chore: remove unnecessary files and update build scripts

clean

@@ -4,6 +4,9 @@ rm -rf venv
 rm -rf .venv
 rm -rf .env_*
 rm -rf src/transcribe_anything/venv
+rm -rf src/transcribe_anything/transcribe_anything.egg-info
+rm -rf src/transcribe_anything.egg-info
+rm -rf src/transcribe_anything/__pycache__
 # remove all *.pyc files
 find . -name "*.pyc" -exec rm -rf {} \;
 # remove all *.egg files

lint

@@ -7,25 +7,25 @@ echo activating venv...
 set -e
 
 echo installing pip dependencies
-pip install -r requirements.txt
+
 
 echo running black...
-black src tests --exclude 'src/transcribe_anything/venv'
+uv run black src tests --exclude 'src/transcribe_anything/venv'
 
 echo running isort...
-isort src tests
+uv run isort src tests
 
 echo re-running black now
-black src tests  --exclude 'src/transcribe_anything/venv'
+uv run black src tests  --exclude 'src/transcribe_anything/venv'
 
 # does --clean exist
 if [[ $* == *--no-ruff* ]]; then
     echo skipping ruff...
 else
     echo running ruff...
-    ruff check --fix src tests
+    uv run ruff check --fix src tests
 fi
 
 
 echo running mypy...
-mypy src tests --exclude 'src/transcribe_anything/venv'
+uv run mypy src tests --exclude 'src/transcribe_anything/venv'

pyproject.toml

@@ -4,35 +4,46 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "transcribe-anything"
+version = "2.3.7"  # Update this manually or configure setuptools-scm for automatic versioning
 readme = "README.md"
-maintainers = [
-  {name = "Zachary Vorhies", email = "dont@email.me"}
+description = "Uses Whisper AI to transcribe speech from video and audio files. Also accepts URLs for YouTube, Rumble, BitChute, clear file links, etc."
+requires-python = ">=3.10"
+keywords = ["transcribe", "openai", "whisper"]
+license = { text = "BSD-3-Clause" }
+dependencies = [
+    "static-ffmpeg>=2.7",
+    "yt-dlp>=2025.1.15",
+    "appdirs>=1.4.4",
+    "disklru>=1.0.7",
+    "FileLock",
+    "webvtt-py==0.4.6",
+    "uv-iso-env>=1.0.33",
 ]
 
-description = "Uses whisper AI to transcribe speach from video and audio files. Also accepts urls for youtube, rumble, bitchute, clear file, etc."
-requires-python = ">=3.10"
-keywords = ["transcribe", "open ai", "whisper"]
-license = { text = "BSD 3-Clause License" }
-classifiers = ["Programming Language :: Python :: 3"]
-dynamic = ["dependencies"]
-# Change this with the version number bump.
-version = "2.3.0"
+maintainers = [{ name = "Zachary Vorhies", email = "dont@email.me" }]
+
+[project.urls]
+homepage = "https://github.com/zackees/transcribe-anything"
 
 [tool.setuptools]
 package-dir = {"" = "src"}
 
+[tool.setuptools.packages.find]
+where = ["src"]
+include = ["transcribe_anything*"]
+exclude = ["tests*", "docs*", "examples*"]
+namespaces = false  # Prevent setuptools from looking elsewhere
+
+
+[project.scripts]
+transcribe_anything = "transcribe_anything._cmd:main"
+transcribe-anything = "transcribe_anything._cmd:main"
+
 [tool.ruff]
 line-length = 200
 
 [tool.pylint."MESSAGES CONTROL"]
-good-names = [
-    "c",
-    "i",
-    "ok",
-    "id",
-    "e",
-    "f"
-]
+good-names = ["c", "i", "ok", "id", "e", "f"]
 disable = [
     "missing-function-docstring",
     "missing-module-docstring"
@@ -45,9 +56,6 @@ profile = "black"
 ignore_missing_imports = true
 disable_error_code = ["import-untyped"]
 
-[tool.setuptools.dynamic]
-dependencies = {file = ["requirements.txt"]}
-
-[project.scripts]
-transcribe_anything = "transcribe_anything._cmd:main"
-transcribe-anything = "transcribe_anything._cmd:main"
+[tool.black]
+line-length = 200
+target-version = ['py310']

setup.py

@@ -1,19 +1,6 @@
 
-
-
-import os
-
 from setuptools import setup
 
-URL = "https://github.com/zackees/transcribe-anything"
-KEYWORDS = "isolated environment, uv pip python build tools"
-HERE = os.path.dirname(os.path.abspath(__file__))
-
 if __name__ == "__main__":
-    setup(
-        keywords=KEYWORDS,
-        url=URL,
-        package_data={"": ["assets/example.txt"]},
-        include_package_data=True)
-
+    setup()
 

src/transcribe_anything/_cmd.py

@@ -1,5 +1,5 @@
 """
-    Entry point for running the transcribe-anything prgram.
+Entry point for running the transcribe-anything prgram.
 """
 
