fmda/moisture_rnn_test.py

   1 import numpy as np
   2 from moisture_rnn import staircase_spatial
   3
   4
   5 # Test Function Notes:
   6 # ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
   7
   8 # Testing for a single Functions:
   9 # 1. Helper function to create data objects, eg X & y
  10 # 2. Function to print X & y in human readable way
  11 # 3. Call target function eg staircase_spatial to create XX, yy, n_seqs
  12 # 4. Print previous output in human readable way (as similar to images in Overleaf FMDA w Recurrent Notebooks - Example 4)
  13 # 5. Compare to expectations* (eg see Overleaf FMDA w Recurrent Notebooks - Example 4) human or code
  14
  15
  16 # DRAFT UTIL FOR PRINTING:
  17 # rounded_padded_X0 = round_and_pad(X[0,:,:])
  18 # rounded_padded_X1 = round_and_pad(X[params['batch_size']-1,:,:])
  19 # sample_0_str = f"[{' '.join(rounded_padded_X0[0])}]"
  20 # sample_1_str = f"[{' '.join(rounded_padded_X1[0])}]"
  21 # row_length = len(f"{sample_0_str}  ...  {sample_1_str}")
  22
  23
  24 # # Manually adjust padding to match the row length
  25 # total_header_length = len("Sample 0") + len(f"Sample {params['batch_size']-1}") + len("...")
  26 # spaces_needed = row_length - total_header_length
  27 # padding = " " * (spaces_needed // 2)
  28
  29 # print("Batch 0:")
  30 # print(f"Sample 0{padding}...{padding}Sample {params['batch_size']-1}")
  31
  32 # for row0, row1 in zip(rounded_padded_X0, rounded_padded_X1):
  33 #     print(f"[{' '.join(row0)}]  ...  [{' '.join(row1)}]")
  34
  35
  36
  37 def staircase_spatial_test(total_time_steps, features, batch_size, timesteps, n_locs = 10):
  38     """
  39     Parameters:
  40     -----------
  41
  42
  43
  44     Run staircase_spatial with:
  45     -----------
  46     X : list of numpy arrays of length n_locs
  47         A list where each element is a numpy array containing features for a specific location. The shape of each array is `(total_time_steps, features)`.
  48
  49     y : list of numpy arrays of length n_locs
  50         A list where each element is a numpy array containing the target values for a specific location. The shape of each array is `(total_time_steps,)`.
  51
  52
  53     Returns:
  54     -----------
  55
  56     """
  57
  58
  59     #
  60     print("staircase_spatial_test inputs")
  61     print(f"total_time_steps: {total_time_steps}")
  62     print(f"features: {features}")
  63     print(f"batch_size: {batch_size}")
  64     print(f"timesteps: {timesteps}")
  65     print(f"n_locs: {n_locs}")
  66
  67
  68     # Create test arrays
  69     X = []
  70     y = []
  71
  72     for i in range(0, n_locs):
  73         Xi = np.arange(i, i+total_time_steps)
  74         yi = np.arange(i, i+total_time_steps)
  75         X.append(Xi)
  76         y.append(yi)
  77
  78
  79     print(X)
  80     print(y)
  81
  82
  83
  84
  85