text-blackout-memes

Find out what are possibilities to create a meme by crossing-out letters from text on image.

A part of text-blackout meme series.

Not funny	Funny

slangs.txt contains sample phrases web-crawled from polish slang dictionary used to check their presence in ordered input string provided to python script.

Requirements

Python 3.X

$ python setup.py install

Pytesseract installation

Install tesseract

tl;dr

Linux

$ sudo apt install tesseract-ocr
$ sudo apt install libtesseract-dev 

MacOS

$ brew install tesseract

Windows

• Create config.json file in main folder with content presented in config.example.json file
• Download EXE here
• Specify path to tesseract executable installed in config.json

Using Docker

Image available at https://hub.docker.com/repository/docker/hvvka/text-blackout-memes

Build locally

$ docker build -t memes_image .
$ docker run -it --name memes_container memes_image

Usage notes

find_words

Checks input string for a given file with dictionary words and outputs those which are present in string + their positions.

I.e. default_string in find_words.py for given slangs.txt outputs output.txt.

3 arguments are possible, but optional:

• --input – input string provided from image, default: wałpzwpowodzoadinisttojewódzkizarządmelioracjiiurządzeńwodnywrocławulmatejki5aazrozkopywaniawbijanisszkadzaadarninyiinnychnpodstaapaez

• --dict – file with dictionary words, default: resources/slangs.txt

• --out – output file, default: resources/output.txt

If dictionary file is not present, then it uses polish_slang_crawler script to fetch words. The script can we modified to use different source.

Use output file to find the most suitable phrase and modify the image manually.

image_analyzer

Manual image modification is not required with this one.

1. Run image_anaylzer.py and wait for gigabytes of data with mentzen memes