crispyx 0.2.4__tar.gz → 0.2.5__tar.gz

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (92) hide show
  1. {crispyx-0.2.4/src/crispyx.egg-info → crispyx-0.2.5}/PKG-INFO +5 -5
  2. {crispyx-0.2.4 → crispyx-0.2.5}/README.md +4 -4
  3. {crispyx-0.2.4 → crispyx-0.2.5}/src/crispyx/__init__.py +18 -0
  4. {crispyx-0.2.4 → crispyx-0.2.5}/src/crispyx/_namespaces.py +90 -0
  5. {crispyx-0.2.4 → crispyx-0.2.5}/src/crispyx/_version.py +1 -1
  6. {crispyx-0.2.4 → crispyx-0.2.5}/src/crispyx/data.py +759 -220
  7. {crispyx-0.2.4 → crispyx-0.2.5}/src/crispyx/de.py +983 -95
  8. {crispyx-0.2.4 → crispyx-0.2.5}/src/crispyx/glm.py +63 -104
  9. {crispyx-0.2.4 → crispyx-0.2.5}/src/crispyx/plotting.py +2 -1
  10. {crispyx-0.2.4 → crispyx-0.2.5}/src/crispyx/pseudobulk.py +85 -9
  11. {crispyx-0.2.4 → crispyx-0.2.5}/src/crispyx/qc.py +393 -1
  12. {crispyx-0.2.4 → crispyx-0.2.5/src/crispyx.egg-info}/PKG-INFO +5 -5
  13. {crispyx-0.2.4 → crispyx-0.2.5}/src/crispyx.egg-info/SOURCES.txt +8 -0
  14. crispyx-0.2.5/tests/test_assign_guides.py +88 -0
  15. crispyx-0.2.5/tests/test_atomic_outputs.py +168 -0
  16. crispyx-0.2.5/tests/test_check_data.py +201 -0
  17. crispyx-0.2.5/tests/test_collapse_genes.py +122 -0
  18. {crispyx-0.2.4 → crispyx-0.2.5}/tests/test_concat_h5ad.py +1 -1
  19. {crispyx-0.2.4 → crispyx-0.2.5}/tests/test_convert_format.py +4 -1
  20. {crispyx-0.2.4 → crispyx-0.2.5}/tests/test_data_helpers.py +9 -0
  21. {crispyx-0.2.4 → crispyx-0.2.5}/tests/test_nb_glm.py +27 -0
  22. crispyx-0.2.5/tests/test_nb_offset.py +386 -0
  23. crispyx-0.2.5/tests/test_pseudobulk_sparse_output.py +95 -0
  24. {crispyx-0.2.4 → crispyx-0.2.5}/tests/test_rank_genes_groups_outputs.py +17 -0
  25. crispyx-0.2.5/tests/test_recover_counts.py +106 -0
  26. crispyx-0.2.5/tests/test_shrink_lfc_arrays.py +79 -0
  27. {crispyx-0.2.4 → crispyx-0.2.5}/tests/test_stream_on_fast_axis.py +6 -5
  28. {crispyx-0.2.4 → crispyx-0.2.5}/tests/test_two_group_deseq2.py +16 -0
  29. {crispyx-0.2.4 → crispyx-0.2.5}/LICENSE +0 -0
  30. {crispyx-0.2.4 → crispyx-0.2.5}/pyproject.toml +0 -0
  31. {crispyx-0.2.4 → crispyx-0.2.5}/setup.cfg +0 -0
  32. {crispyx-0.2.4 → crispyx-0.2.5}/src/crispyx/_checkpoint.py +0 -0
  33. {crispyx-0.2.4 → crispyx-0.2.5}/src/crispyx/_disk.py +0 -0