 # flake8: noqa E501
@@ -59,11 +59,7 @@ def parse_arguments() -> argparse.Namespace:
     """Parse arguments."""
     whisper_options = get_whisper_options()
     device = get_computing_device()
-    help_str = (
-        f'transcribe_anything is using a "{device}" device.'
-        " Any unrecognized args are assumed to be for whisper"
-        " ai and will be passed as is to whisper ai."
-    )
+    help_str = f'transcribe_anything is using a "{device}" device.' " Any unrecognized args are assumed to be for whisper" " ai and will be passed as is to whisper ai."
     parser = argparse.ArgumentParser(description=help_str)
     parser.add_argument(
         "url_or_file",
@@ -72,10 +68,7 @@ def parse_arguments() -> argparse.Namespace:
     )
     parser.add_argument(
         "--query-gpu-json-path",
-        help=(
-            "Query the GPU and store it in the given path,"
-            " warning takes a long time on first load!"
-        ),
+        help=("Query the GPU and store it in the given path," " warning takes a long time on first load!"),
         type=Path,
     )
     parser.add_argument(
@@ -119,18 +112,12 @@ def parse_arguments() -> argparse.Namespace:
     )
     parser.add_argument(
         "--diarization_model",
-        help=(
-            "Name of the pretrained model/ checkpoint to perform diarization."
-            + " (default: pyannote/speaker-diarization). Only works for --device insane."
-        ),
+        help=("Name of the pretrained model/ checkpoint to perform diarization." + " (default: pyannote/speaker-diarization). Only works for --device insane."),
         default="pyannote/speaker-diarization-3.1",
     )
     parser.add_argument(
         "--timestamp",
-        help=(
-            "Whisper supports both chunked as well as word level timestamps. (default: chunk)."
-            + " Only works for --device insane."
-        ),
+        help=("Whisper supports both chunked as well as word level timestamps. (default: chunk)." + " Only works for --device insane."),
         choices=["chunk", "word"],
         default=None,
     )
@@ -163,10 +150,7 @@ def main() -> int:
     if args.model == "large-legacy":
         args.model = "large"
     elif args.model == "large":
-        print(
-            "Defaulting to large-v3 model for --model large,"
-            + " use --model large-legacy for the old model"
-        )
+        print("Defaulting to large-v3 model for --model large," + " use --model large-legacy for the old model")
         args.model = "large-v3"
     elif args.model is None and args.device == "insane":
         print("Defaulting to large-v3 model for --device insane")
@@ -187,9 +171,7 @@ def main() -> int:
     # For now, just stuff --diarization_model and --timestamp into unknown
     if args.diarization_model:
         if args.device != "insane":
-            print(
-                "--diarization_model only works with --device insane. Ignoring --diarization_model"
-            )
+            print("--diarization_model only works with --device insane. Ignoring --diarization_model")
         else:
             unknown.append(f"--diarization_model {args.diarization_model}")
 

src/transcribe_anything/api.py

@@ -1,5 +1,5 @@
 """
-    Api for using transcribe_anything from python. Allows bulk processing.
+Api for using transcribe_anything from python. Allows bulk processing.
 """
 
 # pylint: disable=too-many-arguments,broad-except,too-many-locals,unsupported-binary-operation,too-many-branches,too-many-statements,disable=notimplemented-raised,unused-variable,line-too-long
@@ -42,14 +42,7 @@
 
 CACHE_FILE = os.path.join(user_config_dir("transcript-anything", "cache", roaming=True))
 
-PERMS = (
-    stat.S_IRUSR
-    | stat.S_IRGRP
-    | stat.S_IROTH
-    | stat.S_IWOTH
-    | stat.S_IWUSR
-    | stat.S_IWGRP
-)
+PERMS = stat.S_IRUSR | stat.S_IRGRP | stat.S_IROTH | stat.S_IWOTH | stat.S_IWUSR | stat.S_IWGRP
 
 
 class Device(Enum):
@@ -81,9 +74,7 @@ def make_temp_wav() -> str:
     """
     Makes a temporary mp3 file and returns the path to it.
     """
-    tmp = tempfile.NamedTemporaryFile(  # pylint: disable=consider-using-with
-        suffix=".wav", delete=False
-    )
+    tmp = tempfile.NamedTemporaryFile(suffix=".wav", delete=False)  # pylint: disable=consider-using-with
 
     tmp.close()
 
