Merge pull request TreB1eN#8 from tomasic/feat/cps/log-generator

feat/cps/log generator

.gitignore

@@ -136,3 +136,7 @@ dmypy.json
 
 # npm stuff
 node_modules/
+
+# ingnore channel log files
+
+autocoder/cps/logs/*.csv

autocoder/cps/README.md

@@ -30,12 +30,12 @@ pip install -r requirements.txt
 Generate Interface
 
 ```
-usage: generate.py [-h] [-c CONFIG_FILE] [-o OUTPUT_FILE] [-r SEED] [-s SIZE] [-p SHOW_PLOT]
+usage: generate.py [-h] [-c CONFIG_FILE] [-o OUTPUT_PATH] [-p SHOW_PLOT]
 
 optional arguments:
   -h, --help            show this help message and exit
   -c CONFIG_FILE, --config CONFIG_FILE set configuration file yaml
-  -o OUTPUT_FILE, --output OUTPUT_FILE set output channel log file
+  -o OUTPUT_PATH, --output OUTPUT_PATH set output channel log path
   -p SHOW_PLOT, --plot SHOW_PLOT show the distribution histogram plot
 ```
 

autocoder/cps/generate.py

@@ -5,14 +5,20 @@
 import time
 import random
 import argparse
+import os
 
 
-def generate(output_file, show_plot):
-    loc, scale, size = 20, 5, 1000  # mean , standard deviation and size
-    measurements = np.random.normal(loc, scale, size)
+def generate(output_path, show_plot, name, distribution, interval):
+    measurements = handle_distribution(**distribution)
+    intervals = handle_distribution(**interval)
 
-    save_to_csv(measurements, output_file)
+    log_array = np.hstack((intervals, measurements))
+
+    output_file = os.path.join(output_path, name + "_log.csv")
+    save_to_csv(log_array, output_file)
     if (show_plot):
+        scale = distribution["scale"]
+        loc = distribution["loc"]
         import matplotlib.pyplot as plt
 
         count, bins, ignored = plt.hist(measurements, 30, density=True)
@@ -22,16 +28,33 @@ def generate(output_file, show_plot):
         plt.show()
 
 
+def handle_distribution(distribution_type, arguments):
+    if hasattr(np.random, distribution_type):
+        res = (getattr(np.random, distribution_type)(**arguments))
+
+        return res.reshape(arguments['size'], 1)
+
+
+def create_dirs(filepath):
+    if not os.path.exists(filepath):
+        try:
+            os.makedirs(filepath)
+        except OSError as exc:  # Guard against race condition
+            raise
+
+
 def save_to_csv(measurements, output_file):
-    np.savetxt(output_file, measurements, delimiter=",")
+    fmt = ",".join(["%d"] + ["%10.6e"] * (measurements.shape[1]-1))
+    np.savetxt(output_file, measurements, delimiter=",",
+               fmt=fmt, header='Interval, Measurement')
 
 
 def parse_arguments():
     parser = argparse.ArgumentParser()
     parser.add_argument("-c", "--config", type=str, dest='config_file', help="set configuration file yaml",
                         default="autocoder/cps/generator.yaml")
-    parser.add_argument("-o", "--output", type=str, dest='output_file',
-                        help="set output channel log file", default="channel_log.csv")
+    parser.add_argument("-o", "--output", type=str, dest='output_path',
+                        help="set output channel log path", default=os.path.join(os.path.dirname(os.path.abspath(__file__)), "logs"))
     parser.add_argument("-p", "--plot", type=bool, dest='show_plot',
                         help="show the distribution histogram plot", default=False)
 
@@ -40,15 +63,17 @@ def parse_arguments():
 
 def main():
     args = parse_arguments()
-    seed = 0
+    create_dirs(args.output_path)
     with open(args.config_file, 'r') as stream:
         config_info = yaml.load_all(stream, Loader=yaml.Loader)
         for config in config_info:
-            print(config["run"]["generator"]["random_seed"])
             seed = config["run"]["generator"]["random_seed"]
+            np.random.seed(seed)
+            channels = config["run"]["generator"]["facility"]["channels"]
 
-    np.random.seed(seed)
-    generate(args.output_file, args.show_plot)
+            for channel in channels:
+                generate(args.output_path, args.show_plot,
+                         **channel["channel"])
 
 
 if __name__ == "__main__":

autocoder/cps/generator.yaml

@@ -7,31 +7,31 @@ run: # the name of the scenario
         - channel: # each channel generates measurements with a different distribution
             name: facility_temperature
             distribution:
-              -distribution:
-                type: normal # using the same name for distribution and params as in numpy
-                arguments: # normal(loc=0.0, scale=1.0, size=None)
-                  loc: 20.0
-                  scale: 5.0
-                  size: 1000
+              distribution_type: normal # using the same name for distribution and params as in numpy
+              arguments: # normal(loc=0.0, scale=1.0, size=None)
+                loc: 20.0
+                scale: 5.0
+                size: 1000
             interval:
-              -distribution:
-                type: choice
-                options: [1, 2, 3, 4, 5, 6, 7, 8]
-                probabilities: [0.1, 0.1, 0.1, 0.2, 0.2, 0.1, 0.1, 0.1]
+              distribution_type: choice
+              arguments:
+                a: [1, 2, 3, 4, 5, 6, 7, 8]
+                p: [0.1, 0.1, 0.1, 0.2, 0.2, 0.1, 0.1, 0.1]
+                size: 1000
         - channel: # read data is type converted (Date objects)
             name: facility_humidity
             distribution:
-              -distribution:
-                type: normal
-                arguments: #normal(loc=0.0, scale=1.0, size=None)
-                  loc: 45.0
-                  scale: 2.0
-                  size: 1000
+              distribution_type: normal
+              arguments: #normal(loc=0.0, scale=1.0, size=None)
+                loc: 45.0
+                scale: 2.0
+                size: 1000
             interval:
-              -distribution:
-                type: choice
-                options: [1, 2, 3, 4, 5, 6, 7, 8]
-                probabilities: [0.1, 0.1, 0.1, 0.2, 0.2, 0.1, 0.1, 0.1]
+              distribution_type: choice
+              arguments:
+                a: [1, 2, 3, 4, 5, 6, 7, 8]
+                p: [0.1, 0.1, 0.1, 0.2, 0.2, 0.1, 0.1, 0.1]
+                size: 1000
 
 test: # the test scenario is the same as the run scenario
   generator: