These ideas were initially compiled for GSoC 2021 but as SMC wasn’t included it is kept for anyone to work on.
Please reply here if you want to include an idea. In your response, please include these details:
Project name
Context and background
Complexity
Mentor(s) and their availability
Skills that the student will require to be able to work on the project
Related links and reading material
If you have any questions or want to discuss, please talk in #smc-project on freenode (matrix link). (You can, of course, use the development category of this discourse for long form discussions)
dictmaker is a standalone web server app for publishing online dictionaries like Olam and Alar. It can be used to build dictionaries for any language, where language specific text processing and search semantics are added as independent plugins. The app is written in Go and data is stored in a couple of Postgres tables. The project’s README describes the concepts and architecture in detail.
It would be very useful to have an admin frontend / dashboard on top of dictmaker for managing dictionary entries (add, remove, search, edit, moderate public submissions). For instance, on Olam, there a number of public submissions (new words, definitions, and corrections) that come through regularly. A UI tailor made for easily moderating such submissions can add great value to a dictionary website.
What is involved
Minor changes to the Postgres table schema.
Addition of CRUD handlers and SQL operations to the Go app (eg: add, remove, delete, approve/reject).
A web based dashboard as a management frontend to the CRUD operations (preferably in VueJS).
Skills
Understanding of web apps and CRUD in general.
Go, HTML + Javascript. Understanding of a frontend framework like VueJS.
Low to medium complexity.
Mentoring
Hoping that there are candidates with the prerequisite know-how, I should be able spend ~2-3 hours hands-on per week with them.
Jotting down ideas from Indic Keyboard discussion group and bug tracker
Indic keyboard ideas
Emoji Search and suggestions
Indic Keyboard supports a large number of emojis. There is a basic search feature in the beta version. These deliverables maybe considered:
Improve accuracy and ease of emoji search (search by multiple keywords, multilingual search?)
Better suggestions strip view: For long words the current suggestion view is not good. Especially for Malayalam, making long agglutinated words are good in current Indic Keyboard.
The project attempts to create a platform for collecting free licensed speech corpus for various research purposes. To help crowdsourcing the speech, a simple mobile friendly web application is provided.