  34. {crispyx-0.2.4 → crispyx-0.2.5}/src/crispyx/_grouping.py +0 -0
  35. {crispyx-0.2.4 → crispyx-0.2.5}/src/crispyx/_h5codec.py +0 -0
  36. {crispyx-0.2.4 → crispyx-0.2.5}/src/crispyx/_irls.py +0 -0
  37. {crispyx-0.2.4 → crispyx-0.2.5}/src/crispyx/_kernels.py +0 -0
  38. {crispyx-0.2.4 → crispyx-0.2.5}/src/crispyx/_memory.py +0 -0
  39. {crispyx-0.2.4 → crispyx-0.2.5}/src/crispyx/_messages.py +0 -0
  40. {crispyx-0.2.4 → crispyx-0.2.5}/src/crispyx/_preflight.py +0 -0
  41. {crispyx-0.2.4 → crispyx-0.2.5}/src/crispyx/_provenance.py +0 -0
  42. {crispyx-0.2.4 → crispyx-0.2.5}/src/crispyx/_size_factors.py +0 -0
  43. {crispyx-0.2.4 → crispyx-0.2.5}/src/crispyx/_statistics.py +0 -0
  44. {crispyx-0.2.4 → crispyx-0.2.5}/src/crispyx/batch.py +0 -0
  45. {crispyx-0.2.4 → crispyx-0.2.5}/src/crispyx/dimred.py +0 -0
  46. {crispyx-0.2.4 → crispyx-0.2.5}/src/crispyx/hvg.py +0 -0
  47. {crispyx-0.2.4 → crispyx-0.2.5}/src/crispyx/profiling.py +0 -0
  48. {crispyx-0.2.4 → crispyx-0.2.5}/src/crispyx/residual.py +0 -0
  49. {crispyx-0.2.4 → crispyx-0.2.5}/src/crispyx/sample.py +0 -0
  50. {crispyx-0.2.4 → crispyx-0.2.5}/src/crispyx.egg-info/dependency_links.txt +0 -0
  51. {crispyx-0.2.4 → crispyx-0.2.5}/src/crispyx.egg-info/requires.txt +0 -0
  52. {crispyx-0.2.4 → crispyx-0.2.5}/src/crispyx.egg-info/top_level.txt +0 -0
  53. {crispyx-0.2.4 → crispyx-0.2.5}/tests/test_batch_process.py +0 -0
  54. {crispyx-0.2.4 → crispyx-0.2.5}/tests/test_cgroup_memory_budget.py +0 -0
  55. {crispyx-0.2.4 → crispyx-0.2.5}/tests/test_compare_controls.py +0 -0
  56. {crispyx-0.2.4 → crispyx-0.2.5}/tests/test_compress_h5ad.py +0 -0
  57. {crispyx-0.2.4 → crispyx-0.2.5}/tests/test_de_edge_cases.py +0 -0
  58. {crispyx-0.2.4 → crispyx-0.2.5}/tests/test_de_lowexpr_filter.py +0 -0
  59. {crispyx-0.2.4 → crispyx-0.2.5}/tests/test_de_resume.py +0 -0
  60. {crispyx-0.2.4 → crispyx-0.2.5}/tests/test_dimred.py +0 -0
  61. {crispyx-0.2.4 → crispyx-0.2.5}/tests/test_disk_dispatch.py +0 -0
  62. {crispyx-0.2.4 → crispyx-0.2.5}/tests/test_downsample_counts.py +0 -0
  63. {crispyx-0.2.4 → crispyx-0.2.5}/tests/test_glm_structured.py +0 -0
  64. {crispyx-0.2.4 → crispyx-0.2.5}/tests/test_groupby_alias.py +0 -0
  65. {crispyx-0.2.4 → crispyx-0.2.5}/tests/test_hvg.py +0 -0
