joblib

operation/parallel_inference.py

@@ -2,6 +2,13 @@ import os
 import numpy as np
 from azureml.core import Model
 import joblib
+#import argparse
+
+#parser = argparse.ArgumentParser()
+#parser.add_argument("--id-feature", type=str, dest='id_feature', help='ID Freature')
+#args = parser.parse_args()
+#id_feat = str(args.id_feature)
+#print('id feature',  id_feat)
 
 
 def init():
@@ -13,13 +20,8 @@ def init():
     model = joblib.load(model_path)
     
 def run(mini_batch):
-#     # This runs for each batch
-#     resultList = model.predict(mini_batch)
-#     ind = mini_batch.index
-#     return resultList.tolist()
+    mini_batch.set_index('Van_Stock_Proposal_Detail_Id', inplace=True)
     index_list = list(mini_batch.index)
     y_pred = model.predict(mini_batch).tolist()
     score = model.score_samples(mini_batch).tolist()
-    
-    
     return(list(zip(index_list, y_pred, score)))

operation/parallel_overhead.py

@@ -3,7 +3,7 @@ from azureml.core import Model, Run
 import argparse
 import numpy as np
 import iJungle
-import pickle
+import joblib
 
 run = Run.get_context()
 
@@ -14,6 +14,7 @@ parser = argparse.ArgumentParser()
 
 # Input Data
 parser.add_argument("--input-data", type=str, dest='input_data', help='Overhead dataset')
+parser.add_argument("--id-feature", type=str, dest='id_feature', help='ID Freature')
 
 # Hyper parameters
 parser.add_argument('--trees', type=int, dest='trees', default=100, help='Number of trees')
@@ -21,6 +22,8 @@ parser.add_argument('--subsample-size', type=int, dest='subsample_size', default
 
 # Add arguments to args collection
 args = parser.parse_args()
+id_feat = str(args.id_feature)
+print('id feature',  id_feat)
 
 # Log Hyperparameter values
 trees = np.int(args.trees)
@@ -33,6 +36,7 @@ run.log('subsample_size',  subsample_size)
 # Load training data
 print("Loading Data...")
 W = run.input_datasets['overhead_data'].to_pandas_dataframe() # Get the training data from the estimator input
+W.set_index(id_feat, inplace=True)
 
 # Load iFor_list pickle
 print("Loading pickle...")
@@ -40,9 +44,7 @@ model_name = 'iJungle_light_' + str(trees) + '_' + str(subsample_size)
 print(model_name)
 model_path = Model.get_model_path(model_name)
 print(model_path)
-with open(model_path, 'rb') as infile:
-    iFor_list = pickle.load(infile)
-
+iFor_list = joblib.load(model_path)
 
 # Evaluation
 print("Starting evaluation ...")
@@ -50,8 +52,7 @@ os.makedirs(iJungle._MODEL_DIR, exist_ok=True)
 results = iJungle.model_eval_fun(W, iFor_list)
 results_filename = os.path.join(iJungle._MODEL_DIR, model_name + '_results.pkl')
 print("Writing results:", results_filename)
-with open(results_filename, 'wb') as outfile:
-    pickle.dump(results, outfile)
+joblib.dump(value=results, filename=results_filename)
 
 # Log dummy metric
 run.log('Dummy', np.float(0))

operation/parallel_train.py

@@ -12,6 +12,7 @@ parser = argparse.ArgumentParser()
 
 # Input Data
 parser.add_argument("--input-data", type=str, dest='input_data', help='training dataset')
+parser.add_argument("--id-feature", type=str, dest='id_feature', help='ID Freature')
 parser.add_argument("--max-subsample-size", type=int, dest='max_sss', help='Max subsample size')
 parser.add_argument("--train-size", type=float, dest='train_size', help='Train size')
 
@@ -21,6 +22,8 @@ parser.add_argument('--subsample-size', type=int, dest='subsample_size', default
 
