alexandrebarral / vigrep

ViGrep

It's a cli to extract text from video by a regex.

Index 📖

• How it works?
• Dependencies
• Install python requirements
• Usage
• Example
• Output

How it works? ⚙️

• Excecute the cli informing your file
• Get the matched words separeted by video seconds in a output.txt file
• Be happy

Dependencies ⚠️

We are using tesseract, it's a free open source Google's OCR, you must install it in your system before continue.

Tesseract installation instructions

Install python requirements 🚧

pip install -r requirements.txt

Usage 🎬

python main.py -h

usage: main.py [-h] [-i INPUT] [-r [REGEX]] [-o OUTPUT] [-v VERBOSE]

optional arguments:
  -h, --help            show this help message and exit
  -i INPUT, --input INPUT
                        path to input video file
  -r [REGEX], --regex [REGEX]
                        regex to find text in video frames
  -o OUTPUT, --output OUTPUT
                        path to (optional) output video file
  -v VERBOSE, --verbose VERBOSE
                        If 1(True) will be printed in output file the text
                        used in regex

Example 👀

python main.py --input "video.mkv" --regex "((https?:\/\/(www)?)?[a-zA-Z0-9][a-zA-Z0-9-]+[a-zA-Z0-9]\.[^\s]{2,}|www\.[a-zA-Z0-9][a-zA-Z0-9-]+[a-zA-Z0-9]\.[^\s]{2,})" --verbose 1

Output 💩

Verbose disabled

frame: 1 | second: 0.001 | matches: [('stackoverflow.com/quest', '', '')]
frame: 2 | second: 0.031 | matches: [('stackoverflow.com/questions', '', '')]
frame: 3 | second: 0.064 | matches: [('stackoverflow.com/questions', '', '')]
frame: 4 | second: 0.096 | matches: [('stackoverflow.com/questions', '', '')]

Verbose enabled

 text: <text extracted by tessaract in the current frame>
frame: 1 | second: 0.001 | matches: [('stackoverflow.com/quest', '', '')]
-------------------------------------------------- 

 text: <text extracted by tessaract in the current frame> 
frame: 2 | second: 0.031 | matches: [('stackoverflow.com/questions', '', '')]
-------------------------------------------------- 

 text: <text extracted by tessaract in the current frame>
frame: 3 | second: 0.064 | matches: [('stackoverflow.com/questions', '', '')]