  66. {crispyx-0.2.4 → crispyx-0.2.5}/tests/test_in_memory_input.py +0 -0
  67. {crispyx-0.2.4 → crispyx-0.2.5}/tests/test_irls.py +0 -0
  68. {crispyx-0.2.4 → crispyx-0.2.5}/tests/test_memory_dispatch.py +0 -0
  69. {crispyx-0.2.4 → crispyx-0.2.5}/tests/test_messages.py +0 -0
  70. {crispyx-0.2.4 → crispyx-0.2.5}/tests/test_normalisation.py +0 -0
  71. {crispyx-0.2.4 → crispyx-0.2.5}/tests/test_normalized_effects.py +0 -0
  72. {crispyx-0.2.4 → crispyx-0.2.5}/tests/test_normalized_effects_se.py +0 -0
  73. {crispyx-0.2.4 → crispyx-0.2.5}/tests/test_obs_writes.py +0 -0
  74. {crispyx-0.2.4 → crispyx-0.2.5}/tests/test_pickle.py +0 -0
  75. {crispyx-0.2.4 → crispyx-0.2.5}/tests/test_plotting.py +0 -0
  76. {crispyx-0.2.4 → crispyx-0.2.5}/tests/test_preflight_dispatch.py +0 -0
  77. {crispyx-0.2.4 → crispyx-0.2.5}/tests/test_profiling.py +0 -0
  78. {crispyx-0.2.4 → crispyx-0.2.5}/tests/test_pseudobulk_batch.py +0 -0
  79. {crispyx-0.2.4 → crispyx-0.2.5}/tests/test_pseudobulk_memory.py +0 -0
  80. {crispyx-0.2.4 → crispyx-0.2.5}/tests/test_pseudobulk_profiles.py +0 -0
  81. {crispyx-0.2.4 → crispyx-0.2.5}/tests/test_qc_parity.py +0 -0
  82. {crispyx-0.2.4 → crispyx-0.2.5}/tests/test_read_backed.py +0 -0
  83. {crispyx-0.2.4 → crispyx-0.2.5}/tests/test_residual_effects.py +0 -0
  84. {crispyx-0.2.4 → crispyx-0.2.5}/tests/test_result_provenance.py +0 -0
  85. {crispyx-0.2.4 → crispyx-0.2.5}/tests/test_scanpy_parity.py +0 -0
  86. {crispyx-0.2.4 → crispyx-0.2.5}/tests/test_standardize.py +0 -0
  87. {crispyx-0.2.4 → crispyx-0.2.5}/tests/test_subsample.py +0 -0
  88. {crispyx-0.2.4 → crispyx-0.2.5}/tests/test_temp_file_cleanup.py +0 -0
  89. {crispyx-0.2.4 → crispyx-0.2.5}/tests/test_wilcoxon_dispatch.py +0 -0
  90. {crispyx-0.2.4 → crispyx-0.2.5}/tests/test_wilcoxon_labels.py +0 -0
  91. {crispyx-0.2.4 → crispyx-0.2.5}/tests/test_wilcoxon_stratified.py +0 -0
  92. {crispyx-0.2.4 → crispyx-0.2.5}/tests/test_workflow.py +0 -0
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.4
2
2
  Name: crispyx
3
- Version: 0.2.4
3
+ Version: 0.2.5
4
4
  Summary: Memory-efficient streaming analysis of large-scale CRISPR and Perturb-seq screens on disk-backed AnnData files
5
5
  Author: Jin-Hong Du
6
6
  License: Modified MIT License
@@ -145,7 +145,7 @@ Dynamic: license-file
145
145
 
