Open
Conversation
nhannht
added a commit
to nhannht/obsidian-historica
that referenced
this pull request
Apr 15, 2026
- vendor/chrono: fork of wanasit/chrono as git submodule on feat/vi-locale branch; contains full VI locale (11 parsers, 3 refiners, 41 tests) - package.json: switch chrono-node to file:./vendor/chrono so VI locale is available at runtime - tsconfig.json: exclude vendor/ from root tsc to prevent test-file errors - ChronoParser.ts: add setupCustomChronoVi() and vi case in getParserForLanguage(); vie already mapped in FRANC_TO_LOCALE PR to upstream: wanasit/chrono#641 Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
wanasit
reviewed
Apr 19, 2026
Owner
wanasit
left a comment
wanasit
left a comment
There was a problem hiding this comment.
Thanks for the change. Could you remove some of the changes I mentioned in the comment?
Owner
There was a problem hiding this comment.
I don't think this test and its fixtures suit the project.
- It doesn't test specific patterns we want to support.
- Ensure that it's 85%+ accuracy is a corpus is not useful, and would be difficult to debug if failed.
- Embedding a corpus also make the project too large.
Author
There was a problem hiding this comment.
Thank you for the feedback — completely agree on all three points.
Removed test/vi/vi_corpus.test.ts and the entire test/vi/fixtures/ directory (19 Wikipedia articles + curated JSON). The corpus added ~13k lines and wasn't testing specific parser patterns.
The VI locale now has 92 targeted unit tests across 16 suites, covering all 11 parsers and 3 refiners — including strict mode, forwardDate, isCertain() assertions, and negative cases. All tests follow the same testSingleCase/testUnexpectedResult conventions used by the EN locale.
| "outDir": "dist/cjs", | ||
| "module": "commonjs" | ||
| "module": "commonjs", | ||
| "typeRoots": ["./node_modules/@types"] |
Owner
There was a problem hiding this comment.
Please do not modify the compiler option and dependencies in the same commit. If you think this necessary or good for the project, please create a separate CL.
Author
There was a problem hiding this comment.
Done — reverted tsconfig.build.json to match upstream. The typeRoots change was needed to resolve a build issue on my end, but I understand the preference to keep compiler/dependency changes in a separate CL. I'll submit that separately if needed.
Implements full Vietnamese date/time parsing for chrono-node with EN-locale parity. Parsers cover all major Vietnamese temporal patterns extracted from 19 Wikipedia war/history articles (1132 annotated fixtures). Parsers: - VIStandardParser — ngày D tháng M năm YYYY / D tháng M năm YYYY - VIMonthYearParser — tháng M năm YYYY - VIYearParser — năm YYYY, năm N TCN (BC) - VICasualDateParser — hôm nay, hôm qua, ngày mai, ngày kia, bây giờ - VICasualTimeParser — buổi sáng/trưa/chiều/tối/đêm, nửa đêm - VIWeekdayParser — thứ Hai–CN, t2–t7/cn abbreviations - VITimeExpressionParser — X giờ Y phút, HH:MM, lúc/vào prefixes, meridiem - VITimeUnitAgoFormatParser — N ngày/tuần/tháng/năm trước/qua - VITimeUnitLaterFormatParser — N ngày/tuần/tháng/năm sau/nữa/tới - VITimeUnitWithinFormatParser — trong (vòng) N ngày/tuần/tháng - VITimeUnitCasualRelativeFormatParser — tuần này/trước/tới, tháng sau Refiners: - VIMergeDateTimeRefiner, VIMergeDateRangeRefiner, VIMergeWeekdayComponentRefiner Common parsers inherited: ISOFormatParser, SlashDateFormatParser (DD/MM/YYYY) Tests: 10 test files, 41 cases, all passing. Wikipedia corpus: 19 articles, 1132 annotated date fixtures in test/vi/fixtures/wikiwars_vi_curated.json Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- VIStandardParser: improve boundary handling - VICasualTimeParser: fix meridiem mapping and implied hours - VITimeExpressionParser: correct colon-format minute group - VIWeekdayParser: fix modifier detection - vi_standard.test.ts: remove unreliable ngày-32 partial-match assertion - vi_weekday.test.ts: align modifier expectations - wikiwars_vi_curated.json: minor annotation correction
… typo Non-capturing (?:...) on modifier group meant match[MODIFIER_GROUP] was always undefined, making next/last weekday logic dead code. Changed to a capturing group so modifier text is available. Also corrected the typo 'quả' (fruit) → 'qua' (past) in the last-weekday branch. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
…er group index TIME_UNITS_PATTERN contains zero capturing groups (repeatedTimeunitPattern strips them all to non-capturing). Group layout is: 1 = prefix modifier, 2 = unit (prefix form) 3 = unit (suffix form), 4 = suffix modifier match[5] was off by one, so suffix-form 'tuần trước' / 'tháng qua' always fell through to an undefined modifier and produced a future date instead of past. Fixed to match[4]. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
In chrono's 12-hour convention, AM = hours 0–11, PM = hours 12–23. Noon (12:00) is PM. The AM assignment caused noon to be interpreted as midnight in downstream meridiem-aware code. Added inline comment to explain the convention for future readers. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
