v1.4.1: documentation and reader guidance¶
Release date: 22 October 2026. Git tag: v1.4.1.
Changes¶
The README distinguishes target-depth stability from fluid-density constraints over the whole hole. It introduces the project and its use without requiring development-history knowledge, and its links work from PyPI.
The scientific guides identify immediate breakout formation and the 400 °C spallation restriction as assumptions, distinguish illustrative circulation and creep cases from site programmes, and correct the explanation of supercritical water. The summary and two-page PDF include the assumed-delay opening brackets and the remaining programme failures. The PDF has a neutral filename and a visible preparation date.
The API reference explains accepted ranges and provenance labels, local-helper limitations and compatibility fields. The new-site guide and output guide explain inputs and a real JSON export, including pressure units, density bounds and schedule identity. Data notes correct stale descriptions and qualify equipment claims. Historical reviews remain available through the archive index.
The link checker covers the current data guides, scripts guide and changelog, and resolves absolute main-branch GitHub links against the checkout. Three added regressions check that coverage. The earlier version-specific documentation checker is labelled as historical.
Calculation and data preservation¶
All 29 public exports and their signatures are unchanged. Executable statements match the reviewed baseline across all 37 package and example Python files after excluding docstrings. The calculation modules contain documentation edits only. Package fingerprints include those bytes, so cached study fingerprints can change; stored evidence and the fingerprint scheme are preserved.
All 15 source numeric data files and all 71 numerical benchmark files retain their recorded bytes. The original result, golden and historical-review records remain unchanged. The v1.4.0 release record has a dated publication follow-up.
Benchmark dataset 1.0 remains an unpublished dataset draft with its own version. Its guide copies and provenance have been refreshed without running models or changing expected values and tolerances. It has no separate dataset DOI yet.
Verification¶
The final local fast suite passed 404 tests in 22.38 seconds, with 106 slow tests excluded. API, link, benchmark, distribution and fresh-install checks passed. The final release-metadata checks confirm version 1.4.1 in the package and citation. Current links are checked across 46 Markdown files. The test workflow checks the fast suite on Python 3.11 to 3.14 and installation from a wheel.
The complete v1.4.0 suite passed 507 tests, including 106 slow tests, on candidate abdb56caec44e8d3a43905e869e87983f5fda38d in CI run 39660954178. This is inherited physical-regression evidence. No new physical experiment or complete slow-suite run was required for this documentation patch; executable-code and numerical-input equivalence were checked separately.
Installation and citation¶
python -m pip install borethm==1.4.1
The GitHub release and PyPI page identify this version. The all-versions software DOI remains the project citation. The benchmark dataset has a separate citation and deposit process.