videopython 0.2.0__tar.gz → 0.2.1__tar.gz

videopython-0.2.1/.gitignore

@@ -0,0 +1,140 @@
+# Byte-compiled / optimized / DLL files
+__pycache__/
+*.py[cod]
+*$py.class
+
+# C extensions
+*.so
+
+# Distribution / packaging
+.Python
+build/
+develop-eggs/
+dist/
+downloads/
+eggs/
+.eggs/
+lib/
+lib64/
+parts/
+sdist/
+var/
+wheels/
+pip-wheel-metadata/
+share/python-wheels/
+*.egg-info/
+.installed.cfg
+*.egg
+MANIFEST
+
+# PyInstaller
+#  Usually these files are written by a python script from a template
+#  before PyInstaller builds the exe, so as to inject date/other infos into it.
+*.manifest
+*.spec
+
+# Installer logs
+pip-log.txt
+pip-delete-this-directory.txt
+
+# Unit test / coverage reports
+htmlcov/
+.tox/
+.nox/
+.coverage
+.coverage.*
+.cache
+nosetests.xml
+coverage.xml
+*.cover
+*.py,cover
+.hypothesis/
+.pytest_cache/
+
+# Translations
+*.mo
+*.pot
+
+# Django stuff:
+*.log
+local_settings.py
+db.sqlite3
+db.sqlite3-journal
+
+# Flask stuff:
+instance/
+.webassets-cache
+
+# Scrapy stuff:
+.scrapy
+
+# Sphinx documentation
+docs/_build/
+
+# PyBuilder
+target/
+
+# Jupyter Notebook
+.ipynb_checkpoints
+
+# IPython
+profile_default/
+ipython_config.py
+
+# pyenv
+.python-version
+
+# pipenv
+#   According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control.
+#   However, in case of collaboration, if having platform-specific dependencies or dependencies
+#   having no cross-platform support, pipenv may install dependencies that don't work, or not
+#   install all needed dependencies.
+#Pipfile.lock
+
+# PEP 582; used by e.g. github.com/David-OConnor/pyflow
+__pypackages__/
+
+# Celery stuff
+celerybeat-schedule
+celerybeat.pid
+
+# SageMath parsed files
+*.sage.py
+
+# Environments
+.env
+.venv
+env/
+venv/
+ENV/
+env.bak/
+venv.bak/
+
+# Spyder project settings
+.spyderproject
+.spyproject
+
+# Rope project settings
+.ropeproject
+
+# mkdocs documentation
+/site
+
+# mypy
+.mypy_cache/
+.dmypy.json
+dmypy.json
+
+# type checker
+.pyre/
+.mypy_cache/
+
+# Random shit
+*.ipynb
+.vscode
+*.csv
+
+# Data directories
+data/downloaded/*.mp4
+data/exported/*.mp4
+!data/exported/example.mp4

videopython-0.2.1/PKG-INFO

@@ -0,0 +1,130 @@
+Metadata-Version: 2.3
+Name: videopython
+Version: 0.2.1
+Summary: Minimal video generation and processing library.
