Merge pull request #135 from kathyxchen/pytorch-140-update

Minor changes and updating Selene to be compatible with PyTorch 1.4.0

Author: kathyxchen

README.md

@@ -9,7 +9,9 @@ Selene is a Python library and command line interface for training deep neural n
 We recommend using Selene with Python 3.6 or above. 
 Package installation should only take a few minutes (less than 10 minutes, typically ~2-3 minutes) with any of these methods (conda, pip, source). 
 
-**Install [PyTorch](https://pytorch.org/get-started/locally/).** If you have an NVIDIA GPU, install a version of PyTorch that supports it--Selene will run much faster with a discrete GPU.
+**First, install [PyTorch](https://pytorch.org/get-started/locally/).** If you have an NVIDIA GPU, install a version of PyTorch that supports it--Selene will run much faster with a discrete GPU. 
+The library is currently compatible with PyTorch versions between 0.4.1 and 1.4.0.
+We will continue to update Selene to be compatible with the latest version of PyTorch.
 
 ### Installing selene with [Anaconda](https://www.anaconda.com/download/) (for Linux):
 

docs/source/overview/cli.md

@@ -172,6 +172,9 @@ evaluate_model: !obj:selene_sdk.EvaluateModel {
     features: !obj:selene_sdk.utils.load_features_list {
         input_path: /path/to/features_list.txt
     },
+    use_features_ord: !obj:selene_sdk.utils.load_features_list {
+        input_path: /path/to/features_subset_ordered.txt
+    },
     trained_model_path: /path/to/trained/model.pth.tar,
     batch_size: 64,
     n_test_samples: 640000,
@@ -190,6 +193,7 @@ evaluate_model: !obj:selene_sdk.EvaluateModel {
 - `report_gt_feature_n_positives`: Default is 10. In total, each class/feature must have more than `report_gt_feature_n_positives` positive examples in the test set to be considered in the performance computation. The output file that reports each class's performance will report 'NA' for classes that do not have enough positive samples.
 - `use_cuda`: Default is False. Specify whether CUDA-enabled GPUs are available for torch to use.  
 - `data_parallel`: Default is False. Specify whether multiple GPUs are available for torch to use.
+- `use_features_ord`: Default is None. Specify an ordered list of features for which to run the evaluation. The features in this list must be identical to or a subset of `features`, and in the order you want the resulting `test_targets.npz` and `test_predictions.npz` to be saved.
 
 #### Additional notes
 Similar to the `train_model` configuration, any arguments that you find in [the documentation](https://selene.flatironinstitute.org/selene.html#evaluatemodel) that are not present in the function-type value's arguments are automatically instantiated and passed in by Selene.

selene_sdk/evaluate_model.py

@@ -3,6 +3,7 @@
 """
 import logging
 import os
+import warnings
 
 import numpy as np
 import torch
@@ -57,6 +58,11 @@ class EvaluateModel(object):
     data_parallel : bool, optional
         Default is `False`. Specify whether multiple GPUs are available
         for torch to use during training.
+    use_features_ord : list(str) or None, optional
+        Default is None. Specify an ordered list of features for which to
+        run the evaluation. The features in this list must be identical to or
+        a subset of `features`, and in the order you want the resulting
+        `test_targets.npz` and `test_predictions.npz` to be saved.
 
     Attributes
     ----------
@@ -88,7 +94,8 @@ def __init__(self,
                  n_test_samples=None,
                  report_gt_feature_n_positives=10,
                  use_cuda=False,
-                 data_parallel=False):
+                 data_parallel=False,
+                 use_features_ord=None):
         self.criterion = criterion
 
         trained_model = torch.load(
@@ -103,11 +110,26 @@ def __init__(self,
 
         self.sampler = data_sampler
 
-        self.features = features
-
         self.output_dir = output_dir
         os.makedirs(output_dir, exist_ok=True)
 
