Merge pull request #23 from mskcc/feature/add-py-support

add additional python version support and ci config

Author: john-ziegler

.travis.yml

@@ -3,6 +3,8 @@ python:
     - 2.7
     - 3.5
     - 3.6
+    - 3.7
+    - 3.8
 before_install:
     - sudo apt-get -y update
 install:

analyze.py

@@ -28,7 +28,7 @@
 from torch.autograd import Variable
 from sklearn import metrics
 import traceback
-
+import pkg_resources
 from data.generate_vectors.create_data import create_data
 from main.evaluate_sample import run_eval
 
@@ -38,6 +38,12 @@
 
 def main():
     parser = argparse.ArgumentParser(description="MiMSI Analysis")
+    parser.add_argument(
+        "--version",
+        action="store_true",
+        default=False,
+        help="Display current version of MiMSI",
+    )
     parser.add_argument(
         "--no-cuda",
         action="store_true",
@@ -125,6 +131,10 @@ def main():
     )
 
     args = parser.parse_args()
+    if args.version:
+        print("MiMSI Case Analysis CLI version - " + pkg_resources.require("MiMSI")[0].version)
+        return 
+
     case_list, tumor_bam, normal_bam, case_id, norm_case_id, ms_list, save_loc, cores, saved_model, no_cuda, seed, save, save_format, name, covg, confidence, use_attention = (
         args.case_list,
         args.tumor_bam,

data/data_loader.py

@@ -105,7 +105,7 @@ def __getitem__(self, idx):
 
         sample_id, label = self.__parsefilename__(self.data_file_list[idx])
         final_bags = []
-        bags = self._preprocess(np.load(data_file))
+        bags = self._preprocess(np.load(data_file, allow_pickle=True, encoding="bytes"))
         for bag in bags:
 
             bag = [

data/generate_vectors/create_data.py

@@ -24,6 +24,7 @@
 import warnings
 import pysam
 import traceback
+import pkg_resources
 from glob import glob
 
 from .bam2tensor import Bam2Tensor
@@ -296,7 +297,7 @@ def create_data(
             raise Exception(
                 "Label column in case list can be empty or contain one the values +1 (MSI) or -1 (MSS)."
             )
-        cases = cases.replace({pd.np.nan: None})
+        cases = cases.replace({np.nan: None})
         for index, row in cases.iterrows():
             tumor_id, normal_id, tumor_bam, normal_bam, label = map(
                 lambda x: row[x] if x in row and row[x] else None,
@@ -340,6 +341,12 @@ def create_data(
 
 def main():
     parser = argparse.ArgumentParser(description="MiMSI Vector Generation Utility")
+    parser.add_argument(
+        "--version",
+        action="store_true",
+        default=False,
+        help="Display current version of MiMSI",
+    )
 
     single_sample_group = parser.add_argument_group("Single Sample Mode")
     single_sample_group.add_argument(
@@ -389,6 +396,11 @@ def main():
     )
 
     args = parser.parse_args()
+
+    if args.version:
+        print("MiMSI Vector Creation CLI version - " + pkg_resources.require("MiMSI")[0].version)
+        return 
+
     case_list, tumor_bam, normal_bam, case_id, m_list, save_loc, covg, cores, norm_case_id = (
         args.case_list,
         args.tumor_bam,

main/evaluate_sample.py

@@ -22,6 +22,7 @@
 import torch
 import torch.utils.data as data_utils
 import torch.optim as optim
+import pkg_resources
 from torch.autograd import Variable
 from sklearn import metrics
 
@@ -90,7 +91,7 @@ def sample_name_parser(sample):
         try:
             os.mkdir(save_loc)
         except OSError as e:
-            if e.errno != os.errno.EEXIST:
+            if e.errno != errno.EEXIST:
                 print("Exception when creating directory to store final results.")
                 raise
 
@@ -191,6 +192,12 @@ def run_eval(
 
 def main():
     parser = argparse.ArgumentParser(description="MiMSI Sample(s) Evalution Utility")
+    parser.add_argument(
+        "--version",
+        action="store_true",
+        default=False,
+        help="Display current version of MiMSI",
+    )
     parser.add_argument(
         "--no-cuda",
         action="store_true",
@@ -250,6 +257,10 @@ def main():
 
     args = parser.parse_args()
 
+    if args.version:
+        print("MiMSI Case Evaluation Utility version - " + pkg_resources.require("MiMSI")[0].version)
+        return 
+
     saved_model, vector_location, no_cuda, seed, save, save_format, save_loc, name, coverage, confidence, use_attention = (
         args.model,
         args.vector_location,

main/mi_msi_train_test.py

@@ -23,16 +23,23 @@
 import torch
 import torch.utils.data as data_utils
 import torch.optim as optim
+import pkg_resources
 from torch.autograd import Variable
 from sklearn import metrics
 from copy import deepcopy
-from .data.data_loader import MSIBags
-from .model.mi_msi_model import MSIModel
+from data.data_loader import MSIBags
+from model.mi_msi_model import MSIModel
 
 
 def main():
     parser = argparse.ArgumentParser(
         description='MiMSI - A Multiple Instance Learning Model for detecting microsatellite instability in NGS data')
+    parser.add_argument(
+        "--version",
+        action="store_true",
+        default=False,
+        help="Display current version of MiMSI",
+    )
     parser.add_argument('--epochs', type=int, default=40, metavar='N',
                         help='Number of epochs to train (default: 40)')
     parser.add_argument('--lr', type=float, default=0.0001, metavar='LR',
@@ -54,6 +61,11 @@ def main():
 
 
     args = parser.parse_args()
+
+    if args.version:
+        print("MiMSI Model Training & Testing CLI version - " + pkg_resources.require("MiMSI")[0].version)
+        return 
+
     # Training settings
     epochs, lr, reg, seed, name, train_location, test_location, save = args.epochs, args.lr, args.reg, args.seed, args.name, args.train_location, args.test_location, args.save
     cuda = not args.no_cuda and torch.cuda.is_available()

requirements.txt

@@ -1,13 +1,18 @@
-future==0.18.2
+future; python_version <= '2.7'
 nose>=1.3.7
-numpy==1.16.1
-pandas==0.24.2
-Pillow==5.3.0
+numpy==1.16.5
+pandas==0.24.2; python_version <= '2.7'
+pandas
+pillow>=8.1.1; python_version >= '3.6'
+pillow
 pysam==0.16.0.1
 python-dateutil==2.8.1
-scikit-learn==0.20.0
-scipy==1.1.0
+scikit-learn>=0.20.0; python_version <= '2.7'
+scikit-learn
+scipy>=1.1.0; python_version <= '2.7'
+scipy
 six==1.11.0
-sklearn==0.0
+sklearn==0.0; python_version <= '2.7'
+sklearn
 torch==1.4.0
 torchvision==0.5.0

setup.py

@@ -29,7 +29,7 @@ def most_recent_tag():
 
 setup(
     name="MiMSI",
-    version=most_recent_tag(),
+    version="v0.4.1",
     description="A deep, multiple instance learning based classifier for identifying Microsatellite Instability from NGS",
     url="https://github.com/mskcc/mimsi",
     author="John Ziegler",