Pipeline for Extracting Metadata¶

A pipeline for building an interactive website from media files.

Offers similar functionality to the command line interface from within Python.

class dvt.pipeline.csv.VideoCsvPipeline(finput, dirout=None, diff_cutoff=10, cut_min_length=30, frequency=0, path_to_faces=None, path_to_audio=None, path_to_subtitle=None)[source]¶

Bases: dvt.abstract.Pipeline

Contains annotators to return structured data about video inputs.

finput¶

path to the input video file

Type:	str

dirout¶

output directory. If set to None (default), will be a directory named “dvt-output-data” in the current working directory

Type:	str

diff_cutoff¶

difference cutoff value; integer from 0-256; higher values produce fewer cuts.

Type:	int

cut_min_length¶

minimum length of a detected cut in frames; higher values produce few cuts.

Type:	int

freq¶

set to a positive integer to select images based on frequency rather than detecting cuts; integer gives frequency of the sampling

Type:	int

path_to_faces¶

Path to directory containing protype faces (optional).

Type:	str

path_to_audio¶

Path to a wav file with audio data. See tutorial on the commandline interface for more details.

Type:	str

path_to_subtitle¶

Path to a src file with subtitle data. See tutorial on the commandline interface for more details.

Type:	str

static get_argparser()[source]¶: Return an argument parser class that can be used from the CLI.

run()[source]¶: Run the pipeline over the input video object.