[openvoiceos-bot]

Problem

test/test_chunking.py::TestChunking::test_word fails on dev (c57316f, and at c8e1e10 before #18, on 3.11 and 3.14): ['2', 'and', '3/4'] != ['2', 'and', '3', '/', '4']. Nobody saw it because build_tests.yml builds and installs the wheel and runs no pytest, and license_tests.yml runs the licence file alone. Evidence: knowledge/wiki/audits/gate-ledger/merged-run-quebra_frases-20.md.

Which side is right

The test. Two assertions written in 2021 (3988f62, e461391) predate two merged tokenizer changes:

• extend tokenizer regex #1 (783a0b6, 2023) added / to the in-number separators of _WORD_REGEX, (?:\d+(?:[,.:\/]\d*)+), with date: 11/11/2023 11/11 as its stated test string. A digit run joined by / is one token, the same as 1.5 and 13:00.
• fix negative float #3 (720185e, "fix negative float", 2023) added -? in front of it, so -2 and -2.5 are one token.

Both were merged and neither touched the test. The consumer agrees with the code: ovos-number-parser reads a fraction as one 3/4 token and splits it on / itself (look_for_fractions(split_list), "list created by splitting on '/'"). Reverting the code instead would split 11/11/2023 into five tokens and -2.5 into two.

Change

• test_word states the contract: 2 and 3/4 is ['2', 'and', '3/4'], on 11/11/2023 at 13:00 keeps the date and the time whole, -2 and -2.5 are one token. Counter-cases beside them: a / between spaces or between letters, and a - set apart by a space, are their own tokens.
• build_tests.yml calls OpenVoiceOS/gh-automations/.github/workflows/build-tests.yml@dev with test_path: 'test/test_chunking.py test/test_tokenization.py' on 3.10 to 3.14, on push and pull request to dev. license_tests.py stays with its own workflow, which installs lichecker.

Checks

• Fail before: at dev the same file gives 1 failed, 21 passed (this run and the ledger's).
• After: python -m pytest test/test_chunking.py test/test_tokenization.py, 22 passed, Python 3.11.
• The caller YAML parses; the two inputs it names exist on the shared workflow.

This PR was written by an AI agent (Claude) and is not reviewed by a human yet.

🤖 Generated with Claude Code

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on base/target branches other than the default branch.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Advanced

Run ID: eb027a13-bfaa-47fe-af24-abc057823c99

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[github-actions]

Beep! Automated checks have reached 100% completion. 🔋

I've aggregated the results of the automated checks for this PR below.

🔨 Build Tests

The build pipeline has finished its work. 🏁

✅ All versions pass

Python	Build	Install	Tests	pytest
3.10	✅	✅	✅	22 passed in 0.07s
3.11	✅	✅	✅	22 passed in 0.09s
3.12	✅	✅	✅	22 passed in 0.07s
3.13	✅	✅	✅	22 passed in 0.07s
3.14	✅	✅	✅	22 passed in 0.11s

---

Catch you at the next merge! 🌊