 # Add arguments to args collection
 args = parser.parse_args()
+id_feat = str(args.id_feature)
+print('id feature',  id_feat)
 
 # Log Hyperparameter values
 trees = np.int(args.trees)
@@ -41,7 +44,7 @@ run.log('train_size',  train_size)
 # Load training data
 print("Loading Data...")
 df = run.input_datasets['training_data'].to_pandas_dataframe() # Get the training data from the estimator input
-
+df.set_index(id_feat, inplace=True)
 
 print("Starting training ...")
 model_filename = iJungle.model_train_fun(df, trees, subsample_size, train_size, max_sss)

src/iJungle/config.py

@@ -1,2 +1,2 @@
-__version__ = '0.1.66'
+__version__ = '0.1.73'
 _MODEL_DIR = 'outputs'

src/iJungle/train.py

@@ -3,7 +3,7 @@ from iJungle.config import _MODEL_DIR
 
 import random
 from sklearn.ensemble import IsolationForest
-import pickle
+import joblib
 import os
 import numpy as np
 import pandas as pd
@@ -67,9 +67,8 @@ def model_train_fun(df, trees=100, subsample_size=8192, train_size = 0.2, max_ss
                 print("{}/{}".format(counter, int(df_len/max_sss+1)))
             
         filename = 'iJungle_light_' + str(trees) + '_' + str(subsample_size) + '.pkl'
-        with open(os.path.join(_MODEL_DIR, filename), 'wb') as outfile:
-            pickle.dump(iFor_list, outfile)
-
+        joblib.dump(value=iFor_list, filename=os.path.join(_MODEL_DIR, filename))
+ 
         return(filename)
     except Exception as err:
         # TODO: Implement logger
@@ -135,7 +134,7 @@ def grid_eval(df, subsample_list = [4096, 2048, 1024, 512],
         W = df.iloc[my_indexes[:df_len]]
         
         results_dic = {}
-        ## Evaluation with stored models as external files(pickle format)
+        ## Evaluation with stored models as external files(joblib format)
         for i, subsample_size in enumerate(subsample_list):
             results_dic_t = {}
             for j, trees in enumerate(trees_list):
@@ -143,16 +142,14 @@ def grid_eval(df, subsample_list = [4096, 2048, 1024, 512],
                 # TODO: Implement logger
                 if verbose:
                     print('Reading ' + filename)
-                with open(os.path.join(_MODEL_DIR, filename), 'rb') as infile:
-                    iFor_list = pickle.load(infile)
+                iFor_list = joblib.load(os.path.join(_MODEL_DIR, filename))
                 results_dic_t[str(trees)] = model_eval_fun(W, iFor_list, verbose)
             results_dic[str(subsample_size)] = results_dic_t
         
         filename_results = 'iJungle_light_results_overhead.pkl'
         
         results = pd.DataFrame(results_dic)
-        with open(os.path.join(_MODEL_DIR, filename_results), 'wb') as outfile:
-            pickle.dump(results, outfile)
+        joblib.dump(value=results, filename=os.path.join(_MODEL_DIR, filename_results))
         return(results)
     except Exception as err:
         # TODO: Implement logger
@@ -165,8 +162,7 @@ def get_grid_eval_results(verbose = True):
     if os.path.exists(picklename):
         if verbose:
             print("Reading ", picklename)
-        with open(picklename, 'rb') as pickle_in:
-            results = pickle.load(pickle_in)
+        results = joblib.load(picklename)
         return(results)
     else:
         raise Exception("grid_eval has not have been executed")
@@ -211,9 +207,7 @@ def best_iforest(results, verbose=True):
     picklename = os.path.join(_MODEL_DIR,'iJungle_light_' + str(trees) + '_' + str(subsample_size) + '.pkl')
     if verbose:
         print('Reading ' + picklename)
-        
-    with open(picklename,"rb") as pickle_in:  
-        iFor_list = pickle.load(pickle_in)
+    iFor_list = joblib.load(picklename)    
 
     model = iFor_list[best_iF_k]
     if verbose: