmd_list():
    """List files in the dataset."""
    if not DATASET_ID or not HF_TOKEN:
        print("ERROR: Set HF_DATASET_ID and HF_TOKEN environment variables.")
        sys.exit(1)

    with tempfile.TemporaryDirectory() as tmp:
        clone_repo(tmp)
        result = run(f'cd "{tmp}" && git ls-files')
        if result.stdout.strip():
            print("Files in dataset:")
            for line in result.stdout.strip().split("\n"):
                print(f"  {line}")
        else:
            print("Dataset is empty (no tracked files).")


def cmd_watch(directory=None):
    """Watch a local directory for changes and auto-push."""
    import hashlib

    watch_dir = Path(directory or ".")
    if not watch_dir.exists():
        print(f"ERROR: Directory {watch_dir} does not exist.")
        sys.exit(1)

    print(f"Watching {watch_dir}/ for changes... (Ctrl+C to stop)")

    file_hashes = {}

    def get_hashes():
        hashes = {}
        for f in watch_dir.rglob("*"):
            if f.is_file() and ".git" not in f.parts:
                rel = f.relative_to(watch_dir)
                try:
                    hashes[str(rel)] = hashlib.md5(f.read_bytes()).hexdigest()
                except:
                    pass
        return hashes

    file_hashes = get_hashes()

    while True:
        time.sleep(5)
        new_hashes = get_hashes()

        if new_hashes != file_hashes:
            changed = []
            for k, v in new_hashes.items():
                if file_hashes.get(k) != v:
                    changed.append(k)
            for k in list(file_hashes.keys()):
                if k not in new_hashes:
                    changed.append(f"{k} (deleted)")

            print(f"\nChanges detected: {', '.join(changed)}")
            print("Pushing to dataset...")

            try:
                with tempfile.TemporaryDirectory() as tmp:
                    clone_repo(tmp)

                    # Copy changed files
                    for f in watch_dir.rglob("*"):
                        if f.is_file() and ".git" not in f.parts:
                            rel = f.relative_to(watch_dir)
                            dst = Path(tmp) / rel
                            dst.parent.mkdir(parents=True, exist_ok=True)
                            shutil.copy2(f, dst)

                    push_to_repo(tmp, f"Auto-push: {', '.join(changed[:3])}{'...' if len(changed) > 3 else ''}")
            except Exception as e:
                print(f"Push failed: {e}")

            file_hashes = new_hashes


def main():
    if len(sys.argv) < 2:
        print(__doc__)
        sys.exit(0)

    cmd = sys.argv[1]

    if cmd == "push":
        if len(sys.argv) < 3:
            print("Usage: sync-client.py push <file1> [file2] ...")
            sys.exit(1)
        cmd_push(sys.argv[2:])

    elif cmd == "pull":
        output = sys.argv[2] if len(sys.argv) > 2 else None
        cmd_pull(output)

    elif cmd == "list":
        cmd_list()

    elif cmd == "watch":
        directory = sys.argv[2] if len(sys.argv) > 2 else None
        cmd_watch(directory)

    else:
        print(f"Unknown command: {cmd}")
        print("Commands: push, pull, list, watch")
        sys.exit(1)


if __name__ == "__main__":
    main()