Upload reupload_dataset.py

reupload_dataset.py

@@ -0,0 +1,248 @@
+from huggingface_hub import login, hf_hub_download, list_repo_files
+from datasets import Dataset, Audio, DatasetDict
+import json
+import shutil
+import os
+import sys
+import readchar
+import re
+import subprocess
+
+# This script has been designed to run on Windows, but *should* also run on linux too.
+# このスクリプトはWindowsでの実行を想定していますが、Linuxでも動作するはずです。
+
+# Configuration - fill these out
+# 設定 - ここに入力してください
+SOURCE_REPO_ID = "OOPPEENN/56697375616C4E6F76656C5F44617461736574"
+SOURCE_REPO_SUBFOLDER = "GalGame"
+
+HF_READ_TOKEN = ""
+HF_WRITE_TOKEN = ""
+
+TARGET_REPO_ID = "username/dataset-name" # dataset the original data will be uploaded to.
+
+ARCHIVE_PASSWORD = "" # password for 7z archives
+
+STATUS_FILE = "upload_status.json" # filename to track upload progress for the 'continue' option.
+
+
+def display_menu(selected_index, options):
+    os.system('cls' if os.name == 'nt' else 'clear')
+    print("="*50)
+    print("      Dataset reuploader")
+    print("="*50)
+    print("Use up and down arrow keys to navigate. enter to select\n")
+
+    for i, option in enumerate(options):
+        if i == selected_index:
+            print(f"  > {option}")
+        else:
+            print(f"    {option}")
+
+    print("\n" + "="*50)
+
+def download_process_and_upload(archive_filename):
+    # Assumes 7zip is installed and 7z utility is in path
+    # 7zipがインストールされており、7zユーティリティがパスに含まれていることを前提としています
+
+    downloaded_path = None
+    extract_dir = None
+    try:
+        # download:
+        print(f"-> Downloading '{archive_filename}' from {SOURCE_REPO_ID}...")
+        downloaded_path = hf_hub_download(
+            repo_id=SOURCE_REPO_ID, filename=archive_filename, repo_type="dataset",
+            subfolder=SOURCE_REPO_SUBFOLDER, token=HF_READ_TOKEN, local_dir="."
+        )
+        print(f"   Successfully downloaded to: {downloaded_path}")
+
+        # prepar new data
+        archive_name_without_ext = os.path.splitext(os.path.basename(archive_filename))[0]
+        sanitized_config_name = re.sub(r'[\s\W]+', '_', archive_name_without_ext).strip('_')
+        extract_dir = f"./{sanitized_config_name}_extracted"
+        if os.path.exists(extract_dir): shutil.rmtree(extract_dir)
+        os.makedirs(extract_dir)
+
+        # extract archive
+        print(f"\n-> Extracting {archive_filename} using 7zip...")
+
+        command = [
+            '7z', 'x', downloaded_path,
+            f'-o{extract_dir}',
+            f'-p{ARCHIVE_PASSWORD}',
+            '-y'
+        ]
+
+        result = subprocess.run(command, check=True, capture_output=True, text=True)
+        print("   Extraction complete.")
+
+        print("-> Searching for index.json...")
+        index_path, base_dir = None, None
+        for root, _, files in os.walk(extract_dir):
+            if 'index.json' in files:
+                index_path, base_dir = os.path.join(root, 'index.json'), root
+                break
+
+        if not index_path: raise FileNotFoundError("Could not find index.json in extracted files.")
+
+        print("-> Loading index.json...")
+        with open(index_path, 'r', encoding='utf-8') as f:
+            metadata = json.load(f)
+
+        print("-> Creating new dataset")
+        dataset_entries = []
+        for item in metadata:
+            text = item.get("Text")
+            if not text: continue
+            audio_path, voice_id = None, None
+            if "FilePath" in item:
+                relative_path = item["FilePath"].replace('\\', os.path.sep)
+                audio_path = os.path.join(base_dir, relative_path)
+                voice_id = os.path.splitext(os.path.basename(relative_path))[0]
+            else:
+                speaker, temp_voice_id = item.get("Speaker"), item.get("Voice")
+                if speaker and temp_voice_id:
+                    audio_path = os.path.join(base_dir, speaker.strip(), f"{temp_voice_id.strip()}.ogg")
+                    voice_id = temp_voice_id.strip()
+            if audio_path and voice_id and os.path.exists(audio_path):
+                dataset_entries.append({"audio": audio_path, "text": text, "audio_ID": f"{voice_id}__{sanitized_config_name}"})
+
+        if not dataset_entries:
+            print(f"   [WARNING] No valid audio files were found for {archive_filename}. Skipping upload.")
+            return True
+
+        new_dataset = Dataset.from_list(dataset_entries).cast_column("audio", Audio(sampling_rate=48000))
+        print(f"   Processed {len(new_dataset)} valid entries.")
+
+        new_dataset_dict = DatasetDict({"train": new_dataset})
+
+        print(f"\n-> Pushing {len(new_dataset)} new rows to hub as configuration '{sanitized_config_name}'...")
