From a3324b5e04d5a8bff9f37a64653aadbfee3c4ff0 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ren=C3=A9=20Knaebel?= <rknaebel@uni-potsdam.de>
Date: Sat, 8 Jul 2017 11:53:03 +0200
Subject: [PATCH] add pauls config test (TMP)

---
 main.py | 49 ++++++++++++++++++++++++++++++++++++++++++-------
 1 file changed, 42 insertions(+), 7 deletions(-)

diff --git a/main.py b/main.py
index 5e751f0..9cbc986 100644
--- a/main.py
+++ b/main.py
@@ -19,8 +19,8 @@ parser.add_argument("--test", action="store", dest="test_data",
 # parser.add_argument("--h5data", action="store", dest="h5data",
 #                     default="")
 #
-parser.add_argument("--model", action="store", dest="model",
-                    default="model_x")
+parser.add_argument("--models", action="store", dest="models",
+                    default="models/model_x")
 
 # parser.add_argument("--pred", action="store", dest="pred",
 #                     default="")
@@ -80,6 +80,38 @@ args = parser.parse_args()
 # session = tf.Session(config=config)
 
 
+def main_paul_best():
+    char_dict = dataset.get_character_dict()
+    user_flow_df = dataset.get_user_flow_data(args.train_data)
+
+    param = models.pauls_networks.best_config
+    print(param)
+
+    print("create training dataset")
+    domain_tr, flow_tr, client_tr, server_tr = dataset.create_dataset_from_flows(
+        user_flow_df, char_dict,
+        max_len=args.domain_length,
+        window_size=args.window)
+    client_tr = np_utils.to_categorical(client_tr, 2)
+    server_tr = np_utils.to_categorical(server_tr, 2)
+
+    embedding, model = models.get_models_by_params(param)
+
+    model.compile(optimizer='adam',
+                  loss='categorical_crossentropy',
+                  metrics=['accuracy'])
+
+    model.fit([domain_tr, flow_tr],
+              [client_tr, server_tr],
+              batch_size=args.batch_size,
+              epochs=args.epochs,
+              shuffle=True,
+              validation_split=0.2)
+
+    embedding.save(args.models + "_embd.h5")
+    model.save(args.models + "_clf.h5")
+
+
 def main_hyperband():
     char_dict = dataset.get_character_dict()
     user_flow_df = dataset.get_user_flow_data(args.train_data)
@@ -137,13 +169,13 @@ def main_train():
     client_tr = np_utils.to_categorical(client_tr, 2)
     server_tr = np_utils.to_categorical(server_tr, 2)
 
-    shared_cnn = network.get_embedding(len(char_dict) + 1, args.embedding, args.domain_length,
-                                       args.hidden_char_dims, kernel_size, args.domain_embedding, 0.5)
-    shared_cnn.summary()
+    embedding = network.get_embedding(len(char_dict) + 1, args.embedding, args.domain_length,
+                                      args.hidden_char_dims, kernel_size, args.domain_embedding, 0.5)
+    embedding.summary()
 
     model = network.get_model(cnnDropout, flow_tr.shape[-1], args.domain_embedding,
                               args.window, args.domain_length, filters, kernel_size,
-                              cnnHiddenDims, shared_cnn)
+                              cnnHiddenDims, embedding)
     model.summary()
 
     model.compile(optimizer='adam',
@@ -157,7 +189,8 @@ def main_train():
               shuffle=True,
               validation_split=0.2)
 
-    model.save(args.model)
+    embedding.save(args.models + "_embd.h5")
+    model.save(args.models + "_clf.h5")
 
 
 def main_test():
@@ -206,6 +239,8 @@ def main():
         main_visualization()
     if "score" in args.modes:
         main_score()
+    if "paul" in args.modes:
+        main_paul_best()
 
 
 if __name__ == "__main__":