Fix imports for model initialization and allow subsetting of clustering data types

.gitignore

@@ -0,0 +1,174 @@
+# Byte-compiled / optimized / DLL files
+__pycache__/
+*.py[cod]
+*$py.class
+
+# C extensions
+*.so
+
+# Distribution / packaging
+.Python
+build/
+develop-eggs/
+dist/
+downloads/
+eggs/
+.eggs/
+lib/
+lib64/
+parts/
+sdist/
+var/
+wheels/
+share/python-wheels/
+*.egg-info/
+.installed.cfg
+*.egg
+MANIFEST
+
+# PyInstaller
+#  Usually these files are written by a python script from a template
+#  before PyInstaller builds the exe, so as to inject date/other infos into it.
+*.manifest
+*.spec
+
+# Installer logs
+pip-log.txt
+pip-delete-this-directory.txt
+
+# Unit test / coverage reports
+htmlcov/
+.tox/
+.nox/
+.coverage
+.coverage.*
+.cache
+nosetests.xml
+coverage.xml
+*.cover
+*.py,cover
+.hypothesis/
+.pytest_cache/
+cover/
+
+# Translations
+*.mo
+*.pot
+
+# Django stuff:
+*.log
+local_settings.py
+db.sqlite3
+db.sqlite3-journal
+
+# Flask stuff:
+instance/
+.webassets-cache
+
+# Scrapy stuff:
+.scrapy
+
+# Sphinx documentation
+docs/_build/
+
+# PyBuilder
+.pybuilder/
+target/
+
+# Jupyter Notebook
+.ipynb_checkpoints
+
+# IPython
+profile_default/
+ipython_config.py
+
+# pyenv
+#   For a library or package, you might want to ignore these files since the code is
+#   intended to run in multiple environments; otherwise, check them in:
+# .python-version
+
+# pipenv
+#   According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control.
+#   However, in case of collaboration, if having platform-specific dependencies or dependencies
+#   having no cross-platform support, pipenv may install dependencies that don't work, or not
+#   install all needed dependencies.
+#Pipfile.lock
+
+# UV
+#   Similar to Pipfile.lock, it is generally recommended to include uv.lock in version control.
+#   This is especially recommended for binary packages to ensure reproducibility, and is more
+#   commonly ignored for libraries.
+#uv.lock
+
+# poetry
+#   Similar to Pipfile.lock, it is generally recommended to include poetry.lock in version control.
+#   This is especially recommended for binary packages to ensure reproducibility, and is more
+#   commonly ignored for libraries.
+#   https://python-poetry.org/docs/basic-usage/#commit-your-poetrylock-file-to-version-control
+#poetry.lock
+
+# pdm
+#   Similar to Pipfile.lock, it is generally recommended to include pdm.lock in version control.
+#pdm.lock
+#   pdm stores project-wide configurations in .pdm.toml, but it is recommended to not include it
+#   in version control.
+#   https://pdm.fming.dev/latest/usage/project/#working-with-version-control
+.pdm.toml
+.pdm-python
+.pdm-build/
+
+# PEP 582; used by e.g. github.com/David-OConnor/pyflow and github.com/pdm-project/pdm
+__pypackages__/
+
+# Celery stuff
+celerybeat-schedule
+celerybeat.pid
+
+# SageMath parsed files
+*.sage.py
+
+# Environments
+.env
+.venv
+env/
+venv/
+ENV/
+env.bak/
+venv.bak/
+
+# Spyder project settings
+.spyderproject
+.spyproject
+
+# Rope project settings
+.ropeproject
+
+# mkdocs documentation
+/site
+
+# mypy
+.mypy_cache/
+.dmypy.json
+dmypy.json
+
+# Pyre type checker
+.pyre/
+
+# pytype static type analyzer
+.pytype/
+
+# Cython debug symbols
+cython_debug/
+
+# PyCharm
+#  JetBrains specific template is maintained in a separate JetBrains.gitignore that can
+#  be found at https://github.com/github/gitignore/blob/main/Global/JetBrains.gitignore
+#  and can be added to the global gitignore or merged into this file.  For a more nuclear
+#  option (not recommended) you can uncomment the following to ignore the entire idea folder.
+#.idea/
+
+# Ruff stuff:
+.ruff_cache/
+
+# PyPI configuration file
+.pypirc

README.md

@@ -94,6 +94,11 @@ from sccello.src.model_prototype_contrastive import PrototypeContrastiveForMaske
 model = PrototypeContrastiveForMaskedLM.from_pretrained("katarinayuan/scCello-zeroshot", output_hidden_states=True)
 ```
 
+or run to ensure the model can load properly
+```
+python ./sccello/script/run_load_model.py 
+```
+
 * for linear probing tasks (see details in sccello/script/run_cell_type_classification.py)
 ```
 from sccello.src.model_prototype_contrastive import PrototypeContrastiveForSequenceClassification

sccello/script/run_cell_type_clustering.py

@@ -36,6 +36,8 @@ def parse_args():
     parser.add_argument("--model_source", type=str, default="model_prototype_contrastive")
 
     parser.add_argument("--indist", type=int, default=0)
+    parser.add_argument('--sample_outdist', action='store_true', help='Samples the out of distribution')
+    parser.add_argument('--num_samples', type=int, default=10, help='If sample_outdist is set, this sets the number of samples')
     parser.add_argument("--normalize", type=int, default=0)
     parser.add_argument("--pass_cell_cls", type=int, default=0)
 
