Merge branch 'feat-updated-plotter' into 'master'

Updated embody plotter to handle testing more fluently

See merge request tithei/pet-rating!7

scripts/plot_image.py

@@ -2,12 +2,14 @@
 Script for testing embody drawing results from onni.utu.fi
 
 Install requirements:
+pip install tqdm
 pip install numpy
 pip install matplotlib
 
 Usage:
 Export data from onni.utu.fi and after that run the script in the same folder
-by passing exported_file as a parameter to the script
+by passing exported_file as a parameter to the script (NOTE that you must have
+the default background image <dummy_600.png> also in the same path as the script).
 
 $ python plot_image.py <exported_file>.csv
 
@@ -15,22 +17,21 @@ Program prints header of the file, from which you must select column where the
 image data is. After you have selected the right column, program prints the 
 drawing results from embody answers.
 
-If you want the program to draw default embody image to the background, then 
-you must put a copy of the 'dummy_6000.png' -file (this is the same that is used 
-in onni.utu.fu) to the same path as the script.
-
 '''
 
 import copy
-import numpy as np
-import matplotlib.pyplot as plt
 import csv
 import sys
+import subprocess
+
+from tqdm import tqdm
+import numpy as np
+import matplotlib.pyplot as plt
 
 csv.field_size_limit(sys.maxsize)
 
 
-def show_images(images, cols=1, titles=None):
+def show_one_image_per_answer(images, cols=1, titles=None):
     """Display a list of images in a single figure with matplotlib.
     Parameters
     ---------
@@ -44,7 +45,6 @@ def show_images(images, cols=1, titles=None):
     """
 
     # default embody image for the background
-
     try:
         background = True
         default_img = plt.imread("./dummy_600.png")
@@ -57,12 +57,14 @@ def show_images(images, cols=1, titles=None):
     my_cmap.set_bad(alpha=0)
 
     assert((titles is None) or (len(images) == len(titles)))
+
     n_images = len(images)
+
     if titles is None:
         titles = ['Image (%d)' % i for i in range(1, n_images + 1)]
     fig = plt.figure()
-    for n, (image, title) in enumerate(zip(images, titles)):
 
+    for n, (image, title) in enumerate(zip(images, titles)):
         a = fig.add_subplot(cols, np.ceil(n_images/float(cols)), n + 1)
 
         # draw points from users answers
@@ -78,34 +80,89 @@ def show_images(images, cols=1, titles=None):
     plt.show()
 
 
+def show_images(images):
+    """Display all data from list of images in a single figure with matplotlib.
+    Parameters
+    ---------
+    images: List of np.arrays compatible with plt.imshow.
+    """
+
+    # default embody image for the background
+
+    try:
+        background = True
+        default_img = plt.imread("./dummy_600.png")
+    except FileNotFoundError:
+        background = False
+
+    # get a copy of the gray color map
+    my_cmap = copy.copy(plt.cm.get_cmap('gray'))
+    # set how the colormap handles 'bad' values
+    my_cmap.set_bad(alpha=0)
+    n_images = len(images)
+    fig = plt.figure()
+    all_images = np.zeros(shape=(602,207)) 
+    for n, (image, title) in enumerate(zip(images, titles)):
+        all_images += image
+
+    plt.imshow(all_images, cmap=my_cmap)
+
+    if background:
+        plt.imshow(default_img, extent=[0, 200, 600, 0], alpha=0.33)
+
+    fig.set_size_inches(np.array(fig.get_size_inches()) * n_images)
+    plt.show()
+
+
 if __name__ == '__main__':
 
+    filename = sys.argv[1]
+
+    skipped = 0
+    empty = 0
+    rows = 0
     images = []
     titles = []
 
-    # filename = 'experiment_1_2020-05-20.csv'
-    filename = sys.argv[1]
+    # count the file length
+    output = subprocess.check_output(f"wc -l {filename}", shell=True, stderr=subprocess.STDOUT)
+    wc = output.decode("utf-8").split(" ")[0]
 
     with open(filename, 'r+') as csvfile:
-        for row_no, row in enumerate(csv.reader(csvfile, delimiter=';')):
+        csv_rows = csv.reader(csvfile, delimiter=';')
+
+        for row_no, row in enumerate(tqdm(csv_rows, total=int(wc))):
+            rows += 1
 
             # parse header
             if row_no == 0:
                 for column, title in enumerate(row):
-                    print("Column (no. {}): {}".format(column, title))
+                    if "embody_question" in title:
+                        print("Column (ID {}): {}".format(column, title))
 
-                print('Enter the column number which has image data:')
+                print('\nEnter the column ID from which you want to see image data:')
                 x = int(input())
+                print("\n...processing images...\n")
                 continue
 
             try:
+                # skip empty rows (no answer at all)
+                if not row[x]:
+                    skipped += 1
+                    continue
+
                 np_array = np.array(eval(row[x]))
+
+                # skip empty answers (user hasn't drawed on the picture at all)
+                if np.all((np_array == 0)):
+                    empty += 1
+                    continue
             except NameError:
                 print(
-                    "Column didn't contain image data. Try again with different column number.")
-            except SyntaxError:
+                    "Column didn't contain image data. Try again with different column ID.")
+            except SyntaxError as err:
                 continue
-            except IndexError:
+            except IndexError as err:
                 continue
 
             np_array = np.transpose(np_array)
@@ -114,4 +171,12 @@ if __name__ == '__main__':
             # add id of the answerer to the image
             titles.append(row[0])
 
-    show_images(images, titles=titles)
+        print(f"\nExperiment started {rows} times")
+        print(f"from which users has skipped {skipped + empty} times this question\n")
+
+        # show all answers from one column in one image
+        show_images(images)
+
+        # show all answer from one column in own images
+        # OBS: this works decently only with small amount of answers
+        # show_one_image_per_answer(images, titles=titles)