+Project-URL: Homepage, https://github.com/bartwojtowicz/videopython/
+Project-URL: Repository, https://github.com/bartwojtowicz/videopython/
+Project-URL: Documentation, https://github.com/bartwojtowicz/videopython/
+Author-email: Bartosz Wójtowicz <bartoszwojtowicz@outlook.com>, Bartosz Rudnikowicz <bartoszrudnikowicz840@gmail.com>, Piotr Pukisz <piotr.pukisz@gmail.com>
+License: Apache-2.0
+License-File: LICENSE
+Keywords: editing,generation,movie,opencv,python,video,videopython
+Classifier: License :: OSI Approved :: Apache Software License
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Requires-Python: <3.13,>=3.10
+Requires-Dist: numpy>=1.25.2
+Requires-Dist: opencv-python>=4.9.0.80
+Requires-Dist: pillow>=10.3.0
+Requires-Dist: pydub>=0.25.1
+Requires-Dist: tqdm>=4.66.3
+Provides-Extra: dev
+Requires-Dist: black==24.3.0; extra == 'dev'
+Requires-Dist: isort==5.12.0; extra == 'dev'
+Requires-Dist: mypy==1.8.0; extra == 'dev'
+Requires-Dist: pydub-stubs==0.25.1.1; extra == 'dev'
+Requires-Dist: pytest==7.4.0; extra == 'dev'
+Requires-Dist: types-pillow==10.2.0.20240213; extra == 'dev'
+Requires-Dist: types-tqdm==4.66.0.20240106; extra == 'dev'
+Provides-Extra: generation
+Requires-Dist: accelerate>=0.29.2; extra == 'generation'
+Requires-Dist: diffusers>=0.26.3; extra == 'generation'
+Requires-Dist: torch>=2.1.0; extra == 'generation'
+Requires-Dist: transformers>=4.38.1; extra == 'generation'
+Description-Content-Type: text/markdown
+
+# About
+
+Minimal video generation and processing library.
+
+## Setup 
+
+### Install ffmpeg
+```bash
+# Install with brew for MacOS:
+brew install ffmpeg
+# Install with apt-get for Ubuntu:
+sudo apt-get install ffmpeg
+```
+
+### Install with pip
+```bash
+pip install videopython[generation]
+```
+> You can install without `[generation]` dependencies for basic video handling and processing. 
+> The funcionalities found in `videopython.generation` won't work.
+
+## Basic Usage
+
+### Video handling
+
+```python
+from videopython.base.video import Video
+
+# Load videos and print metadata
+video1 = Video.from_path("tests/test_data/fast_benchmark.mp4")
+print(video1)
+
+video2 = Video.from_path("tests/test_data/slow_benchmark.mp4")
+print(video2)
+
+# Define the transformations
+from videopython.base.transforms import CutSeconds, ResampleFPS, Resize, TransformationPipeline
+
+pipeline = TransformationPipeline(
+    [CutSeconds(start=1.5, end=6.5), ResampleFPS(fps=30), Resize(width=1000, height=1000)]
+)
+video1 = pipeline.run(video1)
+video2 = pipeline.run(video2)
+
+# Combine videos, add audio and save
+from videopython.base.transitions import FadeTransition
+
+fade = FadeTransition(effect_time_seconds=3.0)
+video = fade.apply(videos=(video1, video2))
+video.add_audio_from_file("tests/test_data/test_audio.mp3")
+
+savepath = video.save()
+```
+
+### Video Generation
+
+> Using Nvidia A40 or better is recommended for the `videopython.generation` module.
+```python
+# Generate image and animate it
+from videopython.generation import ImageToVideo
+from videopython.generation import TextToImage
+from videopython.generation import TextToMusic
+
+image = TextToImage().generate_image(prompt="Golden Retriever playing in the park")
+video = ImageToVideo().generate_video(image=image, fps=24)
+
+# Video generation directly from prompt
+from videopython.generation import TextToVideo
+video_gen = TextToVideo()
+video = video_gen.generate_video("Dogs playing in the snow")
+for _ in range(10):
+    video += video_gen.generate_video("Dogs playing in the snow")
+
+# Cut the first 2 seconds
+from videopython.base.transforms import CutSeconds
