Update api_encoding_train_test.ipynb and ray as optional import

Author: niklases

pypef/dca/plmc_encoding.py

@@ -46,11 +46,15 @@
 logger = logging.getLogger('pypef.dca.encoding')
 
 import numpy as np
-import ray
 from tqdm import tqdm
 import pickle
 
+from pypef.main import USE_RAY
+if USE_RAY:
+    import ray
+
 from pypef.utils.variant_data import amino_acids
+from pypef.ml.parallelization import ConditionalDecorator
 
 _SLICE = np.s_[:]
 
@@ -741,7 +745,7 @@ def get_encoded_sequence(
     return encoded_seq
 
 
-@ray.remote
+@ConditionalDecorator(ray.remote, USE_RAY)
 def _get_data_parallel(
         variants: list,
         sequences: list,

pypef/hybrid/hybrid_run.py

@@ -14,7 +14,10 @@
 
 import logging
 logger = logging.getLogger('pypef.dca.dca_run')
-import ray
+
+from pypef.main import USE_RAY
+if USE_RAY:
+    import ray
 
 from pypef.utils.variant_data import read_csv, get_wt_sequence
 from pypef.dca.plmc_encoding import save_plmc_dca_encoding_model
@@ -34,7 +37,7 @@ def run_pypef_hybrid_modeling(arguments):
         model_type = 'undefined'
     if model_type in ['GREMLIN', 'DCAHybridModel'] and threads > 1:
         logger.info(f'No (Ray) parallelization for {model_type} model...')
-    elif model_type not in ['GREMLIN', 'DCAHybridModel'] and threads > 1:
+    elif model_type not in ['GREMLIN', 'DCAHybridModel'] and threads > 1 and USE_RAY:
         ray.init()
         logger.info(f'Using {threads} threads for running...')
     if model_type == 'DCAHybridModel':

pypef/main.py

@@ -286,6 +286,7 @@
                                     A2M format [default: False].
 """
 
+USE_RAY = False
 
 from sys import argv, version_info
 from pypef import __version__

pypef/ml/ml_run.py

@@ -15,10 +15,14 @@
 import os
 from os import listdir
 from os.path import isfile, join
-import ray
+
 import logging
 logger = logging.getLogger('pypef.ml.ml_run')
 
+from pypef.main import USE_RAY
+if USE_RAY:
+    import ray
+
 from pypef.ml.parallelization import aaindex_performance_parallel
 from pypef.ml.regression import (
     read_models, formatted_output, performance_list, save_model, predict, predict_ts
@@ -35,7 +39,7 @@ def run_pypef_pure_ml(arguments):
     """
     threads = abs(arguments['--threads']) if arguments['--threads'] is not None else 1
     threads = threads + 1 if threads == 0 else threads
-    if threads > 1:
+    if threads > 1 and USE_RAY:
         ray.init()
     if arguments['--show']:
         if arguments['MODELS'] != str(5):

pypef/ml/parallelization.py

@@ -21,9 +21,12 @@
 import matplotlib
 matplotlib.use('Agg')
 import os
-import ray
 import warnings
 
+from pypef.main import USE_RAY
+if USE_RAY:
+    import ray
+
 from pypef.utils.variant_data import get_sequences_from_file
 from pypef.ml.regression import (
     full_aaidx_txt_path, path_aaindex_dir, AAIndexEncoding, get_regressor_performances
@@ -34,7 +37,19 @@
 warnings.filterwarnings(action='ignore', category=UserWarning, module='sklearn')
 
 
-@ray.remote
+class ConditionalDecorator(object):
+    def __init__(self, dec, condition):
+        self.decorator = dec
+        self.condition = condition
+
+    def __call__(self, func):
+        if not self.condition:
+            # Return the function unchanged, not decorated
+            return func
+        return self.decorator(func)
+
+
+@ConditionalDecorator(ray.remote, USE_RAY)
 def parallel(
         d,
         core,