Translate papers, photos, books between 64 languages.
GPL-3.0 License
Benefits:
Here you have the bear minimum if you know Python. Tutorial with additional information about the process, copyrights and lawful bindings for translations can be found on my blog.
Two docx documents opened next to each other:
The following code generates 2 docx files of text scanned from photos. It may also generate translation from .docx file using Google Cloud API (It's free for the first 3 months). If you want to translate more documents or in a specific way, try using Google Document Translation API - it will be quicker.
On initializing you choose whether you want to translate a folder with pictures or a single document file.
If you select the photos, then they need to be located within a folder placed next to translate module (folder). The flow of work is:
In order to make it work a few steps have to be fulfilled:
pip install -r requirements.txt
export GOOGLE_APPLICATION_CREDENTIALS="/path-to-credentials/credentials.json"
gcloud auth login
Then we need to get auth token obtained from command line. In order to do that run:
gcloud auth application-default print-access-token
You should recieve something like:
ya29.c.Kp8BAQhYC2hyHsFuRlpjrJnT0XkT[...]axi0Z_3y1-xdUaodmnnjkmlquF3ol8CInVDYPdJO8jMr3h-alcQUayKAJD_e7cuw...........
Using this token update GOOGLE_API_TOKEN variable in main.py and run it.
TODO: using .env file to load variables
python3 translate
You will be prompted for folder name, you can type herman_hesse which is a name of example folder that comes with this repo.
> python3 translate
Enter name of the folder with images of scanned book: herman_hesse
You can run the script within a folder with multiple books to translate. For example:
herman_hesse/ # Folder with pictures of pages from example book
another_book/ # Folder with pictures of pages from another book
translate / __init__.py
translate / __main__.py
translate / languages.py
LICENSE
README.md
Books are protected with copyrights! Translate only books that are on the public domain. If you want to thank me and/or know more, visit blog post for this code.
If you have questions or request feel free to reach me via contact form.