A simple wordcloud API based on amueller/word_cloud.
Wordcloud API is a simple, customizable wordcloud generation API that doesn't require Internet access, can be self-hosted and returns a word cloud.
- Looking for a simple API that could return customized wordcloud images didn't turn out well.
- Quickchart doesn't have wordcloud within their open-sourced codebase. My use case involved hosting an API within a self-contained, isolated (no Internet access) environment.
- Community version of Quickchart APIs is restricted by a rate-limit of 1 wordcloud per sec per IP address. Also, I didn't want to pay for unlimited hits. Solution: Build my own API ;).
The docker image rohanlekhwani/wordcloud-api has all the dependencies installed and can be pulled and set to run instantly.
docker pull rohanlekhwani/wordcloud-api
In case, you need to change parameters to the wordcloud class and build a new image on-premise, a Dockerfile has been provided.
Run
docker run -p 8000:8000 rohanlekhwani/wordcloud-api
The wordcloud can be generated by making a GET request to the /wordcloud endpoint with wordlist within the text parameter.
http://localhost:8000/wordcloud?text="english hindi spanish portuguese german mandarin"
This would return the wordcloud JPEG image.
In case, you don't want to host it on-premise, I have a hosted version at the following URL
https://rohanlekhwani.pythonanywhere.com/wordcloud
This API endpoint theoretically does not have any rate limits (Yay!) unlike Quickchart APIs. Feel free to use it in your projects.
Currently, customizations only for the word list has been provided. I've hardcoded other customizations according to my use case.
A major scope of improvement is to make use of all parameters of the Wordcloud class within the API endpoint. For example, the current min_font_size is hardcoded to 2. This can be accepted as a parameter within the API endpoint.
http://localhost:8000/wordcloud?text="english hindi spanish portuguese german mandarin"&min_font_size=10
Accepting feature requests and bugs on the issues page.
Please star this repository if this project helped you!