Repository navigation
i18n(ru): exercise-name pack for all 1,324 exercises - #266
Merged
Merged
Conversation
The third exercise-name pack, built the way pt-BR was: a curated JSON source, a glossary that fixes the terminology, a translate script that only visits IDs the source does not have yet, and a build step that refuses to emit anything less than the complete catalogue. All 1,324 titles. The app shows them the way it shows Portuguese and Hungarian — the Russian title with the English original in parentheses — so nobody loses the term they already search by. The glossary is the part worth reviewing. It decides where Russian has its own established name (hammer curl → молотковое сгибание, good morning → наклон со штангой, skullcrusher → французский жим) and where the gym simply says the English word, in which case inventing a calque would be worse than keeping it (бёрпи, машина Смита, гакк-приседание, EZ-гриф). After the machine pass I normalised what it got inconsistent across batches: sentence case throughout (leaving V-/Y-/EZ- alone), equipment trailing rather than fronted with a comma, and one Russian term per English term where it had drifted (русский твист, выкат). Thirteen titles had dropped a qualifier the English states — the dumbbell in eight wrist curls, assisted, weighted — and one had taken its equipment from the metadata rather than the title. ru-exercise-names.test.js mirrors the pt-BR suite and pins the qualifiers that change what a movement is, so a future batch cannot quietly drop them. These are original translations produced with LLM assistance from the English titles, not copied from another Russian exercise dataset. They have NOT been reviewed line by line by a native speaker — the normalisation pass above and the spot checks were mine, but treat the pack as machine output that a human has tidied, exactly as the pt-BR README asks.
dima-po
force-pushed
the
i18n/ru-exercise-names
branch
from
September 20, 2026 12:14
3444030 to
0a7fe0b
Compare
Contributor
Author
|
Heads-up on overlap with #254, which I found after opening this: it adds de/es/fr/it packs plus the per-language "English name in parentheses" toggle. The two are complementary — a fifth language rather than a competing approach — but they touch the same line:
Whichever you merge first, I will rebase this one on top — just say the word, or merge #254 and I will push the rebase without waiting. |
ilirhushi
added a commit
to ilirhushi/openGym
that referenced
this pull request
Sep 20, 2026
The upstream name packs (DuarteSantos8#245, DuarteSantos8#254, DuarteSantos8#266) only cover the 1,324 exercises shared with upstream; this fork's own 9 additions (landmine row, pike push-up, nordic hamstring curl, meadows row, bird dog, band clamshell, deficit deadlift, band pull-apart, copenhagen plank) had no translation in any of the newly-added languages, failing each pack's completeness test.
Owner
|
Thanks @dima-po, the Russian name pack is in, and the #254 toggles work with it as you expected. Released in v1.3.9: https://github.com/DuarteSantos8/openGym/releases/tag/v1.3.9 |
1 task done
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Russian joins
pt-BRandhuinEXERCISE_NAME_LANGS: all 1,324 titles, shown the way the other two are — Russian first, English original in parentheses, so nobody loses the term they already search by.Built with your pipeline, not a new one:
scripts/exercise-name-sources/ru.jsonpt-BR.jsonscripts/exercise-name-sources/GLOSSARY.ru.mdscripts/translate-ru-exercise-names.mjsscripts/build-ru-exercise-names.mjsfrontend/src/lib/ru-exercise-names.test.jsThe glossary is the part worth reviewing. It draws the line between movements Russian has its own name for (
hammer curl→ молотковое сгибание,good morning→ наклон со штангой,skullcrusher→ французский жим) and the ones where the gym just says the English word, so a calque would read worse than the loanword (бёрпи, машина Смита, гакк-приседание, EZ-гриф).What I fixed after the machine pass, since a batched run drifts:
V-,Y-,EZ-keep their capitals)гантели, жим лёжаwhere the other 1,278 readжим лёжа с гантелямиrussian twist,rollerout)assisted,weighted— and one had taken its equipment from theeqmetadata instead of the titleThe test pins the qualifiers that change what a movement is (equipment, assisted/weighted, male/female), so a future batch cannot quietly drop them again.
npm testinfrontend/passes — 106 files, 1473 tests.On review status, per the caveat in
exercise-name-sources/README.md: these are original translations from the English titles, produced with LLM assistance and tidied by hand. They have not had a line-by-line native-speaker review, so please don't describe them as such. I am happy to do that pass, or to take corrections from Russian-speaking users, if you'd rather merge it as a starting point.