@@ -151,9 +142,7 @@ def get_video_name_from_url(url: str) -> str:
             log_error("yt-dlp failed to get title, using basename instead.")
             return os.path.basename(url)
     except subprocess.CalledProcessError as exc:
-        log_error(
-            f"yt-dlp failed with {exc}, using basename instead\n{exc.stdout}\n{exc.stderr}"
-        )
+        log_error(f"yt-dlp failed with {exc}, using basename instead\n{exc.stdout}\n{exc.stderr}")
         return os.path.basename(url)
     except Exception as exc:
         log_error(f"yt-dlp failed with {exc}, using basename instead.")
@@ -192,10 +181,7 @@ def transcribe(
     static_ffmpeg.add_paths()
     check_python_in_range()
     if not os.path.isfile(url_or_file) and embed:
-        raise NotImplementedError(
-            "Embedding is only supported for local files. "
-            + "Please download the file first."
-        )
+        raise NotImplementedError("Embedding is only supported for local files. " + "Please download the file first.")
     # cache = DiskLRUCache(CACHE_FILE, 16)
     basename = os.path.basename(url_or_file)
     if not basename or basename == ".":  # if url_or_file is a directory
@@ -216,9 +202,7 @@ def transcribe(
     print(f"making dir {output_dir}")
     os.makedirs(output_dir, exist_ok=True)
     tmp_wav = make_temp_wav()
-    assert os.path.isdir(
-        output_dir
-    ), f"Path {output_dir} is not found or not a directory."
+    assert os.path.isdir(output_dir), f"Path {output_dir} is not found or not a directory."
     # tmp_mp3 = os.path.join(output_dir, "out.mp3")
     fetch_audio(url_or_file, tmp_wav)
     assert os.path.exists(tmp_wav), f"Path {tmp_wav} doesn't exist."
@@ -319,9 +303,7 @@ def transcribe(
             except subprocess.CalledProcessError as exc:
                 stdout = exc.stdout
                 stderr = exc.stderr
-                warnings.warn(
-                    f"ffmpeg failed with return code {exc.returncode}\n{stdout}\n{stderr}"
-                )
+                warnings.warn(f"ffmpeg failed with return code {exc.returncode}\n{stdout}\n{stderr}")
                 raise
     print(f"Done! Files were saved to {output_dir}")
     return output_dir

src/transcribe_anything/audio.py

@@ -1,5 +1,5 @@
 """
-    Fetches audio and handles transcoding it for it.
+Fetches audio and handles transcoding it for it.
 """
 
 import os
@@ -12,15 +12,11 @@
 from transcribe_anything.ytldp_download import ytdlp_download
 
 
-def _convert_to_wav(
-    inpath: str, outpath: str, speech_normalization: bool = False
-) -> None:
+def _convert_to_wav(inpath: str, outpath: str, speech_normalization: bool = False) -> None:
     """Converts a file to wav."""
     # static_ffmpeg -y -i C:\Users\niteris\AppData\Local\Temp\tmp3xhzm1sn\out.webm -filter:a "speechnorm=e=12.5:r=0.00001:l=1" -acodec pcm_s16le -ar 44100 -ac 1 C:\Users\niteris\AppData\Local\Temp\tmpu32zsjov.wav
 
-    tmpwav = tempfile.NamedTemporaryFile(  # pylint: disable=consider-using-with
-        suffix=".wav", delete=False
-    )
+    tmpwav = tempfile.NamedTemporaryFile(suffix=".wav", delete=False)  # pylint: disable=consider-using-with
     tmpwav.close()
     tmpwavepath = tmpwav.name
 

src/transcribe_anything/cuda_available.py

@@ -23,10 +23,7 @@ class CudaDevice:
     device_id: int
 
     def __str__(self):
-        return (
-            f"{self.name} - VRAM: {self.vram / (1024 ** 3):.2f} GB, "
-            f"Multiprocessors: {self.multiprocessors}"
-        )
+        return f"{self.name} - VRAM: {self.vram / (1024 ** 3):.2f} GB, " f"Multiprocessors: {self.multiprocessors}"
 
     def to_json(self) -> dict[str, str | int]:
         """Returns a dictionary representation of the object."""

src/transcribe_anything/generate_speaker_json.py

@@ -30,10 +30,7 @@ def to_json(self) -> dict:
 
 def can_combine(chunk1: Chunk, chunk2: Chunk) -> bool:
     """Check if two chunks can be combined."""
-    return (
-        chunk1.speaker == chunk2.speaker
-        and abs(chunk1.timestamp_end - chunk2.timestamp_start) <= 0.1
-    )
+    return chunk1.speaker == chunk2.speaker and abs(chunk1.timestamp_end - chunk2.timestamp_start) <= 0.1
 
 
 def reduce(dat: list[Chunk]) -> list[Chunk]:
@@ -46,9 +43,7 @@ def reduce(dat: list[Chunk]) -> list[Chunk]:
             continue
         last_chunk = out[-1]
         if not can_combine(last_chunk, chunk):
-            chunk.reason = (
-                "speaker-switch" if last_chunk.speaker != chunk.speaker else "timeout"
-            )
+            chunk.reason = "speaker-switch" if last_chunk.speaker != chunk.speaker else "timeout"
             out.append(chunk)
             continue
         # combine