feat: add humanize library to commit0 benchmark

[07Kaustubh]

Summary

Add python-humanize/humanize (v4.15.0) as a new benchmark entry for the commit0 dataset.

Library Details

Field	Value
Original repo	python-humanize/humanize
Purpose	Human-friendly formatting of numbers, dates, file sizes, and time deltas
Python version	3.12
Tests	737 collected, 737 pass at reference commit
Functions stubbed	34 across 5 source modules
Source modules	filesize.py, i18n.py, lists.py, number.py, time.py

Commits

Commit	SHA	Description
reference_commit	2ddb5903cdc1c7e6eb6b083f4f99f73db50aecd9	Tag 4.15.0 — full working implementation
base_commit	7a405e3f49b907db112ff08d9096720a37f60447	Stubs — 34 functions replaced with pass

Dataset Entry

{
  "instance_id": "commit-0/humanize",
  "repo": "commit-0/humanize",
  "original_repo": "python-humanize/humanize",
  "base_commit": "7a405e3f49b907db112ff08d9096720a37f60447",
  "reference_commit": "2ddb5903cdc1c7e6eb6b083f4f99f73db50aecd9",
  "setup": {
    "install": "pip install -e .",
    "packages": null,
    "pip_packages": ["freezegun", "pytest", "pytest-cov"],
    "pre_install": [
      "export SETUPTOOLS_SCM_PRETEND_VERSION=4.15.0",
      "apt-get update && apt-get install -y gettext",
      "bash scripts/generate-translation-binaries.sh"
    ],
    "python": "3.12",
    "specification": "https://humanize.readthedocs.io/"
  },
  "test": {
    "test_cmd": "pytest",
    "test_dir": "tests/"
  },
  "src_dir": "src/humanize/"
}

Verification

• ✅ import humanize succeeds at base_commit (stubs are syntactically valid)
• ✅ 737 tests collected at base_commit
• ✅ Tests fail (not error) at base_commit — stubs return None, producing assertion failures
• ✅ 737/737 tests pass at reference_commit (including all i18n tests with compiled .mo files)
• ✅ Only 5 files in src/humanize/ modified between reference → base; zero test file changes

Changes in This PR

1. commit0/harness/constants.py — Add SPLIT_HUMANIZE, append "humanize" to SPLIT_ALL, add entry to SPLIT dict
2. commit0/data/test_ids/humanize.bz2 — 737 pytest node IDs (bz2-compressed)

Additional Integration Required

For full integration, maintainers will also need to:

1. Fork python-humanize/humanize to the commit-0 GitHub org (stubs branch available at 07Kaustubh/humanize@commit0-stubs)
2. Add dataset row to HuggingFace datasets (commit0/commit0 and wentingzhao/commit0_combined)
3. Build & push Docker image as wentingzhao/humanize:v0

Design Notes

• i18n dependency chain: All i18n utility functions are stubbed (consistent with babel, jinja, marshmallow precedent). Agents must implement i18n.py before number.py/time.py/filesize.py will function.
• _ngettext_noop preserved: Called at module level in number.py to build human_powers tuple — stubbing it would cause ImportError.
• pre_install rationale: gettext + generate-translation-binaries.sh compiles .po → .mo files needed by 59 i18n tests. SETUPTOOLS_SCM_PRETEND_VERSION is needed because the shallow Docker clone lacks git tags.

[Copilot]

Pull request overview

Adds the humanize repository as a new selectable Commit0 benchmark target by registering it in the harness repo splits and providing its precomputed pytest node IDs.

Changes:

• Added humanize to SPLIT_ALL and introduced SPLIT_HUMANIZE, wiring it into the SPLIT mapping.
• Added commit0/data/test_ids/humanize.bz2 containing the (bz2-compressed) pytest node IDs for the benchmark.

Reviewed changes

Copilot reviewed 1 out of 2 changed files in this pull request and generated no comments.

File	Description
commit0/harness/constants.py	Registers the new humanize repo split and includes it in the “all” set.
commit0/data/test_ids/humanize.bz2	Adds compressed pytest node IDs used by the harness to run the benchmark’s tests.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.