+transformed_video = CutSeconds(start_second=0, end_second=2).apply(video.copy())
+
+# Upsample to 30 FPS
+from videopython.base.transforms import ResampleFPS
+transformed_video = ResampleFPS(new_fps=30).apply(transformed_video)
+
+# Resize to 1000x1000
+from videopython.base.transforms import Resize
+transformed_video = Resize(width=1000, height=1000).apply(transformed_video)
+
+# Add generated music
+# MusicGen cannot generate more than 1503 tokens (~30seconds of audio)
+text_to_music = TextToMusic()
+audio = text_to_music.generate_audio("Happy dogs playing together in a park", max_new_tokens=256)
+transformed_video.add_audio(audio=audio)
+
+filepath = transformed_video.save()
+```

{videopython-0.2.0 → videopython-0.2.1}/README.md

@@ -59,6 +59,7 @@ savepath = video.save()
 # Generate image and animate it
 from videopython.generation import ImageToVideo
 from videopython.generation import TextToImage
+from videopython.generation import TextToMusic
 
 image = TextToImage().generate_image(prompt="Golden Retriever playing in the park")
 video = ImageToVideo().generate_video(image=image, fps=24)
@@ -82,5 +83,11 @@ transformed_video = ResampleFPS(new_fps=30).apply(transformed_video)
 from videopython.base.transforms import Resize
 transformed_video = Resize(width=1000, height=1000).apply(transformed_video)
 
+# Add generated music
+# MusicGen cannot generate more than 1503 tokens (~30seconds of audio)
+text_to_music = TextToMusic()
+audio = text_to_music.generate_audio("Happy dogs playing together in a park", max_new_tokens=256)
+transformed_video.add_audio(audio=audio)
+
 filepath = transformed_video.save()
 ```

videopython-0.2.1/pyproject.toml

@@ -0,0 +1,88 @@
+[project]
+name = "videopython"
+version = "0.2.1"
+description = "Minimal video generation and processing library."
+authors = [
+    { name = "Bartosz Wójtowicz", email = "bartoszwojtowicz@outlook.com" },
+    { name = "Bartosz Rudnikowicz", email = "bartoszrudnikowicz840@gmail.com" },
+    { name = "Piotr Pukisz", email = "piotr.pukisz@gmail.com" }
+]
+license = { text = "Apache-2.0" }
+readme = "README.md"
+requires-python = ">=3.10, <3.13"
+keywords = ["python", "videopython", "video", "movie", "opencv", "generation", "editing"]
+classifiers = [
+    "License :: OSI Approved :: Apache Software License",
+    "Programming Language :: Python :: 3",
+    "Programming Language :: Python :: 3.10",
+    "Programming Language :: Python :: 3.11",
+    "Operating System :: OS Independent",
+]
+
+dependencies = [
+    "numpy>=1.25.2",
+    "opencv-python>=4.9.0.80",
+    "pillow>=10.3.0",
+    "pydub>=0.25.1",
+    "tqdm>=4.66.3",
+]
+
+[project.optional-dependencies]
+dev = [
+    "black==24.3.0",
+    "isort==5.12.0",
+    "mypy==1.8.0",
+    "pytest==7.4.0",
+    "types-Pillow==10.2.0.20240213",
+    "types-tqdm==4.66.0.20240106",
+    "pydub-stubs==0.25.1.1",
+]
+generation = [
+    "accelerate>=0.29.2",
+    "diffusers>=0.26.3",
+    "torch>=2.1.0",
+    "transformers>=4.38.1",
+]
+
+[project.urls]
+Homepage = "https://github.com/bartwojtowicz/videopython/"
+Repository = "https://github.com/bartwojtowicz/videopython/"
+Documentation = "https://github.com/bartwojtowicz/videopython/"
+
+[tool.rye]
+managed = true
+dev-dependencies = [
+    "black==24.3.0",
+    "isort==5.12.0",
+    "mypy==1.8.0",
+    "pytest==7.4.0",
+    "types-Pillow==10.2.0.20240213",
+    "types-tqdm==4.66.0.20240106",
+    "pydub-stubs==0.25.1.1",
+]
+
+[tool.rye.scripts]
+test-unit = "pytest"
+test-type = "mypy src"
+test-static = { chain = [
+    "black src -l 120 --check",
+    "isort src --profile black --check"
