v0.1.0a2

What's Changed

• Expand README badges, add scverse footer, retire the changelog by @timtreis in #212
• Adopt the shared scverse-doc theme by @timtreis in #213
• Adopt decoupler 2.3 by @timtreis in #216
• Label the grouping variable in QC plots by @timtreis in #215
• Clarify the overview tutorial from alpha feedback by @timtreis in #217
• docs: explain interactive plots and uv install (alpha feedback) by @timtreis in #223
• build: pin pynndescent<0.7 (scanpy neighbors regression) by @timtreis in #226
• docs: move the screen-quality tutorial to pki by @timtreis in #224
• docs: answer the plate-vs-biology and hit-rigor questions (overview) by @timtreis in #228
• Fix small API and docs issues found on new datasets (#222) by @timtreis in #229
• Parse underscored channel names and OPS well ids (#218, #219) by @timtreis in #230
• QC robustness and True=problem flag convention (#220, #221) by @timtreis in #232
• map threading passthrough; cheaper, clearer evaluate_integration (#225, #227) by @timtreis in #231
• build(deps): bump the pre-commit group with 2 updates by @dependabot[bot] in #214

Full Changelog: v0.1.0a1...v0.1.0a2

v0.1.0a1

What's Changed

• Bump the pre-commit group with 3 updates by @dependabot[bot] in #3
• Add the IO module, a datasets module and tutorials by @Zethson in #5
• Drop the agent scaffolding by @Zethson in #24
• Add BBBC021 with its mechanism-of-action annotations by @Zethson in #25
• Read an ExportToSpreadsheet directory through read_profiles by @Zethson in #26
• Read profiles into one schema and serve all datasets from the registry by @timtreis in #28
• Add preprocessing by @timtreis in #29
• Add tools and metrics by @timtreis in #30
• Add plots and tutorials by @timtreis in #31
• Tidy the documentation metadata, references and contributing guide by @timtreis in #49
• Bump the actions-deps group across 1 directory with 5 updates by @dependabot[bot] in #32
• Add the single cells, one export directory and the images of BR00121438 by @timtreis in #59
• Clean up the merged tree: 33 fixes, the house style, and four measured hot paths by @timtreis in #51
• Pin the three feature_select operations and the silhouette direction by @timtreis in #58
• Draw the hit-calling null from the group and the controls pooled by @timtreis in #62
• Cite with roles throughout, and correct the citations that were wrong by @timtreis in #73
• Carry an exact cell count on every well-level dataset by @timtreis in #75
• Document the datasets and add the script that rebuilds each one by @timtreis in #76
• Read an infinity as missing, and handle ties in y as Chatterjee does by @timtreis in #77
• Make mode="replicability" the mAP-nonrep of Arevalo et al. by @timtreis in #78
• Offer both ECOD aggregations, stream the kernel through numba, and show it on real cells by @timtreis in #80
• Calibrate the composition test against the controls' spread, and offer hit_calling a robust covariance by @timtreis in #88
• Grade dose-response curves with a continuous hit call, and read the OASIS dose series by @timtreis in #87
• Pin all 141 JUMP-Target-2 plates and load one from each source by default by @timtreis in #79
• Lay out pl.plate panels in a grid on one shared color scale by @timtreis in #81
• Read the control scale off the controls that did not fit the transform by @timtreis in #98
• Read the dose response per feature, and find the window it happens in by @timtreis in #97
• Score a map by how many known relationships it recovers by @timtreis in #99
• Give a hand-built object a public way onto the API surface by @timtreis in #101
• Clear errors at the edges, and one read that should not happen by @timtreis in #106
• Align each batch's controls onto the pooled controls with TVN by @timtreis in #100
• Fit the constant model with the optimizer the curves use by @timtreis in #112
• Read the JUMP-Lite embeddings, and document working with them by @timtreis in #102
• Bump codecov/codecov-action from 7.0.0 to 7.1.0 in the actions-deps group by @dependabot[bot] in #108
