Reorganize, update code, and add travis/tox

Closes #3. Begins #2

.travis.yml

@@ -0,0 +1,27 @@
+sudo: false
+cache: pip
+language: python
+python:
+  - "3.6"
+stages:
+  - lint
+  - docs
+jobs:
+  include:
+    # lint stage
+    - stage: lint
+      env: TOXENV=manifest
+    - env: TOXENV=flake8
+    - env: TOXENV=xenon
+    # docs stage
+    - stage: docs
+      env: TOXENV=doc8
+    - env: TOXENV=readme
+    - env: TOXENV=docs
+matrix:
+  allow_failures:
+    - env: TOXENV=xenon
+install:
+  - pip install tox
+script:
+  - tox

LICENSE

@@ -1,6 +1,6 @@
 MIT License
 
-Copyright (c) 2019 
+Copyright (c) GuiltyTargets Developers, 2019
 
 Permission is hereby granted, free of charge, to any person obtaining a copy
 of this software and associated documentation files (the "Software"), to deal

setup.cfg

@@ -2,7 +2,7 @@
 # Setup.py Configuration #
 ##########################
 [metadata]
-name = guilty_reloaded
+name = guiltytargets_phewas
 version = 0.0.1-dev
 description = Change meeeeee
 long_description = file: README.rst
@@ -13,15 +13,17 @@ author_email = [email protected]
 maintainer = Mauricio Pio de Lacerda
 maintainer_email = [email protected]
 
+url = https://github.com/GuiltyTargets/phewas
+
 # License Information
-license = Apache 2.0 License
+license = MIT
 license_file = LICENSE
 
 # Search tags
 classifiers =
     Development Status :: 1 - Planning
     Intended Audience :: Developers
-    License :: OSI Approved :: Apache Software License
+    License :: OSI Approved :: MIT License
     Operating System :: OS Independent
     Programming Language :: Python
     Programming Language :: Python :: 3.6

src/guiltytargets_phewas/__init__.py

@@ -0,0 +1,3 @@
+# -*- coding: utf-8 -*-
+
+"""GuiltyTargets PheWAS Extension."""

src/guilty_phewas/constants.py → src/guiltytargets_phewas/constants.py

@@ -1,6 +1,8 @@
 # -*- coding: utf-8 -*-
 
+"""Constants for GuiltyTargets PheWAS."""
+
 data_dir = 'data'
 disease_abr = ['ad']
 disease_ids_efo = ['EFO_0000249']
-ot_file = 'ot_entrez.txt'
+ot_file = 'ot_entrez.txt'

src/guilty_phewas/network_phewas.py → src/guiltytargets_phewas/network_phewas.py

@@ -7,20 +7,27 @@
 
 import networkx as nx
 import numpy as np
-from ppi_network_annotation.model.gene import Gene
-from pybel.dsl import protein, rna, gene
+from pybel.dsl import gene, protein, rna
+
+from guiltytargets.ppi_network_annotation import Gene
+
+__all__ = [
+    'NetworkNx',
+]
 
 logger = logging.getLogger(__name__)
 
 
 class NetworkNx:
     """Encapsulate a PPI network with differential gene expression, phenotypes and disease association annotation."""
 
-    def __init__(self,
-                 ppi_graph: nx.Graph,
-                 max_adj_p: Optional[float] = None,
-                 max_l2fc: Optional[float] = None,
-                 min_l2fc: Optional[float] = None):
+    def __init__(
+            self,
+            ppi_graph: nx.Graph,
+            max_adj_p: Optional[float] = None,
+            max_l2fc: Optional[float] = None,
+            min_l2fc: Optional[float] = None,
+    ) -> None:
         """Initialize the network object.
 
         :param ppi_graph: A graph of protein interactions.
@@ -63,8 +70,11 @@ def filter_genes(self, relevant_entrez: list) -> None:
         logger.info("In filter_genes()")
         raise Exception('Not ready to filter genes using NetworkX')
 
-    def _add_vertex_attributes(self, genes: List[Gene],
-                               disease_associations: Optional[dict] = None) -> None:
+    def _add_vertex_attributes(
+            self,
+            genes: List[Gene],
+            disease_associations: Optional[dict] = None,
+    ) -> None:
         """Add attributes to vertices.
 
         :param genes: A list of genes containing attribute information.
@@ -158,8 +168,7 @@ def get_attribute_from_indices(self, indices: list, attribute_name: str):
         return list(np.array(self.graph.vs[attribute_name])[indices])
 
     def get_differentially_expressed_genes(self, diff_type: str) -> List:
-        """Get up regulated, down regulated, all differentially expressed or not
-        differentially expressed nodes.
+        """Get up regulated, down regulated, all differentially expressed or not differentially expressed nodes.
 
         :param diff_type: One of `not_diff_expressed`, `diff_expressed`, `up_regulated`, `down_regulated`
         :return: A list of nodes corresponding to diff_type.

src/guilty_phewas/pulearn.py → src/guiltytargets_phewas/pulearn.py

@@ -3,21 +3,22 @@
 from collections import defaultdict
 
 import pandas as pd
-from GAT2VEC import parsers, paths
-from GAT2VEC.evaluation.classification import Classification
 from sklearn import svm
 from sklearn.metrics import accuracy_score, f1_score, roc_auc_score
 from sklearn.model_selection import StratifiedKFold
 
+from guiltytargets.gat2vec import Classification, Gat2Vec, gat2vec_parsers, gat2vec_paths
 
-# TODO move to GAT2VEC (forked)?
+__all__ = [
+    'PULearn',
+]
 
 
+# TODO move to GAT2VEC (forked)?
 class PULearn(Classification):
-    """ """
 
     def evaluate(self, model, label=False, evaluation_scheme="tr", cost_p: float = None, cost_n: float = None):
-        """Evaluates the model according to the given evaluation scheme.
+        """Evaluate the model according to the given evaluation scheme.
 