+        self.features = features
+        self._use_ixs = list(range(len(features)))
+        if use_features_ord is not None:
+            feature_ixs = {f: ix for (ix, f) in enumerate(features)}
+            self._use_ixs = []
+            self.features = []
+
+            for f in use_features_ord:
+                if f in feature_ixs:
+                    self._use_ixs.append(feature_ixs[f])
+                    self.features.append(f)
+                else:
+                    warnings.warn(("Feature {0} in `use_features_ord` "
+                                   "does not match any features in the list "
+                                   "`features` and will be skipped.").format(f))
+            self._write_features_ordered_to_file()
+
         initialize_logger(
             os.path.join(self.output_dir, "{0}.log".format(
                 __name__)),
@@ -130,11 +152,30 @@ def __init__(self,
 
         self._test_data, self._all_test_targets = \
             self.sampler.get_data_and_targets(self.batch_size, n_test_samples)
+        # TODO: we should be able to do this on the sampler end instead of
+        # here. the current workaround is problematic, since
+        # self._test_data still has the full featureset in it, and we
+        # select the subset during `evaluate`
+        self._all_test_targets = self._all_test_targets[:, self._use_ixs]
 
+        # reset Genome base ordering when applicable.
         if (hasattr(self.sampler, "reference_sequence") and
-                isinstance(self.sampler.reference_sequence, Genome) and
-                _is_lua_trained_model(model)):
-            Genome.update_bases_order(['A', 'G', 'C', 'T'])
+                isinstance(self.sampler.reference_sequence, Genome)):
+            if _is_lua_trained_model(model):
+                Genome.update_bases_order(['A', 'G', 'C', 'T'])
+            else:
+                Genome.update_bases_order(['A', 'C', 'G', 'T'])
+
+    def _write_features_ordered_to_file(self):
+        """
+        Write the feature ordering specified by `use_features_ord`
+        after matching it with the `features` list from the class
+        initialization parameters.
+        """
+        fp = os.path.join(self.output_dir, 'use_features_ord.txt')
+        with open(fp, 'w+') as file_handle:
+            for f in self.features:
+                file_handle.write('{0}\n'.format(f))
 
     def _get_feature_from_index(self, index):
         """
@@ -170,7 +211,7 @@ def evaluate(self):
         all_predictions = []
         for (inputs, targets) in self._test_data:
             inputs = torch.Tensor(inputs)
-            targets = torch.Tensor(targets)
+            targets = torch.Tensor(targets[:, self._use_ixs])
 
             if self.use_cuda:
                 inputs = inputs.cuda()
@@ -182,10 +223,11 @@ def evaluate(self):
                 predictions = None
                 if _is_lua_trained_model(self.model):
                     predictions = self.model.forward(
-                        inputs.transpose(1, 2).unsqueeze_(2))
+                        inputs.transpose(1, 2).contiguous().unsqueeze_(2))
                 else:
                     predictions = self.model.forward(
                         inputs.transpose(1, 2))
+                predictions = predictions[:, self._use_ixs]
                 loss = self.criterion(predictions, targets)
 
                 all_predictions.append(predictions.data.cpu().numpy())

selene_sdk/predict/_common.py

@@ -93,7 +93,8 @@ def predict(model, batch_sequences, use_cuda=False):
         inputs = Variable(inputs)
 
         if _is_lua_trained_model(model):
-            outputs = model.forward(inputs.transpose(1, 2).unsqueeze_(2))
+            outputs = model.forward(
+                inputs.transpose(1, 2).contiguous().unsqueeze_(2))
         else:
             outputs = model.forward(inputs.transpose(1, 2))
         return outputs.data.cpu().numpy()

selene_sdk/predict/_variant_effect_prediction.py

@@ -123,6 +123,7 @@ def read_vcf_file(input_path,
                 if not reference_sequence.coords_in_bounds(chrom, start, end):
                     na_rows.append(line)
                     continue
+            alt = alt.replace('.', ',')  # consider '.' a valid delimiter
             for a in alt.split(','):
                 variants.append((chrom, pos, name, ref, a, strand))
 

