commit 4b081c3aad37fcd6a19f3f5858a38e83eddf39bd
parent a124b6dae7aad26959b241609ce8a9f16656cb49
Author: MTRNord <mtrnord1@gmail.com>
Date: Fri, 23 Feb 2024 11:23:09 +0100
More fixes
Diffstat:
4 files changed, 24 insertions(+), 32 deletions(-)
diff --git a/README.md b/README.md
@@ -1,5 +1,6 @@
# transcribe
-A fork for matrix-live transcription@home by fyyd.de
+
+A fork for matrix-live transcription@home by fyyd.de
This is not fully tested. You need a decent CPU. You also need these deps:
@@ -7,15 +8,15 @@ This is not fully tested. You need a decent CPU. You also need these deps:
- git
- yt-dlp
-Next you need to run `./setup.sh` this will ask how many cores you want to use and it will test the setup.
+Next you need to run `./setup.py` this will ask how many cores you want to use and it will test the setup.
It also downloads and compiles the core of this: [whisper.cpp](https://github.com/ggerganov/whisper.cpp)
-To start transcribing you can run `./transcribe.sh`.
+To start transcribing you can run `./transcribe.py`.
You can cancel it with ctrl+c at any time. or create a `.trancribe-stop` file in the parent folder to stop after the current transcription.
You find the vtt outputs in the `./whisper.cpp/output/` folder.
The name contains the video id.
Youtube allows you to upload these:
-- https://support.google.com/youtube/answer/2734796?hl=en#zippy=%2Cupload-a-file
-- https://support.google.com/youtube/answer/2734698#zippy=%2Cadvanced-file-formats
+- <https://support.google.com/youtube/answer/2734796?hl=en#zippy=%2Cupload-a-file>
+- <https://support.google.com/youtube/answer/2734698#zippy=%2Cadvanced-file-formats>
diff --git a/setup.sh b/legacy/setup.sh
diff --git a/transcribe.sh b/legacy/transcribe.sh
diff --git a/transcribe.py b/transcribe.py
@@ -15,7 +15,6 @@ import matplotlib.pyplot as plt
import tqdm
import yt_dlp
from ffmpeg_normalize import FFmpegNormalize
-from matplotlib import font_manager
# Configure logging
logging.basicConfig(
@@ -86,7 +85,7 @@ def setup_whisper() -> None:
check=True,
)
logging.info("Compiling whisper...")
- subprocess.run(["make", "-j", "WHISPER_CLBLAST=1"], cwd=whisper_dir, check=True)
+ subprocess.run(["WHISPER_CLBLAST=1", "make", "-j"], cwd=whisper_dir, check=True)
logging.info("Downloading the model")
subprocess.run(
["./models/download-ggml-model.sh", "medium"], cwd=whisper_dir, check=True
@@ -271,6 +270,7 @@ def transcribe_audio_files(files_directory: str, model: str, threads: int) -> No
srt_exists = os.path.exists(f"output/{base_filename}.srt")
vtt_exists = os.path.exists(f"output/{base_filename}.vtt")
if not (txt_exists and srt_exists and vtt_exists):
+ logging.info(f"Transcribing {file}")
whisper_cmd: List[str] = [
"./main",
"-m",
@@ -291,35 +291,26 @@ def transcribe_audio_files(files_directory: str, model: str, threads: int) -> No
"3.0",
]
- whisper_process: subprocess.CompletedProcess = subprocess.run(
- whisper_cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE
- )
+ whisper_process: subprocess.CompletedProcess = subprocess.run(whisper_cmd)
if whisper_process.returncode != 0:
logging.error("Error transcribing")
else:
logging.info("Transcription successful")
+ backup_file(
+ Path(input_file),
+ Path(input_file.replace("playlist_normalized", "playlist")),
+ Path("./backup"),
+ )
-def backup_files(
- input_directory: str, normalized_directory: str, backup_directory: str
+def backup_file(
+ input_file: Path, normalized_file: Path, backup_directory: Path
) -> None:
- # Backup input files
- input_files: List[str] = [
- file for file in os.listdir(input_directory) if file.endswith(".wav")
- ]
- for file in input_files:
- input_file: str = f"{input_directory}/{file}"
- shutil.move(input_file, f"{backup_directory}/input/{file}")
- logging.info(f"Moved input file {file} to backup directory")
-
- # Backup normalized input files
- normalized_files: List[str] = [
- file for file in os.listdir(normalized_directory) if file.endswith(".wav")
- ]
- for file in normalized_files:
- normalized_file: str = f"{normalized_directory}/{file}"
- shutil.move(normalized_file, f"{backup_directory}/normalized/{file}")
- logging.info(f"Moved normalized file {file} to backup directory")
+ shutil.move(input_file, f"{backup_directory}/input/{input_file}")
+ logging.info(f"Moved input file {input_file} to backup directory")
+
+ shutil.move(normalized_file, f"{backup_directory}/normalized/{normalized_file}")
+ logging.info(f"Moved normalized file {normalized_file} to backup directory")
def run() -> None:
@@ -341,6 +332,9 @@ def run() -> None:
os.makedirs("./playlist", exist_ok=True)
open("./playlist/downloaded.txt", "a").close()
os.makedirs("./playlist_normalized", exist_ok=True)
+ os.makedirs("./backup", exist_ok=True)
+ os.makedirs("./backup/input", exist_ok=True)
+ os.makedirs("./backup/normalized", exist_ok=True)
download_audio_files(playlist_url, "./playlist")
@@ -352,9 +346,6 @@ def run() -> None:
# Transcribe audio files
transcribe_audio_files("./playlist_normalized", model, threads)
- # Cleanup
- backup_files("./playlist", "./output", "./backup")
-
logging.info("Transcription completed!")
# Change back to the original directory