… never used All three relative-unit parsers accepted a strictMode constructor param but innerPattern() always returned the casual PATTERN, making the parameter dead code. Added STRICT_PATTERN (aliased to PATTERN — VI has no unit abbreviations so both modes are identical) and switched innerPattern() to return the correct variant. Matches the API contract established by EN/IT locales. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
STRICT_PATTERN was assigned as an alias for PATTERN (since VI has no unit abbreviations) and then used in a ternary that always evaluated the same branch. Remove the alias and dead conditional; move the explanatory comment onto innerPattern() where it belongs. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
vi_corpus.test.ts: exercises all 1132 WikiWars-VI fixtures covering full_date, month_year, year_only, slash_date, and bc_year expression types. Accuracy: 1132/1132 (100%). VIYearParser: extend pattern to also match bare 'YYYY TCN' without the 'năm' prefix (e.g. '179 TCN'). Previously only 'năm YYYY TCN' was supported, leaving bare BC year expressions unparsed. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
…s, number-words, weekday modifiers New test files: - vi_month_year.test.ts: VIMonthYearParser (previously zero coverage) — tháng M năm YYYY, tháng M/YYYY slash form, month-only implies year, month > 12 rejected - vi_casual_time.test.ts: VICasualTimeParser — trưa=PM, bình minh/sáng sớm=AM, chiều/tối/đêm=PM, nửa đêm=AM, buổi prefix, date+time merge - vi_negative_cases.test.ts: invalid day/month, invalid slash date, bare 4-digit number, phone number false positive Expanded existing files: - vi_weekday.test.ts: next (tới/sau) and last (qua) modifier assertions with concrete expected dates - vi_time_units_ago.test.ts: number-word durations (hai/ba/một) - vi_time_units_later.test.ts: number-word durations (ba/hai) Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- VITimeExpressionParser: move trưa to PM branch (1 giờ trưa → 13:00); add hour=0 guard for 12 giờ sáng → midnight - VIWeekdayParser: fix capturing group for modifier; correct quả→qua typo; add negative lookahead for 'sau khi' conjunction - VITimeUnitCasualRelativeFormatParser: make number optional so bare 'tuần này'/'tháng trước' match; use references helpers in casual date - VIYearParser: support bare 'YYYY TCN' without năm prefix (e.g. 179 TCN) - VIMonthYearParser: wire MONTH_DICTIONARY for word-form months (tháng ba, tháng giêng, tháng chạp); use references.yesterday/tomorrow - VICasualDateParser: add hôm kia (-2 days); use references helpers - README: add vi to supported locales list (lines 40 and 215) - Tests: add vi_date_range, vi_casual_time, vi_month_year, vi_negative_cases; expand vi_time_exp, vi_weekday, vi_time_units_* (15 test files, 77 tests total) Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
…ayParser Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Maintainer feedback: corpus benchmark tests don't fit the project's testing philosophy, and embedding the corpus makes the project too large. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Maintainer wants compiler/dependency changes in a separate CL. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Adds expect(r.index).toBe(...) checks to sentence-embedded test cases across casual, standard, time expression, weekday, slash, ago, and later test files — matching the pattern used in EN/DE test suites. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Closes test coverage gaps identified by comparing EN (154 tests) with VI: - vi_strict.test.ts: verify strict mode rejects casual/weekday-only expressions while accepting standard dates and explicit time units - vi_forward_date.test.ts: verify forwardDate option rolls time-only, weekday, slash date, and month expressions to future dates - vi_casual.test.ts: add isCertain assertions for casual date components - vi_time_exp.test.ts: add isCertain assertions for hour and meridiem - vi_standard.test.ts: add isCertain assertions for full and partial dates - vi_negative_cases.test.ts: add bare numbers, currency, version numbers, hyphenated ranges, and URL-encoded string rejection tests VI locale: 75 → 92 tests across 14 → 16 suites. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Author
|
Hi @wanasit, thank you for the review. I've addressed both comments:
Additionally:
All 92 tests pass. Will mark as ready for review once I've done a final check. |
… drift - VITimeExpressionParser: separate "trưa" handling from "chiều/tối/đêm" — "11 giờ trưa" is 11 AM (approaching noon), not 23:00 - VICasualTimeParser: remove redundant \b from PATTERN — JS \b fails on Vietnamese đ (non-ASCII), breaking standalone "đêm" parsing. AbstractParserWithWordBoundaryChecking already provides left-boundary - Revert package-lock.json to upstream (unrelated lockfile drift) Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
…edge cases - VITimeExpressionParser: reject minute >= 60 (matches AbstractTimeExpressionParser) - vi_casual: add toBeDate() assertions for today/yesterday/tomorrow/now - vi_forward_date: add same-weekday edge case (stays on same day) - vi_strict: add slash dates acceptance test (30/4/1975, 15/3) - vi_negative_cases: add impossible minute test (61, 99) Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Author
|
Pushed two additional commits addressing issues found during self-review:
Test count: 95 tests across 16 suites, all passing. Full suite (610 tests / 130 suites) also clean. |
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.