@@ -59,7 +61,7 @@ def parse_args():
     return args
 
 def solve_clustering(args, all_datasets):
-    trainset, test_data1, test_data2, label_dict = all_datasets
+    _, test_data1, test_data2, _ = all_datasets
 
     args.output_dir = helpers.create_downstream_output_dir(args)
 
@@ -119,10 +121,17 @@ def solve_clustering(args, all_datasets):
     names = CellTypeClassificationDataset.subsets["frac"]
     if args.indist:
         names = [names[0]]
+
+    if args.sample_outdist:
+        names = [names[1]]
 
     for name in names:
         # every data is tested under the same seeded setting
         helpers.set_seed(args.seed)
         args.data_source = f"frac_{name}"
-        all_datasets = data_loading.get_fracdata(name, args.data_branch, args.indist, False)
-        solve_clustering(args, all_datasets)
+        all_datasets = (
+            data_loading.get_fracdata_sample(name, num_samples=args.num_samples)
+            if args.sample_outdist else
+            data_loading.get_fracdata(name, args.data_branch, args.indist, False)
+        )
+        solve_clustering(args, all_datasets)

sccello/script/run_load_model.py

@@ -0,0 +1,8 @@
+import os
+import sys
+EXC_DIR = os.path.dirname(os.path.dirname(os.path.dirname(__file__)))
+sys.path.append(EXC_DIR)
+
+from sccello.src.model_prototype_contrastive import PrototypeContrastiveForMaskedLM
+
+model = PrototypeContrastiveForMaskedLM.from_pretrained("katarinayuan/scCello-zeroshot", output_hidden_states=True)

sccello/script/run_novel_cell_type_classification.py

@@ -65,7 +65,13 @@ def get_cell_type_labelid2nodeid(cell_type_idmap, clid2nodeid):
     # e.g., {'placental pericyte': CL:2000078, ...}
     name2clid = {v.lower(): k for k, v in clid2name.items()}
 
-    cell_type2nodeid = dict([(k, clid2nodeid[name2clid[cell_type2name[k]]]) if cell_type2name[k] in name2clid else (k, -1) for k in cell_type2name])
+    cell_type2nodeid = dict(
+        [
+            (k, clid2nodeid[name2clid[cell_type2name[k]]])
+            if cell_type2name[k] in name2clid else (k, -1)
+            for k in cell_type2name
+        ]
+    )
     return cell_type2nodeid
 
 def load_cell_type_representation(args, model):

sccello/src/data/dataset.py

@@ -32,8 +32,8 @@ class CellTypeClassificationDataset():
     @classmethod
     def create_dataset(cls, subset_name="celltype"):
         assert subset_name in cls.subsets["frac"]
-        valid_data = load_dataset(f"katarinayuan/scCello_ood_{subset_name}_data1")["train"]
-        test_data = load_dataset(f"katarinayuan/scCello_ood_{subset_name}_data2")["train"]
+        valid_data = load_dataset(f"katarinayuan/scCello_ood_{subset_name}_data1", split="train")
+        test_data = load_dataset(f"katarinayuan/scCello_ood_{subset_name}_data2", split="train")
 
         valid_data = valid_data.rename_column("cell_type", "label")
         test_data = test_data.rename_column("cell_type", "label")

sccello/src/utils/data_loading.py

@@ -112,6 +112,20 @@ def get_prestored_data(data_file_name):
     else:
         raise NotImplementedError
 
+def get_fracdata_sample(name, num_proc=12, num_samples=10):
+
+    from sccello.src.data.dataset import CellTypeClassificationDataset
+    data1, data2 = CellTypeClassificationDataset.create_dataset(name)
+    data1 = data1.rename_column("gene_token_ids", "input_ids")
+    data2 = data2.rename_column("gene_token_ids", "input_ids")
+
+    data1, data2 = data1.select(range(num_samples)), data2.select(range(num_samples))
+
+    data1, eval_label_type_idmap = helpers.process_label_type(data1, num_proc, "label")
+    data2, test_label_type_idmap = helpers.process_label_type(data2, num_proc, "label")
+
+    return None, data1, data2, None
+
 def get_fracdata(name, data_branch, indist, batch_effect, num_proc=12):
 
     from sccello.src.data.dataset import CellTypeClassificationDataset

sccello/src/utils/helpers.py

@@ -11,7 +11,6 @@
 EXC_DIR = os.path.dirname(os.path.dirname(os.path.dirname(__file__)))
 
 from sccello.src.utils import logging_util
-from sccello.src.model_prototype_contrastive import PrototypeContrastiveForMaskedLM
 
 
 def set_seed(seed):
@@ -77,6 +76,8 @@ def create_downstream_output_dir(args):
 
 def load_model_inference(args):
 
+    from sccello.src.model_prototype_contrastive import PrototypeContrastiveForMaskedLM
+
     model = eval(args.model_class).from_pretrained(args.pretrained_ckpt, output_hidden_states=True).to("cuda")
     for param in model.bert.parameters():
         param.requires_grad = False