+]}
+
+[build-system]
+requires = ["hatchling"]
+build-backend = "hatchling.build"
+
+[tool.hatch.build.targets.wheel]
+packages = ["src/videopython"]
+
+[tool.hatch.build.targets.sdist]
+include = ["src/videopython", "src/videopython/py.typed"]
+
+[tool.mypy]
+mypy_path = "stubs"
+
+[tool.pytest]
+testpaths = ["src/tests"]
+python_files = ["test_*.py"]
+addopts = "-v --tb=short"

{videopython-0.2.0 → videopython-0.2.1}/src/videopython/base/video.py

@@ -2,14 +2,18 @@ from __future__ import annotations
 
 import shlex
 import subprocess
+import tempfile
 from dataclasses import dataclass
 from pathlib import Path
+from typing import Literal, get_args
 
 import cv2
 import numpy as np
 from pydub import AudioSegment
 
-from videopython.utils.common import check_path, generate_random_name
+from videopython.utils.common import generate_random_name
+
+ALLOWED_VIDEO_FORMATS = Literal["mp4", "avi", "mov", "mkv", "webm"]
 
 
 @dataclass
@@ -166,54 +170,80 @@ class Video:
         split_videos[1].audio = self.audio[audio_midpoint:]
         return split_videos
 
-    def save(self, filename: str | None = None) -> str:
-        """Saves the video.
+    def save(self, filename: str | Path | None = None, format: ALLOWED_VIDEO_FORMATS = "mp4") -> Path:
+        """Saves the video with audio.
 
         Args:
-            filename: Name of the output video file. Generates random UUID name if not provided.
+            filename: Name of the output video file. Generates random name if not provided.
+            format: Output format (default is 'mp4').
+
+        Returns:
+            Path to the saved video file.
         """
         if not self.is_loaded():
-            raise RuntimeError(f"Video is not loaded, cannot save!")
-
-        if filename is None:
-            filename = generate_random_name(suffix=".mp4")
-        filename = check_path(filename, dir_exists=True, suffix=".mp4")
+            raise RuntimeError("Video is not loaded, cannot save!")
 
-        ffmpeg_video_command = (
-            f"ffmpeg -loglevel error -y -framerate {self.fps} -f rawvideo -pix_fmt rgb24"
-            f" -s {self.metadata.width}x{self.metadata.height} "
-            f"-i pipe:0 -c:v libx264 -pix_fmt yuv420p {filename}"
-        )
-
-        ffmpeg_audio_command = (
-            f"ffmpeg -loglevel error -y -i {filename} -f s16le -acodec pcm_s16le "
-            f"-ar {self.audio.frame_rate} -ac {self.audio.channels} -i pipe:0 "
-            f"-c:v copy -c:a aac -strict experimental {filename}_temp.mp4"
-        )
-
-        try:
-            print("Saving frames to video...")
-            subprocess.run(
-                ffmpeg_video_command,
-                input=self.frames.tobytes(),
-                check=True,
-                shell=True,
+        # Check if the format is allowed
+        if format.lower() not in get_args(ALLOWED_VIDEO_FORMATS):
+            raise ValueError(
+                f"Unsupported format: {format}. Allowed formats are: {', '.join(get_args(ALLOWED_VIDEO_FORMATS))}"
             )
-        except subprocess.CalledProcessError as e:
-            print("Error saving frames to video!")
-            raise e
-
-        try:
-            print("Adding audio track...")
-            subprocess.run(ffmpeg_audio_command, input=self.audio.raw_data, check=True, shell=True)
-            Path(filename).unlink()
-            Path(filename + "_temp.mp4").rename(filename)
-        except subprocess.CalledProcessError as e:
-            print(f"Error adding audio track!")
-            raise e
 
-        print(f"Video saved into `{filename}`!")
