Releases: toolforge/tool-spacemedia
Releases · toolforge/tool-spacemedia
spacemedia-0.4.1
New sources
- NASA/ESA James Webb Space Telescope: https://webbtelescope.org & https://esawebb.org
- NASA ASTER (Advanced Spaceborne Thermal Emission and Reflection Radiometer instrument abord Terra satellite): https://asterweb.jpl.nasa.gov
- NASA JPL Photojournal: https://photojournal.jpl.nasa.gov
- NASA SDO (Solar Dynamics Observatory): https://sdo.gsfc.nasa.gov
- NOIRLab: https://noirlab.edu
- Individuals on Flickr: Judy Schmidt, Kevin Gill, Pierre Markuse, Andrea Luck, Harry Stranger
- New US military units; merged all of them in a single "US Space Force/Command" source
- Enable automatic upload for:
- US military (DVIDS & Flickr), DLR, KARI, James Webb Space Telescope (ESA & NASA): all files
- NASA, IAU, ESO, NoirLab, Hubble (ESA & NASA), Individuals : only files published after 2022 for now
- Enable manual upload of other sources
Major features
- Tweet bot activity on Mastodon and Twitter
- Add SDC (Structured Data on Commons) for uploaded files
- Start a complete calculation of perceptual hashes on the whole Commons database to detect duplicates (still ongoing as of May 2023...)
- Report exact duplicate files to Commons administrators by parsing Special:ListDuplicatedFiles (up to 190 files max to avoid spamming them too much)
- Translate non-English text using Google Translate
- Support WebP images files
- Blocklist of terms implying uninteresting content published by US military and NASA
- New remote capabilities to compute hashes on distant computers more powerful than toolforge pods
- Switch spring schedules to toolforge jobs framework and Cloud VPS cronjobs
- Support extracting information from Wikidata (ISS crew members, astronomical objects, telescopes, instruments...)
- Detect and ignore courtesy photos in media published by US military and NASA (from ULA, SpaceX, Lockheed Martin...) using a blocklist of terms in media description and a blocklist of photographers in EXIF metadata
- Look up for (NASA) images on Commons by their id to avoid upload not-exact duplicates and upload high-resolution version if needed
Minor features
- Allow to manually refresh a media by reassessing all its metadata
- Display a video/audio icons above preview images
- New REST endpoint to return commons last timestamp
- New REST enpoint to put a new hash association
- NASA: Extract metadata for ISS and Artemis images
- Initial support of upload in chunks for very large files exceeding memory on Wikimedia servers. Does not work for now :(
Behind the scene
- Update to Java 17 and Spring Boot 2.7
- Update to latest versions of Mediawiki
- Update to MariaDB 10.4: https://phabricator.wikimedia.org/T301949
- Update to breaking changes on hubblesite.org website
- Update to breaking change on DVIDS CDN
- Migrate from Phabricator Diffusion to Wikimedia GitLab
- Migrate from eqiad.wmflabs to wikimedia.cloud
- Hubble and Webb NASA websites handled as a single "STScI" repository
- IAU, ESO, NoirLab websites handled as "Djangoplicity" repositories
- Use JPEG plugin from twelve-monkeys in order to read more files
- Disabled video support on toolforge, requires too much memory
- As usual, lots of general performance/reliability improvements and dependencies upgrades
Full Changelog: spacemedia-0.4.0...spacemedia-0.4.1
v0.4.0
Major features
- Automatic upload of ESA media
Minor features
- General reliability/performance improvements
Behind the scene
- Handling URL change from wmflabs.org to toolforge.org
Full Changelog: spacemedia-0.3.0...spacemedia-0.4.0
v0.3.0
First version performing automatic upload
Features completed
Major features
- Detect of almost-duplicate images using percerptual hashes
- No longer consider Flickr Public Domain Mark eligible for upload, unless clear U.S Public Domain
- Replace Flickr image title consisting of VIRIN identifiers by their album name, if any
- Enforce minimal delay between successive uploads to meet Wikimedia bot requirements
- Upload mode for each agency: disabled (default), manual, automatic
- Handle change of license for Flickr and YouTube media
- Add DVIDS support for U.S Space Command, Space Force, Air Force Space Command, Space and Missile Systems Center
- Enable automatic upload for images of U.S Space Command and Air Force Space Command
- Add YouTube support for Arianespace
- Download of YouTube videos using youtube-dl
- Retrieval of already upload YouTube videos using their id
- Startup page improvements: AJAX loading of statistics, sortable table
Minor features
- Number of working threads configurable
- Nice error pages for HTTP 404, 500, 501 errors
Behind the scene
- Allow to reset problems, duplicates, ignored media, perceptual hashes
- Mirroring from Phabricator to GitHub
- Dependencies upgrade: Spring Boot 2.2.7, Scribe 6.9.0, Flickr4Java 3.0.4, bootstrap 4.5.0
- Retrieval of web dependencies through Webjars
- Store last update date/time for each media
- Store runtime information (last update start time, end time, duration) for each agency
- Store date/time for each problem
- Rework persistence of metadata information
- Lots of bugfixes, performance improvements, reduced memory usage
Features started
- GitHub actions
- SonarCloud analysis
- Code cleanup, test coverage, documentation