Merge branch 'unblock' into 'develop'

Update jaxlib and release v0.2.2

See merge request sacdallago/bio_embeddings!215

.gitlab-ci.yml

@@ -28,6 +28,7 @@ stages:
   <<: *install-deps
   variables:
     SKIP_EXPENSIVE_TESTS: 1
+    SKIP_AVX2_TESTS: 1
   stage: tests
   coverage: '/TOTAL.*\s(\d+\.\d+\%)/'
   script: poetry run pytest -v

CHANGELOG.md

@@ -1,8 +1,11 @@
 # Changelog
 
-## Unreleased
+## 0.2.2
 
  * Added the `esm1v` embedder from [Meier et al. 2021](https://www.biorxiv.org/content/10.1101/2021.07.09.450648v1), which is part of facebook's [esm](https://github.com/facebookresearch/esm). Note that this is an ensemble model, so you need to pass `ensemble_id` with a value from 1 to 5 to select which weights to use.
+ * Added the `bindEmbed21DL` extract protocol which is an ensemble of 5 convolutional neural network that predicts of 3 different types of binding residues (metal, nucleic acids, small molecules).
+ * Fix model download
+ * Update jaxlib to fix pip installation
 
 ## v0.2.1
 

bio_embeddings/extract/bindEmbed21DL/__init__.py

@@ -1 +1,5 @@
-from bio_embeddings.extract.bindEmbed21DL.bindEmbed21DL_annotation_extractor import BindEmbed21DLAnnotationExtractor
+from bio_embeddings.extract.bindEmbed21DL.bindEmbed21DL_annotation_extractor import (
+    BindEmbed21DLAnnotationExtractor,
+)
+
+__all__ = ["BindEmbed21DLAnnotationExtractor"]

bio_embeddings/extract/bindEmbed21DL/binding_residues_cnn.py

@@ -2,9 +2,9 @@
 
 
 class BindingResiduesCNN(nn.Module):
-    """ Convolutional neural network for prediction of 3 different types of binding residues (metal, nucleic acids,
+    """Convolutional neural network for prediction of 3 different types of binding residues (metal, nucleic acids,
     small molecules. Final output is determined by taking the average output probability from 5 different models from
-    5 cross-validation runs """
+    5 cross-validation runs"""
 
     n_features = 1024
     bottleneck_dim = 128
@@ -14,13 +14,22 @@ class BindingResiduesCNN(nn.Module):
     def __init__(self):
         super(BindingResiduesCNN, self).__init__()
         self.conv1 = nn.Sequential(
-            nn.Conv1d(in_channels=self.n_features, out_channels=self.bottleneck_dim, kernel_size=5, stride=1,
-                      padding=2),
+            nn.Conv1d(
+                in_channels=self.n_features,
+                out_channels=self.bottleneck_dim,
+                kernel_size=5,
+                stride=1,
+                padding=2,
+            ),
             nn.ELU(),
             nn.Dropout(self.dropout_rate),
-
-            nn.Conv1d(in_channels=self.bottleneck_dim, out_channels=self.n_classes, kernel_size=5, stride=1,
-                      padding=2),
+            nn.Conv1d(
+                in_channels=self.bottleneck_dim,
+                out_channels=self.n_classes,
+                kernel_size=5,
+                stride=1,
+                padding=2,
+            ),
         )
 
     def forward(self, x):