Merge branch 'main' of https://github.com/YosefLab/scvi-hub-models in…

…to main

.dvc/.gitignore

@@ -0,0 +1,4 @@
+/config.local
+/tmp
+/cache
+/new

.dvc/config

@@ -0,0 +1,5 @@
+[core]
+    autostage = true
+    remote = gdrive_remote
+['remote "gdrive_remote"']
+    url = gdrive://1MbX5Au5QRgagdXUaV9xvgLB9lOEVbfbQ

.gitignore

@@ -28,3 +28,4 @@ __pycache__/
 /.vscode/
 
 /node_modules/
+/test/

data/.gitignore

@@ -1,2 +1,13 @@
 /mouse_thymus_cite.h5mu
-/test.txt
+/mouse_thymus_cite_totalvi
+/haniffa_covid_pbmc.h5mu
+/neurips_bone_marrow_cite.h5mu
+/heart_cell_atlas.h5mu
+/bone_marrow_cite_totalvi
+/heart_cell_atlas_scvi
+/heart_cell_atlas.h5ad
+/hlca_core.h5ad
+/hlca_reference_scanvi
+/test_data.h5ad
+/test_scvi
+/haniffa_covid_pbmc

data/bone_marrow_cite_totalvi.dvc

@@ -0,0 +1,6 @@
+outs:
+- md5: 78ef07339eacb072b5adcbee125c9acc.dir
+  size: 19834273
+  nfiles: 1
+  hash: md5
+  path: bone_marrow_cite_totalvi

data/haniffa_covid_pbmc.dvc

@@ -0,0 +1,6 @@
+outs:
+- md5: 9727ed6842ec135c97bdaf9fe9475a10.dir
+  size: 30431585
+  nfiles: 1
+  hash: md5
+  path: haniffa_covid_pbmc

data/haniffa_covid_pbmc.h5mu.dvc

@@ -0,0 +1,5 @@
+outs:
+- md5: fcb4c382ba3695eba206c4714fca3d45
+  size: 9357617754
+  hash: md5
+  path: haniffa_covid_pbmc.h5mu

data/heart_cell_atlas.h5ad.dvc

@@ -0,0 +1,5 @@
+outs:
+- md5: 767bf9739d979b6a808164528f70802b
+  size: 28528362
+  hash: md5
+  path: heart_cell_atlas.h5ad

data/heart_cell_atlas.h5mu.dvc

@@ -0,0 +1,5 @@
+outs:
+- md5: 767bf9739d979b6a808164528f70802b
+  size: 28528362
+  hash: md5
+  path: heart_cell_atlas.h5mu

data/heart_cell_atlas_scvi.dvc

@@ -0,0 +1,6 @@
+outs:
+- md5: 75670e08aee1039bc163245eb6fc2152.dir
+  size: 3371610
+  nfiles: 1
+  hash: md5
+  path: heart_cell_atlas_scvi

data/hlca_core.h5ad.dvc

@@ -0,0 +1,5 @@
+outs:
+- md5: 923ab873d03cff49cfd53a9063e6c1ed
+  size: 1940424488
+  hash: md5
+  path: hlca_core.h5ad

data/hlca_reference_scanvi.dvc

@@ -0,0 +1,6 @@
+outs:
+- md5: 5c77b351ae8e005c3180d45cf1281586.dir
+  size: 5825950
+  nfiles: 1
+  hash: md5
+  path: hlca_reference_scanvi

data/mouse_thymus_cite.h5mu.dvc

@@ -1,5 +1,5 @@
 outs:
-- md5: 0932a99a0e1571da676b2cc4881d189a
-  size: 6628168757
+- md5: 27ae88708a0e2e42591f8665601ca915
+  size: 1807038906
   hash: md5
   path: mouse_thymus_cite.h5mu

data/mouse_thymus_cite_totalvi.dvc

@@ -0,0 +1,6 @@
+outs:
+- md5: a2e5acb9b535c112698ea6cad43a6a95.dir
+  size: 19781665
+  nfiles: 1
+  hash: md5
+  path: mouse_thymus_cite_totalvi