146
146
  Genome-wide CRISPR screens routinely produce datasets with hundreds of thousands of cells and tens of thousands of genes. Standard single-cell analysis toolkits (Scanpy, Pertpy) load the entire count matrix into memory, requiring large RAM allocations and often making routine workflows impractical on laptops or shared compute nodes.
147
147
 
148
- **crispyx** solves this by streaming data directly from on-disk AnnData (`.h5ad`) files. Quality control, normalisation, pseudo-bulk aggregation, and differential expression all operate without materialising the full matrix.
148
+ **crispyx** runs quality control, normalisation, pseudo-bulk aggregation, differential expression (t-test, Wilcoxon, NB-GLM with covariates, covariate-adjusted residual effects), control checks and PCA/KNN on screens stored as `.h5ad` files, streaming from disk under a memory budget, so that million-cell screens run on a laptop or a cluster node. The API mirrors Scanpy. It starts from an `.h5ad` file; it does not read raw sequencing outputs, and it is not a general single-cell toolkit.
149
149
 
150
150
  **Documentation:** [crispyx.readthedocs.io](https://crispyx.readthedocs.io/en/latest/) — [installation](https://crispyx.readthedocs.io/en/latest/installation.html), [quick start](https://crispyx.readthedocs.io/en/latest/quickstart.html), [tutorials](https://crispyx.readthedocs.io/en/latest/notebooks/crispyx_tutorial.html) and [API reference](https://crispyx.readthedocs.io/en/latest/api.html).
151
151
 
@@ -202,18 +202,18 @@ For the full workflow (normalisation, PCA, pseudo-bulk, NB-GLM, LFC shrinkage, p
202
202
  - **Streaming QC & preprocessing** – Filter cells, perturbations, and genes; normalise and log-transform; select highly variable genes (`seurat_v3`/`mean_dispersion`, control-cells-only by default); CSC-aware streaming with `format_mismatch_policy`; all without loading the full matrix into memory
203
203
  - **Subsampling & downsampling** – Stratified or cluster-sampled cell subsampling (`cx.pp.subsample`, exact count or proportion per stratum, drop or keep small groups) and dependency-free per-cell count thinning (`cx.pp.downsample_counts`, the streaming equivalent of `scanpy.pp.downsample_counts`) for aligning dataset scale and sequencing depth before comparing screens
204
204
  - **Pseudo-bulk aggregation** – Absolute profiles over multiple grouping columns (for example, perturbation × batch), strict count sums or mean log1p expression, optional deterministic bootstrap sampling, and explicit within-batch effect calculation
205
- - **Differential expression** – t-test, Wilcoxon rank-sum (including batch-stratified / van Elteren test via `batch_column`), and negative binomial GLM with covariates and apeGLM LFC shrinkage; multi-core support and adaptive memory management; per-condition low-expression filtering to exclude genes that are near-zero in both groups
205
+ - **Differential expression** – t-test, Wilcoxon rank-sum (including batch-stratified / van Elteren test via `batch_column`), negative binomial GLM with covariates and apeGLM LFC shrinkage, and a fixed-offset NB test for many small perturbations against a large, stratified control set (`cx.tl.nb_offset_test`); multi-core support and adaptive memory management; per-condition low-expression filtering to exclude genes that are near-zero in both groups
206
206
  - **Covariate-adjusted effects** – `cx.tl.residual_effects` (Feng et al., 2026) fits each gene to covariates such as donor, lane or batch on the control cells only, and takes a perturbation's effect as the mean residual of its cells. Unlike the NB GLM, which fits the covariates on counts jointly with each perturbation, the control fit is shared by every perturbation, so a perturbation's own cells cannot shift its covariate adjustment, and all perturbations come from one streaming pass. On a 1.16-million-cell screen it reproduces the published effects and significance calls in under a minute.
207
207
  - **Storage-format aware** – Gene-streaming functions (`wilcoxon_test`, `batch_process`) can convert a CSR source once to a temporary, memory-bounded CSC copy beside the output instead of re-reading the whole matrix once per gene chunk; `format_mismatch_policy="auto"` (the default) measures the source's read throughput and does so only when the re-reads would really cost more than the conversion, so a small local file is streamed as-is and a large one on a network filesystem is converted; `convert_to_csc`/`convert_to_csr` honour `memory_limit_gb` and preserve the value dtype
208
208
  - **Dimension reduction** – Memory-efficient PCA and KNN graph construction on backed data
209
209
  - **Scanpy-compatible API & plotting** – Familiar `cx.pp`, `cx.pb`, `cx.tl`, and `cx.pl` namespaces; Scanpy-style rank genes plots, volcano, MA, PCA, UMAP, QC summaries, and overlap heatmaps
210
- - **Data preparation utilities** – Edit backed metadata without loading X; standardise gene names; normalise perturbation labels; auto-detect metadata columns
210
+ - **Data preparation utilities** – Edit backed metadata without loading X; normalise perturbation labels and take target genes out of guide names (`target_pattern`); assign guides from guide-capture counts (`cx.pp.assign_guides`); recover integer counts from data published as log1p-normalised values (`cx.pp.recover_counts`); sum the rows of a gene split over several (`cx.pp.collapse_genes`); concatenate multi-file screens; check a file for the mistakes that mislead an analysis (`cx.pp.check_data`); auto-detect metadata columns
211
211
  - **HPC-ready** – Resume/checkpoint for long-running jobs; configurable `memory_limit_gb`
212
212
  - **Disk-aware** – Estimates and warns about scratch-disk usage before large writes or CSC/CSR conversions, and `cx.estimate_disk_usage(...)` answers "how much disk will this need?" up front; the memory savings above assume the machine has enough free disk for streaming intermediates and output files
213
213
 