• Bump the pre-commit group with 4 updates by @dependabot[bot] in #107
• Score phenotypic activity against the query's own plate, and correct tutorial 12 by @timtreis in #115
• Draw mAP nulls per call rather than through copairs' shared cache by @timtreis in #116
• Regroup the tutorials by the question they answer, and add a CRISPR page by @timtreis in #117
• Take a backed group's rows from one ordering, and read them as h5py wants them by @timtreis in #119
• Give feature_signature and dose_trajectory the var columns the schema requires by @timtreis in #118
• fix(tl): default consensus to median; drop false pycytominer-parity claim by @timtreis in #120
• fix(tl): calibrate cluster_composition p-values (leave-one-out dispersion + F reference) by @timtreis in #121
• docs: three-tier tutorial structure and a plot-rich ten-minute quickstart by @timtreis in #122
• docs: figures for the CRISPR genetic-screen tutorial by @timtreis in #123
• ci: scrub build-env path leaks from notebook outputs + pre-commit guard by @timtreis in #124
• docs: figures for the mechanism-of-action tutorial by @timtreis in #125
• docs: figures for the hit-calling tutorial by @timtreis in #126
• ci: match temp and ipykernel paths generally in the notebook scrubber by @timtreis in #127
• feat: learned embeddings as first-class citizens (use_rep reducers + variance_carried) by @timtreis in #128
• perf(pp): scale feature_select to large screens (chunked reductions + opt-in windowed correlation) by @timtreis in #130
• feature_select: opt-in decorrelate step for multivariate redundancy by @timtreis in #132
• docs(tl): plots-first figures for the differential and cross-lab tutorials by @timtreis in #135
• fix(tl): well-block permutation null for cell-resolution hit_calling (#68) by @timtreis in #136
• feat(ds): add scallops_arv471, a real optical pooled screen by @timtreis in #142
• feat(pl): optional plotly interactive plots, auto-gated by @timtreis in #143
• feat(tl): widen enrich to decouplers method suite and add a multi-method consensus by @timtreis in #129
• docs: interactive-plot canary (which_features_moved) by @timtreis in #146
• refactor(tests): speed up the default suite (fixture scope, constants, slow marks) by @timtreis in #144
• refactor(tests): aggressively cut the suite (fewer tests, less LOC) by @timtreis in #145
• docs: render all tutorials with interactive plots by @timtreis in #147
• feat(tl): interpretation layer - cluster, ora, gene-set/interaction resources, network enrichment by @timtreis in #149
• feat(tl): per-group multiple-testing correction in ora (padj_by) by @timtreis in #152
• feat(tl): stability-based auto-cut criterion in cluster by @timtreis in #153
• feat(tl): optional stability_window to bound the stability cut sweep by @timtreis in #154
• docs: one API page per module, with datasets under the API by @Zethson in #160
• docs: logo and pertpy/annbatch doc style in the logo colors by @Zethson in #161
• chore: cut restating comments and rationale, one sentence per docstring line by @Zethson in #162
• build(deps): bump the pre-commit group with 2 updates by @dependabot[bot] in #156
• build(deps): bump the actions-deps group with 2 updates by @dependabot[bot] in #157
• fix(tl): drop the cluster ceiling from the windowed stability sweep by @timtreis in #155
• feat(tl): ora corrects only over the sets a group hits (min_overlap) by @timtreis in #158
• feat(ds): canonical Metadata_Perturbation as the replication unit across all loaders by @timtreis in #150
• feat(ds): staged rohban variants via aggregated/feature_selected by @timtreis in #163
• docs: tutorial - interpreting a Cell Painting screen end to end by @timtreis in #159
• Update the cookiecutter template and link to the source from the API by @timtreis in #50
• refactor(ds): drop luad, agnp, miami, amish as functional duplicates by @timtreis in #166
• refactor(ds): generalize staging build+drift machinery (rohban reference) by @timtreis in #167
• feat(ds): host bbbc021 variants + rohban base on S3, loaders fetch by @timtrei...