data/neurips_bone_marrow_cite.h5mu.dvc

@@ -0,0 +1,5 @@
+outs:
+- md5: 3f68254232dfe342a45a612684f36d92
+  size: 881794010
+  hash: md5
+  path: neurips_bone_marrow_cite.h5mu

data/test_data.h5ad.dvc

@@ -0,0 +1,5 @@
+outs:
+- md5: 91bb92467e4f291e0ba0c1663dfabc96
+  size: 1008864
+  hash: md5
+  path: test_data.h5ad

data/test_scvi.dvc

@@ -0,0 +1,6 @@
+outs:
+- md5: 0155f7292347d68142c70584a7607eae.dir
+  size: 309658
+  nfiles: 1
+  hash: md5
+  path: test_scvi

src/scvi_hub_models/__main__.py

@@ -12,7 +12,15 @@
 @click.option("--dry_run", type=bool, default=False, help="Dry run the workflow.")
 @click.option("--config_key", type=str, help="Use a different config file, e.g. for test purpose.")
 @click.option("--save_dir", type=str, help="Directory to save intermediate results (defaults temporary).")
-def run_workflow(model_name: str, dry_run: bool, config_key: str = None, save_dir: str = None) -> None:
+@click.option("--reload_data", type=bool, help="Reload the data or get from DVC.")
+@click.option("--reload_model", type=bool, help="Reload the model or get from DVC.")
+def run_workflow(
+    model_name: str,
+    dry_run: bool,
+    config_key: str = None,
+    save_dir: str = None,
+    reload_data: bool = False,
+    reload_model: bool = False) -> None:
     """Run the workflow for a specific model."""
     from importlib import import_module
     if not config_key:
@@ -22,7 +30,7 @@ def run_workflow(model_name: str, dry_run: bool, config_key: str = None, save_di
     Workflow = workflow_module._Workflow
     config = json_data_store[config_key]
 
-    workflow = Workflow(save_dir=save_dir, dry_run=dry_run, config=config)
+    workflow = Workflow(save_dir=save_dir, dry_run=dry_run, config=config, reload_data=reload_data, reload_model=reload_model)
     workflow.run()
 
 

src/scvi_hub_models/config/haniffa_covid_pbmc.json

@@ -2,9 +2,11 @@
     "model_dir": "haniffa_covid_pbmc",
     "model_class": "TOTALVI",
     "repo_name": "scvi-tools/haniffa_covid_pbmc_totalvi",
+    "reload_data": true,
     "extra_data_kwargs": {
         "reference_adata_cxg_id": "c7775e88-49bf-4ba2-a03b-93f00447c958",
-        "reference_adata_fname": "haniffa_covid_pbmc.h5ad"
+        "reference_adata_fname": "haniffa_covid_pbmc.h5ad",
+        "large_training_file_name": "haniffa_covid_pbmc.h5mu"
     },
     "metadata": {
         "training_data_url": "https://datasets.cellxgene.cziscience.com/5ad66a4f-d619-4cb3-8015-a87c755647b3.h5ad",
@@ -16,6 +18,7 @@
         "description": "CITE-seq to measure RNA and surface proteins in thymocytes from wild-type and T cell lineage-restricted mice to generate a comprehensive timeline of cell state for each T cell lineage.",
         "references": "Steier, Z., Aylard, D.A., McIntyre, L.L. et al. Single-cell multiomic analysis of thymocyte development reveals drivers of CD4+ T cell and CD8+ T cell lineage commitment. Nat Immunol 24, 1579–1590 (2023). https://doi.org/10.1038/s41590-023-01584-0."
     },
+
     "criticism_settings": {
         "n_samples": 3,
         "cell_type_key": "cell_type"

src/scvi_hub_models/config/heart_cell_atlas.json

@@ -2,6 +2,10 @@
     "model_dir": "heart_cell_atlas_scvi",
     "model_class": "SCVI",
     "repo_name": "scvi-tools/heart-cell-atlas-scvi",
+    "extra_data_kwargs": {
+        "reference_adata_fname": "heart_cell_atlas.h5ad",
+        "large_training_file_name": "heart_cell_atlas.h5ad"
+    },
     "metadata": {
         "training_data_url": "https://www.heartcellatlas.org/#DataSources",
         "tissues": ["heart"],

