3.1.0

Changed

• Add --language-threshold CLI argument for considering languages that appear in at least specified percentage of videos in compute_zim_languages (#212)
• Circular back-to-top button and replaced text characters with icon buttons for better visual consistency (#112)
• Allow users to only get specific video from TED instead extract the entire topic or playlist (#237)

Fixed

• Do not preload all subtitles in video.JS (#220)
• Retry video info extract on missing videoData (#222)
• Do not try to download inappropriate speaker image URL (#224)
• Fix pagination text duplication during page navigation (#217)
• Sub title selection should be done according to the language selection (#207)
• Select language is not persistent enough (#208)
• Style : added minimum height for grid (#239)
• Upgrade Video.JS to 8.23.3 (#247)
• Translation of video title and description is not working on PWA reader (#188)

3.0.3

Changed

• Change log level of "Video at {url} has not yet been translated into {requested_lang_code}" messages from warning to debug (way too verbose)

Fixed

• Restore functionality to resist temporary bad TED responses when parsing video pages (#209)

3.0.2

Changed

• Updgrade to zimscraperlib 3.4.0 (including new webm encoder presets to migrate to VP9 instead of VP8) (#204)

Fixed

• Retry logic is still failing because req might be null when timeout occurs (#203)
• Typo in disable-metadata-checks arg in ted2zim-multi (#202)

3.0.1

Fixed

• Change log level from ERROR to WARNING for missing translations (#197)
• Fix HTTP retries to consider any HTTP failure, not only bad HTTP status code (#162)

3.0.0

Added

• New --long-description CLI argument to set the ZIM long description
• New --disable-metadata-check CLI argument to disable the metadata checks which are automated since zimscraperlib 3.x
• When --languages CLI arugment is not passed, no filtering by language is done (#171)

Changed

• Changed default publisher metadata from 'Kiwix' to 'openZIM'
• Validate ZIM metadata as early as possible
• Migrate to zimscraperlib 3.3.2 (including new VideoLowWebm encoder preset version 2)
• Upgrade Python dependencies, including migration to Python 3.12

Fixed

• Fix language metadata computation (#172)
• Fix computation of automatic description and long description
• Fix subtitles time offset (#177)
• Fix rare bug in display of videos title and description on video page
• Fix support for Youtube fallback when download video from TED CDN is not working (#164 + #182)
• Do not include videos which failed to be fetched / processed in the final list of videos on main page (#167, #169)
• Fix video not working on Safari iOS / iPad (#145)

2.1.0

Changed

• fixed search by topic to use new search API instead of broken web page scraping (#149)
• download_link is renamed request_url and can also perform POST requests (in addition to previous GET requests)
• upgrade to Python 3.11 from 3.8
• upgrade to zimscraperlib 2.1 + upgrade all other dependencies
• significant refactoring to adopt openZIM Python conventions
• activate stale bot + add convenient pull requests template

2.0.13

Changed

• download_link now retries all errors but 404
• Updated ogv.js to 1.8.9
• Fixed missing speaker photo (#144)

2.0.12

Changed

• Fixed crash on videos without speakers (#134)
• Adapted for no-namespace ZIM (#139)

2.0.11

• Fixed new video DOM change
• Fixed dependency issue (markupsafe)
• Don't fail on missing whoTheyAre
• Updated scraperlib (1.6.2) to fix mime guessing bug
• Removed inline JS to comply with some CSP (#128)