import requests
import subprocess
import os
import logging
import re
import unicodedata
from datetime import datetime

#This code does gitea to github mirroring with description sanitization

# ----------------- CONFIGURATION -----------------
GITEA_URL = ""
GITEA_TOKEN = ""
GITEA_USER = ""

GITHUB_TOKEN = ""
GITHUB_USER = ""  # your GitHub username/org

CLONE_DIR = "./gitea_repos"  # Temporary folder for cloning
# -------------------------------------------------

# Setup logging
log_filename = f"mirror_log_{datetime.now().strftime('%Y%m%d_%H%M%S')}.log"
logging.basicConfig(
    level=logging.INFO,
    format="%(asctime)s [%(levelname)s] %(message)s",
    handlers=[
        logging.FileHandler(log_filename),
        logging.StreamHandler()
    ]
)

headers_gitea = {"Authorization": f"token {GITEA_TOKEN}"}
os.makedirs(CLONE_DIR, exist_ok=True)

# ----------------- SANITIZER -----------------
def sanitize_description(text: str) -> str:
    if not text:
        return ""

    # Remove ASCII control chars except: tab(9), LF(10), CR(13)
    text = re.sub(r"[\x00-\x08\x0B\x0C\x0E-\x1F]", " ", text)

    # Remove Unicode control characters from all languages
    text = "".join(ch if unicodedata.category(ch) != "Cc" else " " for ch in text)

    # Collapse multiple spaces
    text = re.sub(r"\s+", " ", text).strip()
    return text
# -------------------------------------------------

# Step 1: Get all public Gitea repos
def get_gitea_repos():
    logging.info("Fetching repositories from Gitea...")
    repos = []
    page = 1
    while True:
        url = f"{GITEA_URL}/api/v1/users/{GITEA_USER}/repos?page={page}&limit=100"
        r = requests.get(url, headers=headers_gitea)
        if r.status_code != 200:
            logging.error(f"Failed to fetch Gitea repos: {r.text}")
            break
        data = r.json()
        if not data:
            break
        public_repos = [repo for repo in data if not repo.get("private", False)]
        repos.extend(public_repos)
        page += 1
    logging.info(f"Found {len(repos)} public repositories on Gitea.")
    return repos

# Step 2: Create GitHub repo
def create_github_repo(repo_name, description=""):
    safe_desc = sanitize_description(description)

    url = f"https://api.github.com/user/repos"
    payload = {"name": repo_name, "private": False, "description": safe_desc}
    headers = {"Authorization": f"token {GITHUB_TOKEN}"}

    r = requests.post(url, json=payload, headers=headers)
    if r.status_code == 201:
        logging.info(f"GitHub repo created: {repo_name}")
        return True
    elif r.status_code == 422 and "already exists" in r.text:
        logging.info(f"GitHub repo already exists: {repo_name}")
        return True
    else:
        logging.error(f"Failed to create GitHub repo {repo_name}: {r.status_code} - {r.text}")
        return False

# Step 3: Mirror from Gitea → GitHub
def mirror_repo(repo_name, clone_url):
    local_path = os.path.join(CLONE_DIR, repo_name)
    try:
        if os.path.exists(local_path):
            logging.info(f"Repo {repo_name} already cloned. Fetching updates...")
            subprocess.run(["git", "--git-dir", f"{local_path}/.git", "fetch", "--all"], check=True)
        else:
            logging.info(f"Cloning {repo_name} from Gitea...")
            subprocess.run(["git", "clone", "--mirror", clone_url, local_path], check=True)

        github_url = f"https://{GITHUB_TOKEN}@github.com/{GITHUB_USER}/{repo_name}.git"
        logging.info(f"Pushing {repo_name} to GitHub...")
        subprocess.run(["git", "--git-dir", local_path, "push", "--mirror", github_url], check=True)
        logging.info(f"✅ Finished mirroring {repo_name}")

    except subprocess.CalledProcessError as e:
        logging.error(f"❌ Error mirroring {repo_name}: {e}")

def main():
    gitea_repos = get_gitea_repos()

    for repo in gitea_repos:
        repo_name = repo["name"]
        description = repo.get("description", "")
        clone_url = repo["clone_url"]

        logging.info(f"Processing repo: {repo_name}")
        if create_github_repo(repo_name, description):
            mirror_repo(repo_name, clone_url)

if __name__ == "__main__":
    main()
    logging.info(f"All done! Log saved to {log_filename}")