Summary
This PR adds a full Vietnamese (
vi) locale to chrono-node, bringing it to parity with the English locale in terms of parser and refiner coverage.What's included
11 parsers:
VIStandardParser—ngày D tháng M năm YYYY(with/withoutngàyprefix, optional year)VIMonthYearParser—tháng M năm YYYY/tháng M/YYYYVIYearParser— standalonenăm YYYYand bareYYYY TCN(BC years, nonămprefix required)VICasualDateParser—hôm nay,hôm qua,ngày mai,ngày kia,bây giờVICasualTimeParser—sáng,trưa,chiều,tối,đêm,nửa đêm,bình minh,sáng sớmVITimeExpressionParser—lúc/vào 7 giờ 30 phút chiều,15:30VIWeekdayParser—thứ hai–thứ bảy,chủ nhậtwithtới/sau/quamodifiersVITimeUnitAgoFormatParser—3 ngày trước,1 tháng quaVITimeUnitLaterFormatParser—2 tuần sau,3 ngày nữaVITimeUnitWithinFormatParser—trong vòng 2 giờVITimeUnitCasualRelativeFormatParser—tuần này/trước/sau,tháng này/trước/sau(number optional — bare unit words supported)3 refiners:
VIMergeDateTimeRefiner— merges date + time results separated bylúc/vàoVIMergeDateRangeRefiner— merges date ranges separated by–/đến/tới/vàVIMergeWeekdayComponentRefiner— standard weekday merge16 test files, 95 tests covering standard dates, casual dates/times, time expressions, weekdays, slash dates, date ranges, month/year, year, time units (ago/later/within/casual-relative), strict mode, forward date, negative cases, and isCertain/toBeDate assertions.
Design notes
YEAR_PATTERNsupports BC years viaTCNsuffix (trước Công nguyên);VIYearParseralso handles bareYYYY TCNwithout anămprefix (e.g."179 TCN")parseYear()callsfindMostLikelyADYear()for short years (e.g.1945stays1945,75→1975)VITimeUnitCasualRelativeFormatParseruses an optional-number pattern so bare unit words (tuần này,tháng trước) match without a numeric prefix; defaults to quantity 1strictModeconstructor parameter on all relative unit parsers (consistent with DE/FR)src/index.tsBug fixes included (found during self-review)
VITimeExpressionParser:trưameridiem was incorrectly grouped withchiều/tối/đêm—11 giờ trưareturned 23:00 instead of 11:00 AM. Fixed by separatingtrưahandling: hour < 10 → PM (+12), hour 10-11 → AM (keep as-is), hour 12 → PM (noon)VITimeExpressionParser:12 giờ sángreturned noon instead of midnight — addedif (hour === 12) hour = 0guard matching EN conventionVITimeExpressionParser: added minute validation — reject minute >= 60 (matchingAbstractTimeExpressionParserbehavior)VICasualTimeParser: removed redundant\bfrom PATTERN — JS\bfails on Vietnameseđ(non-ASCII), silently breaking standaloneđêmparsing.AbstractParserWithWordBoundaryCheckingalready provides left-boundary via(\W|^)VIWeekdayParser: modifier group was non-capturing, makingqua(last) undetectable; fixed to capturing group; also correctedquảtypo →quaVITimeUnitCasualRelativeFormatParser: baretuần này/tháng trướcnever matched due to required numeric prefix inTIME_UNITS_PATTERN; replaced with optional-number patternpackage-lock.json: reverted to upstream — removed unrelated lockfile drift (dayjs removal, peer dependency changes)🤖 Generated with Claude Code