Wikimedia Suomi software project ideas
This page is for documenting potential project ideas for Google Summer Code (2025), Outreachy (round 30), thesis works, interns etc.
Ukbot Development
[edit]Ukbot is a bot which counts scores for editing competitions, developed originally by Danmichaelo and now managed by WMNO (Jon Harald Søby) and WMFI (Zache). WMFI/WMNO are currently focusing on more structured development for it. Primary focus is on bugfixes and new features.
- Project repository: https://github.com/WikimediaNorge/UKBot
- Example competition: https://fi.wikipedia.org/wiki/Wikipedia:Viikon_kilpailu/Viikon_kilpailu_2025-07
- Programming language: Python
- Network/hardware requirements: Low
- Existing coding skill requirement: Low (tasks range from documentation to refactoring)
Cat-a-lot development
[edit]Cat-a-lot is a JavaScript gadget mainly used in Wikimedia Commons for categorizing photos. About 5% of daily Wikimedia Commons users use it. Zache from WMFI fixed some bugs in autumn 2024. However, it doesn't have an active developer and there are bugs to be fixed, so this could be suitable if there's interest for somebody to adopt it.
- Documentation: https://commons.wikimedia.org/wiki/Help:Gadget-Cat-a-lot
- Programming language: JavaScript / HTML / CSS
- Network/hardware requirements: Low
- Existing coding skill requirement: Moderate
Imagehash project
[edit]Wikimedia Finland has a project where we are indexing all Wikimedia Commons images using perceptual hashes. The current main task would be to write a Java version of phash/dhash algorithms that would produce the same results as the Python imagehash library. An alternative task could be implementing the image hashing function used by ISCC to Python's imagehash library.
- https://github.com/Wikimedia-Suomi/ImageHash-Toolforge
- https://github.com/JohannesBuchner/imagehash/issues/212
- Programming language: Python / Java
- Network/hardware requirements: Moderate
- Existing coding skill requirement: Moderate to high
Add RDF as export option to Overpass-turbo
[edit]Currently, there is an option to download OpenStreetMap data in GeoJSON format which can be converted to RDF. However, it would be useful to be able to directly download it in RDF format suitable for importing to local SPARQL tools that support GeoSPARQL, such as Apache Jena. The task would involve modifying Overpass-turbo code to add an RDF export option, configuring Apache Jena to support GeoSPARQL 1.1, and then writing a tutorial about how to import data from Overpass-turbo and Wikidata to Jena-Fuseki and querying combined data using GeoSPARQL.
- https://github.com/tyrasd/overpass-turbo
- https://jena.apache.org/documentation/fuseki2
- Programming language: JavaScript (Overpass), Java (Jena), SPARQL
- Network/hardware requirements: Linux
- Existing coding skill requirement: High
- https://overpass-turbo.eu
Misc ideas
[edit]Misc ideas which are on project plans, but doesn't have any timetable.
- Howto query Toolforge replica databases using Ontop SPARQL server and performance analysis. (Java, SQL, SPARQL)
- Ajapaik Flutter App coding - (Flutter OR Python)
- Importing Ajapaik photos and metadata to Wikimedia Commons (Python)
- WikiScore development (editing competition tool by Wikimedia Brazil)
- Wiki Loves Monuments web app by Wikimedia italy development - source code, backend is Django, frontend is React)
- SeulojaBot refactoring so that it would use LLM:s for vandalism/good/bad edit detection (ie. fiwikis autoreview bot, pywikibot)