Merging First official version

Prepare version for official launch with publication

autopeptideml/autopeptideml.py

@@ -1,26 +1,25 @@
-from copy import deepcopy
-from multiprocessing import cpu_count
+import joblib
 import os
-from typing import Dict, List, Optional, Union
 
-import joblib
-import matplotlib
-matplotlib.use('Agg')
+from copy import deepcopy
+from multiprocessing import cpu_count
+from typing import Dict, List, Union
 
-from hestia.partition import ccpart, graph_part
 import matplotlib.pyplot as plt
 import numpy as np
 import optuna
 import pandas as pd
-import scikitplot.metrics as skplt
 import sklearn.metrics
+
 from sklearn.model_selection import StratifiedKFold
+from hestia.partition import ccpart, graph_part
 
 from .data.algorithms import SYNONYMS, SUPPORTED_MODELS
 from .data.metrics import METRICS, METRIC2FUNCTION, THRESHOLDED_METRICS
 from .data.residues import is_canonical
 from .utils.embeddings import RepresentationEngine
-from .utils.training import FlexibleObjective, UniDL4BioPep_Objective, ModelSelectionObjective
+from .utils.training import (FlexibleObjective, UniDL4BioPep_Objective,
+                             ModelSelectionObjective)
 
 
 class AutoPeptideML:
@@ -60,7 +59,7 @@ def autosearch_negatives(
         self,
         df_pos: pd.DataFrame,
         positive_tags: List[str],
-        proportion: float=1.0
+        proportion: float = 1.0
     ) -> pd.DataFrame:
         """Method for searching bioactive databases for peptides
 
@@ -122,7 +121,7 @@ def autosearch_negatives(
                 else:
                     missing += samples_to_draw
                 df_neg = pd.concat([df_neg, subdf])
-        
+
         get_tags = lambda x: ';'.join(sorted(set(
             [column if (x[column] == 1) else df_neg.columns[0]
              for column in self.tags]
@@ -643,6 +642,10 @@ def _fasta2csv(self, dataset: str, outputdir: str) -> pd.DataFrame:
         return df
 
     def _make_figures(self, figures_path: str, truths, preds_proba):
+        import matplotlib
+        matplotlib.use('Agg')
+        import scikitplot.metrics as skplt
+
         preds = preds_proba > 0.5
         new_preds_proba = np.zeros((len(preds_proba), 2))
         new_preds_proba[:, 0] = 1 - preds_proba

autopeptideml/data/algorithms.py

@@ -1,20 +1,21 @@
 from lightgbm import LGBMClassifier
-import sklearn.ensemble
-import sklearn.neighbors
-import sklearn.neural_network
+from sklearn.ensemble import RandomForestClassifier, AdaBoostClassifier
+from sklearn.neighbors import KNeighborsClassifier
+from sklearn.neural_network import MLPClassifier
 import sklearn.svm
 import sklearn.metrics
+from xgboost import XGBClassifier
 
 from ..utils.unidl4biopep.model import Cnn
 
 
 SUPPORTED_MODELS = {
-    'knn': sklearn.neighbors.KNeighborsClassifier,
-    'rfc': sklearn.ensemble.RandomForestClassifier,
+    'knn': KNeighborsClassifier,
+    'rfc': RandomForestClassifier,
     'svm': sklearn.svm.SVC,
-    'mlp': sklearn.neural_network.MLPClassifier,
-    'xgboost': sklearn.ensemble.GradientBoostingClassifier,
-    'adaboost': sklearn.ensemble.AdaBoostClassifier,
+    'mlp': MLPClassifier,
+    'xgboost': XGBClassifier,
+    'adaboost': AdaBoostClassifier,
     'unidl4biopep': Cnn,
     'lightgbm': LGBMClassifier
 }

autopeptideml/main.py

@@ -50,6 +50,7 @@ def parse_cli():
 
     return parser.parse_args()
 
+
 def parse_cli_predict():
     parser = argparse.ArgumentParser()
     parser.add_argument('dataset', type=str)

autopeptideml/utils/embeddings.py

@@ -38,7 +38,7 @@
 }
 
 
-class RepresentationEngine(torch.nn.Module):    
+class RepresentationEngine(torch.nn.Module):
     def __init__(self, model: str, batch_size: int):
         super().__init__()
         self.device = 'cuda' if torch.cuda.is_available() else 'cpu'

setup.py

@@ -19,6 +19,7 @@
     'transformers',
     'lightgbm',
     'mdpdf',
+    'xgboost',
     'hestia-ood'
 ]
 
@@ -51,6 +52,6 @@
     name='autopeptideml',
     packages=find_packages(exclude=['examples']),
     url='https://ibm.github.io/AutoPeptideML/',
-    version='0.3.4',
+    version='1.0.0',
     zip_safe=False,
 )

tests/test_apml.py

@@ -0,0 +1,7 @@
+from autopeptideml import AutoPeptideML
+
+
+def test_load():
+    apml = AutoPeptideML()
+    df = apml.curate_dataset('examples/AB_positives.csv')
+    assert len(df) == 6_583