214
214
  ## Performance
215
215
 
216
- crispyx consistently outperforms Scanpy, Pertpy/PyDESeq2, and edgeR in both speed and memory across a range of CRISPR screen dataset sizes, with results matching Scanpy to Pearson *r* > 0.999:
216
+ Across 12 public CRISPR screens, crispyx completes every dataset where Scanpy, Pertpy/PyDESeq2 and edgeR run out of memory or time, with t-test and Wilcoxon results matching Scanpy to Pearson *r* > 0.999 and NB-GLM statistics agreeing with PyDESeq2 given the same size factors (z correlation 0.997–0.9999):
217
217
 
218
218
  <p align="center">
219
219
  <img src="docs/_static/fig2.png" width="800" alt="Benchmark results across 12 CRISPR screens: (a) dataset sizes, (b) completion status by method, (c) concordance with Scanpy, (d) runtime scaling, (e) peak memory scaling">
@@ -11,7 +11,7 @@
11
11
 
12
12
  Genome-wide CRISPR screens routinely produce datasets with hundreds of thousands of cells and tens of thousands of genes. Standard single-cell analysis toolkits (Scanpy, Pertpy) load the entire count matrix into memory, requiring large RAM allocations and often making routine workflows impractical on laptops or shared compute nodes.
13
13
 
14
- **crispyx** solves this by streaming data directly from on-disk AnnData (`.h5ad`) files. Quality control, normalisation, pseudo-bulk aggregation, and differential expression all operate without materialising the full matrix.
14
+ **crispyx** runs quality control, normalisation, pseudo-bulk aggregation, differential expression (t-test, Wilcoxon, NB-GLM with covariates, covariate-adjusted residual effects), control checks and PCA/KNN on screens stored as `.h5ad` files, streaming from disk under a memory budget, so that million-cell screens run on a laptop or a cluster node. The API mirrors Scanpy. It starts from an `.h5ad` file; it does not read raw sequencing outputs, and it is not a general single-cell toolkit.
15
15
 
16
16
  **Documentation:** [crispyx.readthedocs.io](https://crispyx.readthedocs.io/en/latest/) — [installation](https://crispyx.readthedocs.io/en/latest/installation.html), [quick start](https://crispyx.readthedocs.io/en/latest/quickstart.html), [tutorials](https://crispyx.readthedocs.io/en/latest/notebooks/crispyx_tutorial.html) and [API reference](https://crispyx.readthedocs.io/en/latest/api.html).
17
17
 