+        new_dataset_dict.push_to_hub(
+            repo_id=TARGET_REPO_ID,
+            config_name=sanitized_config_name,
+            token=HF_WRITE_TOKEN
+        )
+
+        print(f"\n##### Success! '{archive_filename}' uploaded as a new configuration #####")
+        return True
+
+    except subprocess.CalledProcessError as e:
+        print(f"\n[ERROR] 7-Zip failed to extract {archive_filename}.")
+        print(f"   Return Code: {e.returncode}")
+        print(f"   Output: {e.stdout}")
+        print(f"   Error Output: {e.stderr}")
+        return False
+    except Exception as e:
+        print(f"\n[ERROR] An error occurred while processing {archive_filename}: {e}")
+        import traceback
+        traceback.print_exc()
+        return False
+    finally:
+        print("-> Cleaning files...")
+        if downloaded_path and os.path.exists(downloaded_path):
+            os.remove(downloaded_path)
+            containing_folder = os.path.dirname(downloaded_path)
+            if os.path.basename(containing_folder) == SOURCE_REPO_SUBFOLDER and not os.listdir(containing_folder):
+                try:
+                    os.rmdir(containing_folder)
+                    parent_folder = os.path.dirname(containing_folder)
+                    if os.path.basename(parent_folder) == SOURCE_REPO_ID.split('/')[1] and not os.listdir(parent_folder):
+                         os.rmdir(parent_folder)
+                except OSError:
+                    pass
+        if extract_dir and os.path.exists(extract_dir):
+            shutil.rmtree(extract_dir)
+        print("   Cleanup complete.")
+
+
+def reupload_single_archive():
+    os.system('cls' if os.name == 'nt' else 'clear')
+    print("="*50); print("      Reupload Single Archive"); print("="*50)
+
+    filename = input("Enter the filename of .7z archive to process:\n> ")
+    if not filename.endswith('.7z'):
+        print("\nInvalid filename. Please provide the full name ending in .7z")
+        return
+
+    download_process_and_upload(filename)
+
+def continue_uploading_archives():
+    os.system('cls' if os.name == 'nt' else 'clear')
+    print("="*50); print("      Continue Uploading Archives"); print("="*50)
+
+    # Get the full list of archives from hub
+    print("\n-> Fetching file list from source repo...")
+    try:
+        all_repo_files = list_repo_files(SOURCE_REPO_ID, repo_type="dataset", token=HF_READ_TOKEN)
+        archive_filenames = sorted([os.path.basename(f) for f in all_repo_files if f.startswith(f"{SOURCE_REPO_SUBFOLDER}/") and f.endswith('.7z')])
+        if not archive_filenames:
+            print("No .7z files found in the repo. Nothing to do.")
+            return
+        print(f"   Found {len(archive_filenames)} total archives.")
+    except Exception as e:
+        print(f"[ERROR] Could not fetch file list: {e}")
+        return
+
+    # load/create the status file
+    status_data = {}
+    if os.path.exists(STATUS_FILE):
+        print(f"-> Found existing status file '{STATUS_FILE}'. Loading progress")
+        with open(STATUS_FILE, 'r') as f:
+            status_data = json.load(f)
+    else:
+        print(f"-> No status file found. Creating '{STATUS_FILE}'.")
+
+    # create list of files to process
+    files_to_process = [f for f in archive_filenames if not status_data.get(f, False)]
+    total_to_process = len(files_to_process)
+    if not files_to_process:
+        print("\nAll archives have already been processed. Nothing to do!")
+        return
+
+    print(f"   {total_to_process} archives remaining to be processed.")
+
+    # Loop through files and process
+    for i, filename in enumerate(files_to_process):
+        print("\n" + "="*50)
+        print(f"(Processing archive {i+1} / {total_to_process})")
+        print("="*50)
+
+        success = download_process_and_upload(filename)
+
+        if success:
+            status_data[filename] = True
+            with open(STATUS_FILE, 'w') as f:
+                json.dump(status_data, f, indent=4)
+            print(f"   Updated status for '{filename}' to complete.")
+        else:
+            print(f"\n[WARNING!] Processing failed for '{filename}'")
+            return # stop process on failure
+
+    print("\n\n" + "="*50)
+    print("      ALL ARCHIVES PROCESSED SUCCESSFULLY!")
+    print("="*50)
+
+def main():
+    print("Logging into huggingface hub with write access...")
+
+    login(token=HF_WRITE_TOKEN, add_to_git_credential=True)
+
+    selected_index = 0
+    menu_options = ["Reupload single archive", "Continue uploading archives", "Exit"]
+    while True:
+        display_menu(selected_index, menu_options)
+        key = readchar.readkey()
+
+        if key == readchar.key.UP: selected_index = (selected_index - 1) % len(menu_options)
+
+        elif key == readchar.key.DOWN: selected_index = (selected_index + 1) % len(menu_options)
+        elif key == readchar.key.ENTER:
+            if selected_index == 0: reupload_single_archive()
+            elif selected_index == 1: continue_uploading_archives()
+            elif selected_index == 2: print("Exiting."); sys.exit(0)
+            input("\nPress Enter to return to the main menu")
+        elif key in (readchar.key.CTRL_C, readchar.key.CTRL_D):
+            print("Exiting... Goodbye."); sys.exit(0)
+
+if __name__ == "__main__":
+    main()