BatchLens Bio v0.2.0

BatchLens Bio 0.2.0 adds a local browser interface and native desktop installers to the metadata-only experimental-design audit tool.

It checks experimental units, sample/batch coverage and explicit additive-model contrasts, with independent/paired support boundaries. Outputs include an offline HTML report, versioned JSON findings, tables and hashed provenance. Seven synthetic demos ship in the package.

Desktop: download the .dmg matching your Mac (arm64 for Apple Silicon, x86_64 for Intel), drag BatchLens Bio into Applications, and open it. Windows x64 users run Setup.exe. Python and audit dependencies are bundled. Windows setup provisions WebView2 if missing; that initial setup requires internet access.

Browser: install the wheel/source archive with python -m pip install <downloaded-file> and run batchlens serve. The default browser opens a protected localhost URL. Drag in sample/design files, click Run Audit, and view the full report. Optional observations/assays, seven examples, HTML export and ZIP downloads are included.

Reports are saved under ~/BatchLens Audits. Raw uploaded files stay in memory. No cloud upload or account is used. The UI limit is 32 MiB combined; the existing scientific resource limits still apply. The original validate/audit/demo commands remain available.

Desktop alpha signing: Mac apps are ad-hoc signed, but not Developer ID signed or Apple-notarized. Windows installers are not Authenticode signed. OS unknown-publisher warnings may appear. macOS 14+ and Windows 10 2004+ / Windows 11 x64 are the targeted desktop platforms; see the attached workflow evidence for the exact systems actually tested.

SHA256SUMS covers every attached installer, package, demo and validation record. Frozen-runtime checks exercise all seven demos and downloads, native-window checks trigger an audit through JavaScript, and the Windows job installs the actual EXE before re-running validation. The attached HTML is a real synthetic demo report.

Limitations: no expression correction, power or causal inference, spatial correlation test, interaction/random-effect model, or guarantee of scientific validity. External user validation and independent scientific review remain pending. Python 3.12/3.13 are the CI targets; platform support should be assessed using the attached successful workflow evidence.

See README, methods, input schema and validation record. GitHub release does not imply PyPI publication; the latter is tracked separately.

Release validation: https://github.com/guatou904/batchlens-bio/actions/runs/34657047203

PyPI 0.2.0 is published. Official package · Verified release hashes and clean installation from official PyPI.

BatchLens Bio v0.1.0

BatchLens Bio 0.1.0 is an exploratory, metadata-only experimental-design audit tool.

Now published on PyPI: python -m pip install batchlens-bio==0.1.0. The PyPI workflow and fresh local official-index installation both passed; wheel/sdist hashes match the GitHub assets.

It checks experimental units, sample/batch coverage and explicit additive-model contrasts, with independent/paired support boundaries. Outputs include an offline HTML report, versioned JSON findings, tables and hashed provenance. Seven synthetic demos ship in the package.

Download the wheel or source archive and install with python -m pip install <downloaded-file>. Run batchlens demo --case confounded-time --out demo --fail-on none, then open demo/report.html. The attached HTML is a real synthetic demo output; SHA256SUMS covers the release files.

Limitations: no expression correction, power or causal inference, spatial correlation test, interaction/random-effect model, or guarantee of scientific validity. External user validation and independent scientific review remain pending. Python 3.12/3.13 are the CI targets; platform support should be assessed using the attached successful workflow evidence.

See README, methods, input schema and validation record. GitHub and official PyPI publication have both been verified.

Release validation: https://github.com/guatou904/batchlens-bio/actions/runs/34640763878