         :param model:
         :param label:
@@ -30,7 +31,7 @@ def evaluate(self, model, label=False, evaluation_scheme="tr", cost_p: float = N
         clf = self.get_classifier()
 
         if not label:
-            embedding = parsers.get_embeddingDF(model)
+            embedding = gat2vec_parsers.get_embeddingDF(model)
 
         if evaluation_scheme == "bsvm":
             if cost_n and cost_p and cost_n > 0 and cost_p > 0:
@@ -44,9 +45,9 @@ def evaluate(self, model, label=False, evaluation_scheme="tr", cost_p: float = N
             for tr in self.TR:
                 print("TR ... ", tr)
                 if label:
-                    model = paths.get_embedding_path_wl(self.dataset_dir, self.output_dir, tr)
+                    model = gat2vec_paths.get_embedding_path_wl(self.dataset_dir, self.output_dir, tr)
                     if isinstance(model, str):
-                        embedding = parsers.get_embeddingDF(model)
+                        embedding = gat2vec_parsers.get_embeddingDF(model)
                 results.update(self.evaluate_tr(clf, embedding, tr))
 
         print("Training Finished")
@@ -89,9 +90,7 @@ def get_biased_predictions(clf, x_train, x_test, y_train, cost_p, cost_n):
         return clf.predict(x_test), clf.predict_proba(x_test)
 
 
-if __name__ == '__main__':
-    from GAT2VEC.gat2vec import Gat2Vec
-
+def main():
     dir_ = "C:/Users/Mauricio/Thesis/bel_data/alzh"
     g2v = Gat2Vec(dir_, dir_, label=False, tr=[0.1, 0.3, 0.5])
     walk_length = 4
@@ -111,3 +110,7 @@ def get_biased_predictions(clf, x_train, x_test, y_train, cost_p, cost_n):
     print('PU created.')
     auc_df = pul.evaluate(model, label=False, evaluation_scheme="bsvm")
     print(auc_df)
+
+
+if __name__ == '__main__':
+    main()

src/guiltytargets_phewas/script.py

@@ -0,0 +1,28 @@
+# -*- coding: utf-8 -*-
+
+from guiltytargets.constants import gat2vec_config
+from guiltytargets.gat2vec import Classification, Gat2Vec
+
+
+def main():
+    dir_ = "C:/Users/Mauricio/Thesis/bel_data/alzh"  # windows
+    # dir_ = "C:\\Users\\Mauricio\\Thesis\\git\\reproduction\\data\\lc"
+
+    classifier = Classification(dir_, dir_, tr=gat2vec_config.training_ratio)
+
+    g2v = Gat2Vec(dir_, dir_, label=False, tr=gat2vec_config.training_ratio)
+    model = g2v.train_gat2vec(
+        gat2vec_config.num_walks,
+        gat2vec_config.walk_length,
+        gat2vec_config.dimension,
+        gat2vec_config.window_size,
+        output=True,
+    )
+
+    classifier = Classification(dir_, dir_, tr=gat2vec_config.training_ratio)
+
+    auc_df = classifier.evaluate(model, label=False, evaluation_scheme="cv")
+
+
+if __name__ == '__main__':
+    main()

src/guilty_phewas/string_ppi_assembler.py → src/guiltytargets_phewas/string_ppi_assembler.py

@@ -68,6 +68,10 @@ def _get_ensembl_id_to_symbol_converter(self) -> Dict:
         return dict(data3[['protein_external_id', 'preferred_name']].values)
 
 
-if __name__ == '__main__':
+def main():
     assembler = StringAssembler()
     assembler.create_adj_file()
+
+
+if __name__ == '__main__':
+    main()

src/guilty_phewas/utils.py → src/guiltytargets_phewas/utils.py

@@ -6,13 +6,13 @@
 from typing import Dict, List, Optional
 
 import bio2bel_phewascatalog
-from guiltytargets.ppi_network_annotation.model.gene import Gene
 import mygene
-import pandas as pd
 import networkx as nx
+import pandas as pd
 from opentargets import OpenTargetsClient
 from pybel.dsl import BaseEntity, gene, protein, rna
 
+from guiltytargets.ppi_network_annotation.model.gene import Gene
 from .network_phewas import NetworkNx
 
 
@@ -67,9 +67,11 @@ def get_significantly_differentiated(gene_list: List[Gene], max_adjp: float):
     """Returns a dictionary only with significantly differentially expressed genes from the gene list."""
     max_adjp = max_adjp or 0.05
 
-    dge = {g.entrez_id or g.symbol: g.log2_fold_change 
-           for g in gene_list 
-           if g.padj < max_adjp}
+    dge = {
+        g.entrez_id or g.symbol: g.log2_fold_change
+        for g in gene_list
+        if g.padj < max_adjp
+    }
 
     return {k: v for k, v in dge.items() if k}
 
@@ -123,8 +125,10 @@ def get_association_scores(disease_id, outpath):
         fields=['association_scoreoverall', 'target.id']
     )
     assoc_simple = [
-        {'id': a['target']['id'],
-         'score': a['association_score']['overall']}
+        {
+            'id': a['target']['id'],
+            'score': a['association_score']['overall']
+        }
         for a in assoc
     ]
     ensembl_list = [a['id'] for a in assoc_simple]
@@ -305,4 +309,3 @@ def _add_attribute_values(value, att_mappings, indices):
         """
         for i in indices:
             att_mappings[i].append(value)
-