Updated readmes

binning/README.md

@@ -16,19 +16,19 @@ First put your dataset txt in form of ```image_name,image_count``` in a ```.txt`
 To start the training run
 
  ```
- python train.py -d ./dataset_txt/Train_nwpu.txt
+ python train.py -d ./dataset_txt/Train_nwpu.txt -f multinomial
  ``` 
 
-If you want to get bins on your dataset, change the -d argument to that path itself.
+If you want to get bins on your dataset, change the ```-d``` argument to that path itself. The ```-f``` argument helps choose between the fitness functions (```multinomial,poisson```).
 
 ### Step 3
 
-After training (Step 2) there are best files generated in the ```select_best``` folder. To print out the top two best binning configurations, run the fowllowing code
+After training (Step 2) there are best files generated in the ```select_best``` folder. To print out the top two best binning configurations, run the following code
 
 ```
-python generate_bins.py
+python generate_bins.py -d ./dataset_txt/Train_nwpu.txt -f multinomial
 ```
-
+Here ```-d``` corresponds to dataset and ```-f``` corresponds to fitness function.
 
 The folder structure of the codes in this folder is:
 ```

binning/bayesian_blocks.py

@@ -320,7 +320,7 @@ def bayesian_blocks(t, x=None, sigma=None,
  break
  ind = last[ind - 1]
  change_points = change_points[i_cp:]
- print("bins",edges[change_points])
+ # print("bins",edges[change_points])
  return edges[change_points]
 
 # X_train = [1,2,3,4,4,5,5,5,5,6,7,8,0,0,0,0,2,3,4,4,5,5,5,5,6,7,8,0,0,0,0]

binning/generate_bins.py

@@ -1,15 +1,27 @@
 from load_data import load_data,load_data_patch
 from select_best import best_2
 import os
+import glob
+import argparse
+
+
+parser = argparse.ArgumentParser()
+# Adding optional argument
+parser.add_argument("-d", "--dataset", help = "Give dataset path in txt",type=str,default="./dataset_txt/Train_nwpu.txt")
+parser.add_argument("-f", "--fitness_function", help = "Select fitness function",type=str,default="multinomial")
+
+# Read arguments from command line
+args = parser.parse_args()
+
 test_ratio_arr=[]
 files = glob.glob('./select_best/*')
 for f in files:
  test_ratio_arr.append(f)
 # run this code after running train.py
 # this code prins out top 2 bins for each fitness fucntion 
 # and its corresponding gamma values
-fitness_funct='multinomial'
-t = load_data(path=[r'Val_new.txt']) 
+fitness_funct=args.fitness_function
+t = load_data(path=[args.dataset]) 
 # t = load_data_patch(path=['crops_bl_nwpu.txt']) 
 # test_ratio_arr =[2053,4106,5133]
 best_2(t,fitness_funct=fitness_funct,test_ratio_arr=test_ratio_arr,gammas=[0.1,0.2,0.3,0.4,0.5,0.6,0.7,0.8,0.9])

binning/select_best.py

@@ -65,7 +65,7 @@ def best_2(data,fitness_funct,test_ratio_arr,gammas):
  # if fitness_funct=="poisson":
  # path ="./select_best/poi_mu_sig_"+str(i)+".json"
  path = i
- print("path",i)
+ # print("path",i)
  with open(path, 'r') as fp:
  testt = json.load(fp)
 
@@ -92,6 +92,6 @@ def best_2(data,fitness_funct,test_ratio_arr,gammas):
  # bins for entire data based on the gamma values 
  print("Best Bins for the corresponding values of prior are ")
  for i in top_2:
- print(list(bayesian_blocks(data,fitness=fitness_funct,gamma=i)))
+ print(" Gamma : ",i," bins : ",list(bayesian_blocks(data,fitness=fitness_funct,gamma=i)))