update dependencies, reinstate compatibility with python<3.9 with typing for List

geneformer/in_silico_perturber.py

@@ -31,6 +31,7 @@ import logging
 import os
 import pickle
 from collections import defaultdict
+from typing import List
 
 import seaborn as sns
 import torch
@@ -873,7 +874,7 @@ class InSilicoPerturber:
         cos_sims_dict: defaultdict,
         cos_sims_data: torch.Tensor,
         filtered_input_data: Dataset,
-        indices_to_perturb: list[list[int]],
+        indices_to_perturb: List[List[int]],
         gene_list=None,
     ):
         if gene_list is not None and cos_sims_data.shape[0] != len(gene_list):

geneformer/perturber_utils.py

@@ -3,6 +3,7 @@ import logging
 import pickle
 import re
 from collections import defaultdict
+from typing import List
 
 import numpy as np
 import pandas as pd
@@ -275,9 +276,9 @@ def remove_indices_from_emb_batch(emb_batch, list_of_indices_to_remove, gene_dim
 def remove_perturbed_indices_set(
     emb,
     perturb_type: str,
-    indices_to_perturb: list[list],
-    tokens_to_perturb: list[list],
-    original_lengths: list[int],
+    indices_to_perturb: List[List],
+    tokens_to_perturb: List[List],
+    original_lengths: List[int],
     input_ids=None,
 ):
     if perturb_type == "overexpress":
@@ -306,7 +307,7 @@ def remove_perturbed_indices_set(
 
 def make_perturbation_batch(
     example_cell, perturb_type, tokens_to_perturb, anchor_token, combo_lvl, num_proc
-) -> tuple[Dataset, list[int]]:
+) -> tuple[Dataset, List[int]]:
     if combo_lvl == 0 and tokens_to_perturb == "all":
         if perturb_type in ["overexpress", "activate"]:
             range_start = 1

requirements.txt

@@ -0,0 +1,20 @@
+anndata>=0.9
+datasets>=2.12
+loompy>=3.0
+matplotlib>=3.7
+numpy>=1.23
+packaging>=23.0
+pandas>=2.0
+pyarrow>=12.0
+pytz>=2023.0
+ray>=2.6
+scanpy>=1.9
+scikit_learn>=1.2
+scipy>=1.10
+seaborn>=0.12
+setuptools>=65.6
+statsmodels>=0.14
+tdigest>=0.5.2
+torch>=2.0.1
+tqdm>=4.65
+transformers>=4.28

setup.py

@@ -11,12 +11,28 @@ setup(
                  predictions in settings with limited data in \
                  network biology.",
     packages=["geneformer"],
+    python_requires=">=3.10",
     include_package_data=True,
     install_requires=[
+        "anndata",
         "datasets",
         "loompy",
+        "matplotlib",
         "numpy",
+        "packaging",
+        "pandas",
+        "pyarrow",
+        "pytz",
+        "ray",
+        "scanpy",
+        "scikit-learn",
+        "scipy",
+        "seaborn",
+        "setuptools",
+        "statsmodels",
         "tdigest",
+        "torch",
+        "tqdm",
         "transformers",
     ],
 )