emojimix-scraper

Just a fun experiment to generate all possible emoji combinations from https://www.google.com/search?q=combine+emojis

How does it work?

After playing around with the app for a bit with Chrome DevTools open, I was able to determine:

1. A file containing all of the emojis shown in the app (https://www.google.com/logos/fnbx/emoji_kitchen/emoji_kitchen_pairs.3.pb)
2. The URL used to submit the combination request
3. The response JSON which includes a subsequent URL with a PNG of the resulting combination

After that, it was pretty straightforward to enumerate and submit all combinations to the API, parse the response, and fetch the resulting image.

By the numbers

• 234 emojis
• 27495 possible combinations with replacement (also called multichoose -- you are allowed to select the same emoji twice)
• not all combinations are supported; the app will sometimes gray-out a subset of emojis
• the resulting PNG dataset occupies ~360MB of disk space

Usage:

• The only Python dependency is the requsts library
• Running main.py will output each combination to stdout along with the resulting filename, or an error if it fails
• Feel free to adjust MAX_WORKERS; a value of 1000 allowed the script to complete in a few minutes on a 2019 MBP Core i9
• I suggest redirecting the output to a text file; this allows you to later grep for an image filename and see which emojis combined to create it (if you can't tell by visual inspection).

NOTE: There may be a bug in this code somewhere; I got 27451 lines of output instead of the expected 27495 (a delta of 44 combinations). PR welcome if you happen to spot an error 😁