Create autobackup.py

autobackup.py

@@ -0,0 +1,52 @@
+import os, subprocess, time, threading
+
+def run_backup():
+    repo_id = os.environ.get("BACKUP_REPO")
+    hf_token = os.environ.get("HF_TOKEN")
+    repo_type = "space" if "/" in repo_id else "dataset"  # auto-detect
+
+    if not repo_id or not hf_token:
+        print("❌ BACKUP_REPO or HF_TOKEN missing, skipping backup")
+        return
+
+    env = os.environ.copy()
+    env["HF_HOME"] = "/tmp/hf_cache"
+    env["XDG_CACHE_HOME"] = "/tmp/xdg_cache"
+    env["TMPDIR"] = "/tmp"
+
+    os.makedirs(env["HF_HOME"], exist_ok=True)
+    os.makedirs(env["XDG_CACHE_HOME"], exist_ok=True)
+    os.makedirs(env["TMPDIR"], exist_ok=True)
+
+    local_path = "/workspace"
+
+    cmd = [
+        "hf", "upload",
+        repo_id,
+        local_path,
+        "--repo-type", repo_type,
+        "--token", hf_token,
+        "--ignore-patterns", "*.git", "__pycache__",
+    ]
+
+    process = subprocess.Popen(
+        cmd,
+        stdout=subprocess.PIPE,
+        stderr=subprocess.STDOUT,
+        env=env,
+        text=True,
+        cwd="/tmp",
+    )
+
+    for line in iter(process.stdout.readline, ''):
+        print("🔄", line.strip())
+
+def schedule_backups():
+    while True:
+        print("⏳ Running auto-backup...")
+        run_backup()
+        print("✅ Backup finished, sleeping 45m...")
+        time.sleep(45 * 60)
+
+# Run in background
+threading.Thread(target=schedule_backups, daemon=True).start()