-        return filename
+        if filename is None:
+            filename = Path(generate_random_name(suffix=f".{format}"))
+        else:
+            filename = Path(filename).with_suffix(f".{format}")
+            filename.parent.mkdir(parents=True, exist_ok=True)
+
+        with tempfile.TemporaryDirectory() as temp_dir:
+            temp_dir_path = Path(temp_dir)
+
+            # Save frames as images
+            for i, frame in enumerate(self.frames):
+                frame_path = temp_dir_path / f"frame_{i:04d}.png"
+                cv2.imwrite(str(frame_path), cv2.cvtColor(frame, cv2.COLOR_RGB2BGR))
+
+            # Save audio to a temporary file
+            temp_audio = temp_dir_path / "temp_audio.wav"
+            self.audio.export(str(temp_audio), format="adts", bitrate="192k")
+
+            # Construct FFmpeg command
+            ffmpeg_command = [
+                "ffmpeg",
+                "-y",  # Overwrite output file if it exists
+                "-r",
+                str(self.fps),  # Set the frame rate
+                "-i",
+                str(temp_dir_path / "frame_%04d.png"),  # Input image sequence
+                "-i",
+                str(temp_audio),  # Input audio file
+                "-c:v",
+                "libx264",  # Video codec
+                "-preset",
+                "medium",  # Encoding preset (tradeoff between encoding speed and compression)
+                "-crf",
+                "23",  # Constant Rate Factor (lower means better quality, 23 is default)
+                "-c:a",
+                "copy",  # Audio codec
+                "-b:a",
+                "192k",  # Audio bitrate
+                "-pix_fmt",
+                "yuv420p",  # Pixel format
+                "-shortest",  # Finish encoding when the shortest input stream ends
+                str(filename),
+            ]
+
+            try:
+                subprocess.run(ffmpeg_command, check=True, capture_output=True, text=True)
+                print(f"Video saved successfully to: {filename}")
+                return filename
+            except subprocess.CalledProcessError as e:
+                print(f"Error saving video: {e}")
+                print(f"FFmpeg stderr: {e.stderr}")
+                raise
+
+    def add_audio(self, audio: AudioSegment, overlay: bool = True, overlay_gain: int = 0, loop: bool = False) -> None:
+        self.audio = self._process_audio(audio=audio, overlay=overlay, overlay_gain=overlay_gain, loop=loop)
 
     def add_audio_from_file(self, path: str, overlay: bool = True, overlay_gain: int = 0, loop: bool = False) -> None:
         new_audio = self._load_audio_from_path(path)
@@ -221,15 +251,19 @@ class Video:
             print(f"Audio file `{path}` not found, skipping!")
             return
 
-        if (duration_diff := round(self.total_seconds - new_audio.duration_seconds)) > 0 and not loop:
-            new_audio = new_audio + AudioSegment.silent(duration_diff * 1000)
-        elif new_audio.duration_seconds > self.total_seconds:
-            new_audio = new_audio[: round(self.total_seconds * 1000)]
+        self.audio = self._process_audio(audio=new_audio, overlay=overlay, overlay_gain=overlay_gain, loop=loop)
+
+    def _process_audio(
+        self, audio: AudioSegment, overlay: bool = True, overlay_gain: int = 0, loop: bool = False
+    ) -> AudioSegment:
+        if (duration_diff := round(self.total_seconds - audio.duration_seconds)) > 0 and not loop:
+            audio = audio + AudioSegment.silent(duration_diff * 1000)
+        elif audio.duration_seconds > self.total_seconds:
+            audio = audio[: round(self.total_seconds * 1000)]
 
         if overlay:
-            self.audio = self.audio.overlay(new_audio, loop=loop, gain_during_overlay=overlay_gain)
-        else:
-            self.audio = new_audio
+            return self.audio.overlay(audio, loop=loop, gain_during_overlay=overlay_gain)
+        return audio
 
     def __add__(self, other: Video) -> Video:
         # TODO: Should it be class method? How to make it work with sum()?
@@ -282,17 +316,26 @@ class Video:
         Args:
             path: Path to video file.