@@ -68,18 +68,18 @@ For the full workflow (normalisation, PCA, pseudo-bulk, NB-GLM, LFC shrinkage, p
68
68
  - **Streaming QC & preprocessing** – Filter cells, perturbations, and genes; normalise and log-transform; select highly variable genes (`seurat_v3`/`mean_dispersion`, control-cells-only by default); CSC-aware streaming with `format_mismatch_policy`; all without loading the full matrix into memory
69
69
  - **Subsampling & downsampling** – Stratified or cluster-sampled cell subsampling (`cx.pp.subsample`, exact count or proportion per stratum, drop or keep small groups) and dependency-free per-cell count thinning (`cx.pp.downsample_counts`, the streaming equivalent of `scanpy.pp.downsample_counts`) for aligning dataset scale and sequencing depth before comparing screens
70
70
  - **Pseudo-bulk aggregation** – Absolute profiles over multiple grouping columns (for example, perturbation × batch), strict count sums or mean log1p expression, optional deterministic bootstrap sampling, and explicit within-batch effect calculation
71
- - **Differential expression** – t-test, Wilcoxon rank-sum (including batch-stratified / van Elteren test via `batch_column`), and negative binomial GLM with covariates and apeGLM LFC shrinkage; multi-core support and adaptive memory management; per-condition low-expression filtering to exclude genes that are near-zero in both groups
71
+ - **Differential expression** – t-test, Wilcoxon rank-sum (including batch-stratified / van Elteren test via `batch_column`), negative binomial GLM with covariates and apeGLM LFC shrinkage, and a fixed-offset NB test for many small perturbations against a large, stratified control set (`cx.tl.nb_offset_test`); multi-core support and adaptive memory management; per-condition low-expression filtering to exclude genes that are near-zero in both groups
72
72
  - **Covariate-adjusted effects** – `cx.tl.residual_effects` (Feng et al., 2026) fits each gene to covariates such as donor, lane or batch on the control cells only, and takes a perturbation's effect as the mean residual of its cells. Unlike the NB GLM, which fits the covariates on counts jointly with each perturbation, the control fit is shared by every perturbation, so a perturbation's own cells cannot shift its covariate adjustment, and all perturbations come from one streaming pass. On a 1.16-million-cell screen it reproduces the published effects and significance calls in under a minute.
73
73
  - **Storage-format aware** – Gene-streaming functions (`wilcoxon_test`, `batch_process`) can convert a CSR source once to a temporary, memory-bounded CSC copy beside the output instead of re-reading the whole matrix once per gene chunk; `format_mismatch_policy="auto"` (the default) measures the source's read throughput and does so only when the re-reads would really cost more than the conversion, so a small local file is streamed as-is and a large one on a network filesystem is converted; `convert_to_csc`/`convert_to_csr` honour `memory_limit_gb` and preserve the value dtype
74
74
  - **Dimension reduction** – Memory-efficient PCA and KNN graph construction on backed data
75
75
  - **Scanpy-compatible API & plotting** – Familiar `cx.pp`, `cx.pb`, `cx.tl`, and `cx.pl` namespaces; Scanpy-style rank genes plots, volcano, MA, PCA, UMAP, QC summaries, and overlap heatmaps
76
- - **Data preparation utilities** – Edit backed metadata without loading X; standardise gene names; normalise perturbation labels; auto-detect metadata columns
76
+ - **Data preparation utilities** – Edit backed metadata without loading X; normalise perturbation labels and take target genes out of guide names (`target_pattern`); assign guides from guide-capture counts (`cx.pp.assign_guides`); recover integer counts from data published as log1p-normalised values (`cx.pp.recover_counts`); sum the rows of a gene split over several (`cx.pp.collapse_genes`); concatenate multi-file screens; check a file for the mistakes that mislead an analysis (`cx.pp.check_data`); auto-detect metadata columns
77
77
  - **HPC-ready** – Resume/checkpoint for long-running jobs; configurable `memory_limit_gb`
78
78
  - **Disk-aware** – Estimates and warns about scratch-disk usage before large writes or CSC/CSR conversions, and `cx.estimate_disk_usage(...)` answers "how much disk will this need?" up front; the memory savings above assume the machine has enough free disk for streaming intermediates and output files
79
79
 
80
80
  ## Performance
81
81
 
82
- crispyx consistently outperforms Scanpy, Pertpy/PyDESeq2, and edgeR in both speed and memory across a range of CRISPR screen dataset sizes, with results matching Scanpy to Pearson *r* > 0.999:
82
+ Across 12 public CRISPR screens, crispyx completes every dataset where Scanpy, Pertpy/PyDESeq2 and edgeR run out of memory or time, with t-test and Wilcoxon results matching Scanpy to Pearson *r* > 0.999 and NB-GLM statistics agreeing with PyDESeq2 given the same size factors (z correlation 0.997–0.9999):
83
83
 