selene_sdk/predict/model_predict.py

@@ -176,6 +176,8 @@ def __init__(self,
         if type(self.reference_sequence) == Genome and \
                 _is_lua_trained_model(model):
             Genome.update_bases_order(['A', 'G', 'C', 'T'])
+        else:  # even if not using Genome, I guess we can update?
+            Genome.update_bases_order(['A', 'C', 'G', 'T'])
         self._write_mem_limit = write_mem_limit
 
     def _initialize_reporters(self,
@@ -424,11 +426,11 @@ def get_predictions_for_bed_file(self,
             batch_ids.append(label+(contains_unk,))
             sequences[ i % self.batch_size, :, :] = encoding
             if contains_unk:
-                warnings.warn("For region {0}, "
-                                "reference sequence contains unknown base(s). "
-                                "--will be marked `True` in the `contains_unk` column "
-                                "of the .tsv or the row_labels .txt file.".format(
-                                  label))
+                warnings.warn(("For region {0}, "
+                               "reference sequence contains unknown "
+                               "base(s). --will be marked `True` in the "
+                               "`contains_unk` column of the .tsv or "
+                               "row_labels .txt file.").format(label))
 
         if (batch_ids and i == 0) or i % self.batch_size != 0:
             sequences = sequences[:i % self.batch_size + 1, :, :]

selene_sdk/sequences/genome.py

@@ -17,7 +17,7 @@
 def _not_blacklist_region(chrom, start, end, blacklist_tabix):
     """
     Check if the input coordinates are not overlapping with blacklist regions.
-    
+
     Parameters
     ----------
     chrom : str
@@ -29,14 +29,14 @@ def _not_blacklist_region(chrom, start, end, blacklist_tabix):
     blacklist_tabix : tabix.open or None, optional
         Default is `None`. Tabix file handle if a file of blacklist regions
         is available.
-    
+
     Returns
     -------
     bool
         False if the coordinates are overlaping with blacklist regions
         (if specified). Otherwise, return True.
-    
-    
+
+
     """
     if blacklist_tabix is not None:
         try:
@@ -203,7 +203,7 @@ class Genome(Sequence):
 
     """
 
-    BASES_ARR = np.array(['A', 'C', 'G', 'T'])
+    BASES_ARR = ['A', 'C', 'G', 'T']
     """
     This is an array with the alphabet (i.e. all possible symbols
     that may occur in a sequence). We expect that
@@ -463,7 +463,7 @@ def get_encoding_from_coords_check_unk(self,
                                  strand='+',
                                  pad=False):
         """Gets the one-hot encoding of the genomic sequence at the
-        queried coordinates and check whether the sequence contains 
+        queried coordinates and check whether the sequence contains
         unknown base(s).
 
         Parameters

selene_sdk/sequences/proteome.py

@@ -71,8 +71,8 @@ class around the `pyfaidx.Fasta` class.
 
     """
 
-    BASES_ARR = np.array(['A', 'R', 'N', 'D', 'C', 'E', 'Q', 'G', 'H', 'I',
-                          'L', 'K', 'M', 'F', 'P', 'S', 'T', 'W', 'Y', 'V'])
+    BASES_ARR = ['A', 'R', 'N', 'D', 'C', 'E', 'Q', 'G', 'H', 'I',
+                 'L', 'K', 'M', 'F', 'P', 'S', 'T', 'W', 'Y', 'V']
     """
     This is an array with the alphabet (i.e. all possible symbols
     that may occur in a sequence). We expect that

selene_sdk/version.py

@@ -1 +1 @@
-__version__ = "0.4.4"
+__version__ = "0.4.5"

setup.py

@@ -25,7 +25,7 @@
 cmdclass = {'build_ext': build_ext}
 
 setup(name="selene-sdk",
-      version="0.4.4",
+      version="0.4.5",
       long_description=long_description,
       long_description_content_type='text/markdown',
       description=("framework for developing sequence-level "
@@ -62,6 +62,5 @@
         "scipy",
         "seaborn",
         "statsmodels",
-        "torch>=0.4.1",
-        "torchvision"
+        "torch>=0.4.1, <=1.4.0",
     ])