update prepare_environment, fix symlinks

README.md

@@ -25,12 +25,12 @@ sudo apt-get update
 sudo apt-get install sox ffmpeg
 ```
 
-
 ## Backup and Restore Volume
 
 p/s: Need to mount backup folder to container
 
 ### Backup
+
 ```bash
 tar -czvf /backup/data_cache_backup.tar.gz /home/user/.cache
 tar -czvf /backup/data_gfpgan_backup.tar.gz /home/user/app/gfpgan
@@ -41,9 +41,9 @@ tar -czvf /backup/data_gfpgan_backup.tar.gz /home/user/app/gfpgan
 ```bash
 mkdir -p /home/user/.cache
 cd /home/user/.cache
-tar xzvf /backup/data_cache_backup.tar.gz --strip 1
+tar -xzvf /backup/data_cache_backup.tar.gz --strip 1
 
 mkdir -p /home/user/app/gfpgan
 cd /home/user/app/gfpgan
-tar xzvf /backup/data_gfpgan_backup.tar.gz --strip 1
+tar -xzvf /backup/data_gfpgan_backup.tar.gz --strip 1
 ```

prepare_environment.py

@@ -1,11 +0,0 @@
-
-# help me write a python script to download a file from remote server and storage in /tmp
-
-# if the file has already exist in /tmp, just skip download step
-
-# the next step, create the volume, and extract the downloaded data into that volume
-
-# the final step, cleanup /tmp
-
-# http://172.16.15.118:9557/repository/atalink-hf-models/backups/data_cache_backup.tar.gz
-# http://172.16.15.118:9557/repository/atalink-hf-models/backups/data_gfpgan_backup.tar.gz

requirements_f5tts.txt

@@ -1,2 +0,0 @@
-f5_tts
-pypinyin

utils/prepare_environment.py

@@ -0,0 +1,79 @@
+
+# help me write a python script to download a file from remote server and storage in /tmp
+
+# if the file has already exist in /tmp, just skip download step
+
+# the next step, create the volume, and extract the downloaded data into that volume
+
+# the final step, cleanup /tmp
+
+# http://172.16.15.118:9557/repository/atalink-hf-models/backups/data_cache_backup.tar.gz
+# http://172.16.15.118:9557/repository/atalink-hf-models/backups/data_gfpgan_backup.tar.gz
+
+
+#https://drive.usercontent.google.com/download?id=1Kuh-kZApPvI6EIBBy5fK1SRy8fGBHg4z&export=download&authuser=0&confirm=t&uuid=302429b1-d1d4-4aba-8320-ec04d80da509&at=AN8xHoqDn4UCrIap5o_SrGrqaD9F:1758299190305
+import os
+import requests
+import shutil
+import subprocess
+import tarfile
+
+DOWNLOADS = [
+    # "http://172.16.15.118:9557/repository/atalink-hf-models/backups/data_cache_backup.tar.gz",
+    # "http://172.16.15.118:9557/repository/atalink-hf-models/backups/data_gfpgan_backup.tar.gz"
+    "https://huggingface.co/hynt/F5-TTS-Vietnamese-ViVoice/resolve/main/config.json"
+]
+
+
+TMP_DIR = "/tmp"
+VOLUME_PREFIX = "test_data_"
+
+def download_file(url, dest_folder):
+    filename = os.path.basename(url)
+    dest_path = os.path.join(dest_folder, filename)
+    if os.path.exists(dest_path):
+        print(f"{filename} already exists at ```{dest_path}```, skipping download.")
+        return dest_path
+    print(f"Downloading {filename} ...")
+    with requests.get(url, stream=True) as r:
+        r.raise_for_status()
+        with open(dest_path, 'wb') as f:
+            shutil.copyfileobj(r.raw, f)
+    print(f"Downloaded {filename} at {dest_path}")
+    return dest_path
+
+def create_volume_and_extract(tar_path, volume_name):
+    # Create Docker volume
+    subprocess.run(["docker", "volume", "create", volume_name], check=True)
+    # Extract tar.gz into the volume using a temporary container
+    print(f"Extracting {tar_path} into volume {volume_name} ...")
+    subprocess.run([
+        "docker", "run", "--rm",                       # Chạy container tạm thời, tự xóa sau khi xong
+        "-v", f"{volume_name}:/data",                  # Mount Docker volume (volume_name) vào thư mục /data trong container
+        "-v", f"{os.path.dirname(tar_path)}:/tmpdata", # Mount thư mục chứa file tar.gz trên host vào /tmpdata trong container
+        "busybox",                                # Image dùng để chạy container (ở đây là Ubuntu 22.04)
+        "bash", "-c",                                  # Chạy lệnh bash trong container
+        f"tar -xzf /tmpdata/{os.path.basename(tar_path)} -C /data" # Lệnh giải nén file tar.gz từ /tmpdata vào /data
+    ], check=True)
+    print(f"Extracted {tar_path} into volume {volume_name}")
+
+def cleanup_tmp(files):
+    for f in files:
+        try:
+            os.remove(f)
+            print(f"Removed {f}")
+        except Exception as e:
+            print(f"Could not remove {f}: {e}")
+
+def main():
+    downloaded_files = []
+    for url in DOWNLOADS:
+        tar_path = download_file(url, TMP_DIR)
+        downloaded_files.append(tar_path)
+        volume_name = VOLUME_PREFIX + os.path.splitext(os.path.splitext(os.path.basename(tar_path))[0])[0]
+        print("🚀 ~ volume_name:", volume_name)
+        # create_volume_and_extract(tar_path, volume_name)
+    # cleanup_tmp(downloaded_files)
+
+if __name__ == "__main__":
+    main()