84
84
  <p align="center">
85
85
  <img src="docs/_static/fig2.png" width="800" alt="Benchmark results across 12 CRISPR screens: (a) dataset sizes, (b) completion status by method, (c) concordance with Scanpy, (d) runtime scaling, (e) peak memory scaling">
@@ -41,6 +41,8 @@ from .data import (
41
41
  normalise_perturbation_labels,
42
42
  read_h5ad_ondisk,
43
43
  read_backed,
44
+ recover_counts,
45
+ collapse_genes,
44
46
  resolve_data_path,
45
47
  standardise_gene_names,
46
48
  write_filtered_subset,
@@ -50,7 +52,10 @@ from .data import (
50
52
  from .de import (
51
53
  RankGenesGroupsResult,
52
54
  nb_glm_test,
55
+ nb_offset_test,
53
56
  shrink_lfc,
57
+ shrink_lfc_arrays,
58
+ ShrunkLFC,
54
59
  t_test,
55
60
  wilcoxon_test,
56
61
  )
@@ -84,12 +89,16 @@ from .pseudobulk import (
84
89
  compute_pseudobulk_effects,
85
90
  )
86
91
  from .qc import (
92
+ Issue,
93
+ assign_guides,
94
+ check_perturbation_data,
87
95
  compare_controls,
88
96
  control_guide_correlation,
89
97
  filter_cells_by_gene_count,
90
98
  filter_genes_by_cell_count,
91
99
  filter_perturbations_by_cell_count,
92
100
  quality_control_summary,
101
+ raise_on_errors,
93
102
  )
94
103
  from .residual import residual_effects
95
104
  from .sample import subsample
@@ -128,6 +137,10 @@ __all__ = [
128
137
  "quality_control_summary",
129
138
  "compare_controls",
130
139
  "control_guide_correlation",
140
+ "assign_guides",
141
+ "check_perturbation_data",
142
+ "Issue",
143
+ "raise_on_errors",
131
144
  "plot_control_guide_correlation",
132
145
  "highly_variable_genes",
133
146
  # Subsampling
@@ -147,7 +160,10 @@ __all__ = [
147
160
  "t_test",
148
161
  "wilcoxon_test",
149
162
  "nb_glm_test",
163
+ "nb_offset_test",
150
164
  "shrink_lfc",
165
+ "shrink_lfc_arrays",
166
+ "ShrunkLFC",
151
167
  # Data utilities
152
168
  "AnnData",
153
169
  "ensure_gene_symbol_column",
@@ -156,6 +172,8 @@ __all__ = [
156
172
  "resolve_data_path",
157
173
  "normalize_total_log1p",
158
174
  "downsample_counts",
175
+ "recover_counts",
176
+ "collapse_genes",
159
177
  "convert_to_csc",
160
178
  "convert_to_csr",
161
179
  "compress_h5ad",
@@ -19,6 +19,8 @@ from .data import (
19
19
  convert_to_csr,
20
20
  concat_h5ad,
21
21
  downsample_counts,
22
+ recover_counts,
23
+ collapse_genes,
22
24
  normalize_total_log1p,
23
25
  read_backed,
24
26
  resolve_data_path,
@@ -26,7 +28,9 @@ from .data import (
26
28
  from .de import (
27
29
  RankGenesGroupsResult,
28
30
  nb_glm_test,
31
+ nb_offset_test,
29
32
  shrink_lfc,
33
+ shrink_lfc_arrays,
30
34
  t_test,
31
35
  wilcoxon_test,
32
36
  )
@@ -56,6 +60,8 @@ from .pseudobulk import (
56
60
  )
57
61
  from .residual import ResidualNull, residual_effects
58
62
  from .qc import (
63
+ assign_guides,
64
+ check_perturbation_data,
59
65
  compare_controls,
60
66
  control_guide_correlation,
61
67
  filter_cells_by_gene_count,
@@ -210,6 +216,21 @@ class _PreprocessingNamespace:
210
216
  """
211
217
  return control_guide_correlation(data, **kwargs)
212
218
 