         """
-        metadata = VideoMetadata.from_path(path)
-        ffmpeg_command = f"ffmpeg -i {path} -f rawvideo -pix_fmt rgb24 -loglevel quiet pipe:1"
+        cap = cv2.VideoCapture(path)
+        if not cap.isOpened():
+            raise ValueError(f"Unable to open video file: {path}")
+
+        fps = cap.get(cv2.CAP_PROP_FPS)
+        frames = []
+
+        while True:
+            ret, frame = cap.read()
+            if not ret:
+                break
+            frame = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)
+            frames.append(frame)
+
+        cap.release()
 
-        # Run the ffmpeg command and capture the stdout
-        ffmpeg_process = subprocess.Popen(shlex.split(ffmpeg_command), stdout=subprocess.PIPE)
-        ffmpeg_out, _ = ffmpeg_process.communicate()
+        if not frames:
+            raise ValueError(f"No frames could be read from the video file: {path}")
 
-        # Convert the raw video data to a NumPy array
-        frames = np.frombuffer(ffmpeg_out, dtype=np.uint8).reshape([-1, metadata.height, metadata.width, 3])
-        fps = metadata.fps
-        return frames, fps
+        return np.array(frames), fps
 
     @property
     def video_shape(self) -> tuple[int, int, int, int]:

{videopython-0.2.0 → videopython-0.2.1}/src/videopython/generation/__init__.py

@@ -1,4 +1,4 @@
-from .audio import TextToSpeech
+from .audio import TextToMusic, TextToSpeech
 from .image import TextToImage
 from .video import ImageToVideo, TextToVideo
 
@@ -7,4 +7,5 @@ __all__ = [
     "TextToSpeech",
     "TextToImage",
     "TextToVideo",
+    "TextToMusic",
 ]

videopython-0.2.1/src/videopython/generation/audio.py

@@ -0,0 +1,56 @@
+import numpy as np
+import torch
+from pydub import AudioSegment
+from transformers import (
+    AutoProcessor,
+    AutoTokenizer,
+    MusicgenForConditionalGeneration,
+    VitsModel,
+)
+
+TEXT_TO_SPEECH_MODEL = "facebook/mms-tts-eng"
+MUSIC_GENERATION_MODEL_SMALL = "facebook/musicgen-small"
+
+
+class TextToSpeech:
+    def __init__(self):
+        self.pipeline = VitsModel.from_pretrained(TEXT_TO_SPEECH_MODEL)
+        self.tokenizer = AutoTokenizer.from_pretrained(TEXT_TO_SPEECH_MODEL)
+
+    def generate_audio(self, text: str) -> AudioSegment:
+        tokenized = self.tokenizer(text, return_tensors="pt")
+
+        with torch.no_grad():
+            output = self.pipeline(**tokenized).waveform
+
+        output = (output.T.float().numpy() * (2**31 - 1)).astype(np.int32)
+        audio = AudioSegment(data=output, frame_rate=self.pipeline.config.sampling_rate, sample_width=4, channels=1)
+        return audio
+
+
+class TextToMusic:
+    def __init__(self) -> None:
+        """
+        Generates music from text using the Musicgen model.
+        Check the license for the model before using it.
+        """
+        self.processor = AutoProcessor.from_pretrained(MUSIC_GENERATION_MODEL_SMALL)
+        self.model = MusicgenForConditionalGeneration.from_pretrained(MUSIC_GENERATION_MODEL_SMALL)
+
+    def generate_audio(self, text: str, max_new_tokens: int) -> AudioSegment:
+        inputs = self.processor(
+            text=[text],
+            padding=True,
+            return_tensors="pt",
+        )
+        audio_values = self.model.generate(**inputs, max_new_tokens=max_new_tokens)
+        sampling_rate = self.model.config.audio_encoder.sampling_rate
+        output = (audio_values[0, 0].float().numpy() * (2**31 - 1)).astype(np.int32)
+
+        audio = AudioSegment(
+            data=output.tobytes(),
+            frame_rate=sampling_rate,
+            sample_width=4,
+            channels=1,
+        )
+        return audio