src/scvi_hub_models/config/human_lung_cell_atlas.json

@@ -1,13 +1,14 @@
 {
-    "model_dir": "hlca_scanvi_reference",
+    "model_dir": "hlca_reference_scanvi",
     "model_class": "SCANVI",
     "repo_name": "scvi-tools/human-lung-cell-atlas-scanvi",
     "extra_data_kwargs": {
         "legacy_model_url": "https://zenodo.org/records/7599104/files/HLCA_reference_model.zip",
         "legacy_model_hash": "a7cd60f4342292b3cba54545bcd8a34decdc8e6b82163f009273d543e7e3910e",
         "legacy_model_dir": "hlca_scanvi_reference_legacy",
         "reference_adata_cxg_id": "066943a2-fdac-4b29-b348-40cede398e4e",
-        "reference_adata_fname": "hlca_core.h5ad"
+        "reference_adata_fname": "hlca_core.h5ad",
+        "large_training_file_name": "hlca_core.h5ad"
     },
     "metadata": {
         "training_data_url": "https://cellxgene.cziscience.com/collections/6f6d381a-7701-4781-935c-db10d30de293",

src/scvi_hub_models/config/mouse_thymus_cite.json

@@ -1,11 +1,10 @@
 {
-    "model_dir": "mouse_thymus_cite",
+    "model_dir": "mouse_thymus_cite_totalvi",
     "model_class": "TOTALVI",
-    "repo_name": "scvi-tools/mouse_thymus_totalvi",
-    "reload_data": true,
+    "repo_name": "scvi-tools/mouse_thymus_cite_totalvi",
     "extra_data_kwargs": {
         "reference_adata_cxg_id": "c14c54f8-85d8-45db-9de7-6ab572cc748a",
-        "reference_adata_fname": "thymus_cite.h5ad",
+        "reference_adata_fname": "mouse_thymus_cite.h5ad",
         "large_training_file_name": "mouse_thymus_cite.h5mu"
     },
     "metadata": {

src/scvi_hub_models/config/neurips_cite.json

@@ -1,9 +1,12 @@
 {
-    "model_dir": "bone_marrow_cite",
+    "model_dir": "bone_marrow_cite_totalvi",
     "model_class": "TOTALVI",
     "repo_name": "scvi-tools/bone_marrow_cite_totalvi",
     "extra_data_kwargs": {
-        "reference_adata_fname": "bmmc_cite.h5ad"
+        "url": "https://www.ncbi.nlm.nih.gov/geo/download/?acc=GSE194122&format=file&file=GSE194122%5Fopenproblems%5Fneurips2021%5Fcite%5FBMMC%5Fprocessed%2Eh5ad%2Egz",
+        "hash": "b9b50fade9349719cba23c97c6515d3501a32ee3735fe95fe51221d2e8a5f361",
+        "reference_adata_fname": "bmmc_cite.h5ad.gz",
+        "large_training_file_name": "neurips_bone_marrow_cite.h5mu"
     },
     "metadata": {
         "training_data_url": "https://www.ncbi.nlm.nih.gov/geo/download/?acc=GSE194122&format=file&file=GSE194122%5Fopenproblems%5Fneurips2021%5Fcite%5FBMMC%5Fprocessed%2Eh5ad%2Egz",

src/scvi_hub_models/config/test_scvi.json

@@ -2,9 +2,11 @@
     "model_dir": "test_scvi",
     "model_class": "SCVI",
     "repo_name": "scvi-tools/test-scvi",
+    "extra_data_kwargs": {
+        "large_training_file_name": "test_data.h5ad"
+    },
     "collection_name": "test",
     "minify_model": false,
-    "extra_data_kwargs": {},
     "metadata": {
         "tissues": ["synthetic"],
         "data_modalities": ["rna"],