Debloat any Android or Google TV over Wi-Fi and get an Apple TV-style home.
Python adbadblockandroid-tvandroid-tv-launcherapple-tvbloatware-removalclaude-codedebloatfire-tvgoogle-tvlauncherprivacysmart-tvsony-braviatclvibe-codingxiaomi
Workflow: Add a press entry
When to use this: an article, or a YouTube video, about this project.
- Add a row
| Outlet | [Headline](<url>) | Date |to the right country section ofdocs/PRESS.md, in the headline's own language. The link must open and show the article: no syndicated copies, no auto-generated listings, nothing that cannot be checked. - Update the counts in every place
tests/test_press_counts.pychecks: thedocs/PRESS.mdintro line, its index line, the card's alt text in everyREADME*.mdthat showsdocs/card.jpg- todayREADME.md("and N more: N articles from N countries"), the DutchREADME.nl.md("nog N media: N artikelen uit N landen") and the GermanREADME.de.md("und N weitere: N Artikel aus N Ländern"), the UkrainianREADME.uk.md("та ще N видань: N статей із N країн") and the TurkishREADME.tr.md("ve N yayın daha: N ülkeden N makale", the country count before the article total) - andMOREintools/readme_featured.py. Every page shows the same picture, so a translation that copies it states the same numbers in its own words: the pages named here are the ones that carry it today, not the rule. The end of each page's story section repeats the three numbers once more, in prose ("The coverage: articles from countries and videos on YouTube and Bilibili, all indocs/PRESS.md"), and the same test reads that sentence too: the page states the count twice, once in the card's alt text and once in that sentence, so a reader who sees one number in the picture and another in the story section is looking at a page that was half updated. Leave the counts out of the sentence and the test says the paragraph no longer carries them. That sentence dates the numbers as well, in a parenthetical just after the press link ("(numbers as of 31 October 2026)", "(Stand: 31. Oktober 2026)", "(stand 31 oktober 2026)", "(станом на 31 жовтня 2026 року)", "(rakamlar 31 Ekim 2026 itibarıyla)", "(dados de 31 de outubro de 2026)"), so the day moves with the sweep: it may not name a day earlier than the newest article row indocs/PRESS.md, the rule step 3 already applies to "views as of", because a count read on the 30th cannot hold an article published on the 31st.tests/test_press_as_of_date.pyreads that parenthetical in each page's own words, off every page whose story paragraph carries the link, and a page that writes none is not judged:README.he.mddates the stars it had in the sentence before, and its coverage sentence ends at the link. That paragraph is also read on the two front pages that have it without the card:README.pt-BR.mdand the HebrewREADME.he.mdshow nodocs/card.jpg, so their numbers are held toREADME.md's alt text and todocs/PRESS.mdon their own, and their coverage clause has to open with a colon like the others ("A repercussão: 412 artigos ...", "סיקור בתקשורת: 412 כתבות ..."), since that colon is what the reader cuts the three numbers at. A page counts as one of them by its paragraph, so a translation that copies the story without the picture is read without the test being edited. Redrawingdocs/featured.pnganddocs/card.jpgneeds Pillow, which the performance requirements install:tools/venv/bin/pip install -r tools/requirements-perf.txt, thentools/venv/bin/python -I tools/readme_featured.py Fredoka.ttf docs/featured.pngandtools/venv/bin/python -I tools/readme_card.py Fredoka.ttf docs/card.jpg. Fredoka is the variable TTF from Google Fonts and its file is not in the repo, so that first argument is the path of the copy you downloaded; each script says so when the path is wrong. If you cannot redraw them, say so in the PR. - A video goes in the
## Videossection at the end of the same page, not in a country section:| Channel | [Title](<url>) | Date | Views |, in the table's date form (27 Oct), newest first. Two numbers above that table are read from the rows bytests/test_press_counts.pyand move with it: the[N videos]count at the end of the intro line (its link points at the section itself), and the day inviews as of <date>- no earlier than the newest row's date, so a video posted after the last read leaves the line stale ("the views are 'as of 27 Oct' and the newest video is dated 28 Oct"). A row added without moving the count fails as "the intro says 229 videos and the Videos section holds 230 rows".
Then run the tests: python3 -m unittest discover -s tests -v.