219
+ def assign_guides(self, guide_counts, guide_names=None, **kwargs):
220
+ """The guide that dominates each cell's guide-capture counts, or NA.
221
+
222
+ See :func:`crispyx.qc.assign_guides`.
223
+ """
224
+ return assign_guides(guide_counts, guide_names, **kwargs)
225
+
226
+ def check_data(self, data, **kwargs):
227
+ """Check a screen's file for labels, genes and counts that would
228
+ mislead an analysis.
229
+
230
+ See :func:`crispyx.qc.check_perturbation_data`.
231
+ """
232
+ return check_perturbation_data(data, **kwargs)
233
+
213
234
  def convert_to_csc(
214
235
  self,
215
236
  data: str | Path | ad.AnnData,
@@ -392,6 +413,60 @@ class _PreprocessingNamespace:
392
413
  verbose=verbose,
393
414
  )
394
415
 
416
+ def recover_counts(
417
+ self,
418
+ data: str | Path | ad.AnnData,
419
+ output_path: str | Path | None = None,
420
+ *,
421
+ n_counts_column: str | None = None,
422
+ chunk_size: int = 4096,
423
+ data_name: str | None = None,
424
+ format_mismatch_policy: Literal["auto", "warn", "convert", "off"] = "auto",
425
+ verbose: int | bool = True,
426
+ ) -> AnnData:
427
+ """Recover the integer counts behind a ``log1p(normalize_total(counts))``
428
+ matrix, streamed.
429
+
430
+ See :func:`crispyx.data.recover_counts` for the full parameter
431
+ documentation.
432
+ """
433
+ return recover_counts(
434
+ data,
435
+ output_path,
436
+ n_counts_column=n_counts_column,
437
+ chunk_size=chunk_size,
438
+ data_name=data_name,
439
+ format_mismatch_policy=format_mismatch_policy,
440
+ verbose=verbose,
441
+ )
442
+
443
+ def collapse_genes(
444
+ self,
445
+ data: str | Path | ad.AnnData,
446
+ output_path: str | Path | None = None,
447
+ *,
448
+ by: "str | pd.Series | Sequence",
449
+ chunk_size: int = 4096,
450
+ data_name: str | None = None,
451
+ format_mismatch_policy: Literal["auto", "warn", "convert", "off"] = "auto",
452
+ verbose: int | bool = True,
453
+ ) -> AnnData:
454
+ """Sum the rows of ``var`` that are one gene, under a caller-supplied
455
+ key, streamed.
456
+
457
+ See :func:`crispyx.data.collapse_genes` for the full parameter
458
+ documentation.
459
+ """
460
+ return collapse_genes(
461
+ data,
462
+ output_path,
463
+ by=by,
464
+ chunk_size=chunk_size,
465
+ data_name=data_name,
466
+ format_mismatch_policy=format_mismatch_policy,
467
+ verbose=verbose,
468
+ )
469
+
395
470
  def concat_h5ad(
396
471
  self,
397
472
  paths: Sequence[str | Path],
@@ -990,6 +1065,21 @@ class _ToolsNamespace:
990
1065
  "Choose from 'wilcoxon', 't-test', or 'nb_glm'."
991
1066
  )
992
1067
 
1068
+ def nb_offset_test(self, data: str | Path | ad.AnnData, **kwargs):
1069
+ """Negative binomial test of each perturbation on its own cells, with
1070
+ the controls as a fixed per-stratum baseline.
1071
+
1072
+ See :func:`crispyx.de.nb_offset_test`.
1073
+ """
1074
+ return nb_offset_test(data, **kwargs)
1075
+
1076
+ def shrink_lfc_arrays(self, lfc, se, *, prior_scale=None, exclude=None, floor=None):
1077
+ """apeGLM shrinkage of fold changes given as arrays, exactly.
1078
+
1079
+ See :func:`crispyx.de.shrink_lfc_arrays`.
1080
+ """
1081
+ return shrink_lfc_arrays(lfc, se, prior_scale=prior_scale, exclude=exclude, floor=floor)
1082
+
993
1083
  def shrink_lfc(
994
1084
  self,
995
1085
  data: str | Path | ad.AnnData,
@@ -2,4 +2,4 @@
2
2
  the distribution's metadata reads it at build time, so a source checkout
3
3
  reports the code it runs, whatever version is installed."""
4
4
 
5
- __version__ = "0.2.4"
5
+ __version__ = "0.2.5"