[DAR-5390][External] Add 'extract video-artifacts' command

.github/workflows/JOB_tests.yml

@@ -48,6 +48,23 @@ jobs:
                      pip install wheel && \
                      pip install --upgrade setuptools && \
                      pip install --editable '.[test,ml,medical,dev, ocv]'"
+
+      - name: Install ffmpeg (Ubuntu)
+        if: matrix.os == 'ubuntu-latest'
+        shell: bash
+        run: |
+          sudo apt-get update
+          sudo apt-get install -y ffmpeg
+
+      - name: Install ffmpeg (macOS)
+        if: matrix.os == 'macos-latest'
+        shell: bash
+        run: brew install ffmpeg
+
+      - name: Install ffmpeg (Windows)
+        if: matrix.os == 'windows-latest'
+        shell: pwsh
+        run: choco install ffmpeg -y
 
       - name: Run pytest
         shell: bash # Stops Windows hosts from using PowerShell

.gitignore

@@ -13,7 +13,6 @@ darwin_py.egg-info/PKG-INFO
 *.jpg
 *.bpm
 *.mov
-*.mp4
 *.txt
 
 # from https://raw.githubusercontent.com/github/gitignore/master/Python.gitignore

Dockerfile

@@ -23,6 +23,7 @@ RUN apt-get update && apt-get install -y --no-install-recommends \
     libssl-dev \
     python3-dev \
     curl \
+    ffmpeg \
     && rm -rf /var/lib/apt/lists/*
 
 # Install Poetry in a known location and add to PATH

README.md

@@ -15,15 +15,16 @@ Darwin-py can both be used from the [command line](#usage-as-a-command-line-inte
 
 Main functions are (but not limited to):
 
-- Client authentication
-- Listing local and remote datasets
-- Create/remove datasets
-- Upload/download data to/from remote datasets
-- Direct integration with PyTorch dataloaders
+-   Client authentication
+-   Listing local and remote datasets
+-   Create/remove datasets
+-   Upload/download data to/from remote datasets
+-   Direct integration with PyTorch dataloaders
+-   Extracting video artifacts
 
 Support tested for python 3.9 - 3.12
 
-##  🏁 Installation
+## 🏁 Installation
 
 ```
 pip install darwin-py
@@ -43,11 +44,14 @@ If you wish to use video frame extraction, then you can use the `ocv` flag to in
 pip install darwin-py[ocv]
 ```
 
+If you wish to use video artifacts extraction, then you need to install [FFmpeg](https://www.ffmpeg.org/download.html)
+
 To run test, first install the `test` extra package
 
 ```
 pip install darwin-py[test]
 ```
+
 ### Development
 
 See our development and QA environment installation recommendations [here](docs/DEV.md)
@@ -132,8 +136,8 @@ For videos, the frame rate extraction rate can be specified by adding `--fps <fr
 
 Supported extensions:
 
-- Video files: [`.mp4`, `.bpm`, `.mov` formats].
-- Image files [`.jpg`, `.jpeg`, `.png` formats].
+-   Video files: [`.mp4`, `.bpm`, `.mov` formats].
+-   Image files [`.jpg`, `.jpeg`, `.png` formats].
 
 ```
 $ darwin dataset push test /path/to/folder/with/images
@@ -174,7 +178,6 @@ A minimal example to download a dataset is provided below and a more extensive o
 
 [./darwin_demo.py](https://github.com/v7labs/darwin-py/blob/master/darwin_demo.py).
 
-
 ```python
 from darwin.client import Client
 

darwin/cli.py

@@ -101,6 +101,18 @@ def _run(args: Namespace, parser: ArgumentParser) -> None:
 
     elif args.command == "convert":
         f.convert(args.format, args.files, args.output_dir)
+    elif args.command == "extract":
+        if args.extract_type == "video-artifacts":
+            f.extract_video_artifacts(
+                source_file=args.source_file,
+                output_dir=args.output_dir,
+                storage_key_prefix=args.storage_key_prefix,
+                fps=args.fps,
+                segment_length=args.segment_length,
+                repair=args.repair,
+            )
+        else:
+            parser.print_help()
     elif args.command == "dataset":
         if args.action == "remote":
             f.list_remote_datasets(args.all, args.team)

darwin/cli_functions.py

@@ -60,16 +60,17 @@
 )
 from darwin.exporter import ExporterNotFoundError, export_annotations, get_exporter
 from darwin.exporter.formats import supported_formats as export_formats
+from darwin.extractor import video
 from darwin.importer import ImporterNotFoundError, get_importer, import_annotations
 from darwin.importer.formats import supported_formats as import_formats
 from darwin.item import DatasetItem
 from darwin.utils import (
+    BLOCKED_UPLOAD_ERROR_ALREADY_EXISTS,
     find_files,
     persist_client_configuration,
     prompt,
     secure_continue_request,
     validate_file_against_schema,
-    BLOCKED_UPLOAD_ERROR_ALREADY_EXISTS,
 )
 
 
@@ -1468,3 +1469,41 @@ def _console_theme() -> Theme:
 
 def _has_valid_status(status: str) -> bool:
     return status in ["new", "annotate", "review", "complete", "archived"]
+
+
+def extract_video_artifacts(
+    source_file: str,
+    output_dir: str,
+    storage_key_prefix: str,
+    *,
+    fps: float = 0.0,
+    segment_length: int = 2,
+    repair: bool = False,
+) -> None:
+    """
+    Generate video artifacts (segments, sections, thumbnail, frames manifest).
+
+    Parameters
+    ----------
+    source_file : str
+        Path to input video file
+    output_dir : str
+        Output directory for artifacts
+    storage_key_prefix : str
+        Storage key prefix for generated files
+    fps : float, optional
+        Desired output FPS (0.0 for native), by default 0.0
+    segment_length : int, optional
+        Length of each segment in seconds, by default 2
+    repair : bool, optional
+        Whether to attempt to repair video if errors are detected, by default False
+    """
+
+    video.extract_artifacts(
+        source_file=source_file,
+        output_dir=output_dir,
+        storage_key_prefix=storage_key_prefix,
+        fps=fps,
+        segment_length=segment_length,
+        repair=repair,
+    )

darwin/extractor/__init__.py

@@ -0,0 +1 @@
+"""Video extraction functionality for Darwin."""