added animation video

.gitignore

@@ -3,10 +3,12 @@
 *.mp4
 *.xz
 *.json
+animation_table.csv
 *.gif
 *.zip
 *.png
 Data-*
+Data/
 drive-*
 __pycache__/
 *.py[cod]

app.py

@@ -1,7 +1,10 @@
+
+import os
+import csv
+import json
 import torch
+import numpy as np
 import gradio as gr
-import json
-import os
 
 from phate import PHATEAE
 from funcs.som import ClusterSOM
@@ -18,6 +21,78 @@ reducer10d.load('models/r10d_3.pth')
 cluster_som = ClusterSOM()
 cluster_som.load("models/cluster_som3.pkl")
 
+def map_som2animation(som_value):
+    mapping = {
+        2: 0,  # walk
+        1: 1,  # trot
+        3: 2,  # gallop
+        5: 3,  # idle
+        4: 3,  # other
+        -1:3, #other
+    }
+    
+    return mapping.get(som_value, None)
+
+def deviation_scores(tensor_data, scale=50):
+    if len(tensor_data) < 5:
+        raise ValueError("The input tensor must have at least 5 elements.")
+    
+    # Extract the side values and reference value from the input tensor
+    side_values = tensor_data[-5:-1].numpy()
+    reference_value = tensor_data[-1].item()
+
+    # Calculate the absolute differences between the side values and the reference
+    absolute_differences = np.abs(side_values - reference_value)
+    
+    # Check for zero division
+    if np.sum(absolute_differences) == 0:
+        # All side values are equal to the reference, so their deviation scores are 0
+        return int(reference_value/20*32768), [0, 0, 0, 0]
+
+    # Calculate the deviation scores for each side value
+    scores = absolute_differences * scale
+    
+    # Clip the scores between 0 and 1
+    clipped_scores = np.clip(scores, 0, 1)
+
+    return int(reference_value/20*32768), clipped_scores.tolist()
+
+def process_som_data(data, prediction):
+    processed_data = []
+
+    for i in range(0, len(data)):
+        TS, scores_list = deviation_scores(data[i][0])
+
+        # If TS is missing (None), interpolate it using surrounding values
+        if TS is None:
+            if i > 0 and i < len(data) - 1:
+                prev_TS = processed_data[-1][1]
+                next_TS = deviation_scores(data[i + 1][0])[0]
+                TS = (prev_TS + next_TS) // 2
+            elif i > 0:
+                TS = processed_data[-1][1]  # Use the previous TS value
+            else:
+                TS = 0  # Default to 0 if no surrounding values are available
+
+
+        # Set Gait, State, and Condition
+
+        #0-walk 1-trot 2-gallop 3-idle
+        gait = map_som2animation(prediction[0][0])
+        state = 0
+        condition = 0
+
+        # Calculate Shape, Color, and Danger values
+        shape_values = scores_list
+        color_values = scores_list
+        danger_values = [1 if score == 1 else 0 for score in scores_list]
+
+        # Create a row with the required format
+        row = [gait, TS, state, condition] + shape_values + color_values + danger_values
+        processed_data.append(row)
+
+    return processed_data
+
 # ml inference
 def get_som_mp4(file, slice_select, reducer=reducer10d, cluster=cluster_som):
 
@@ -38,10 +113,23 @@ def get_som_mp4(file, slice_select, reducer=reducer10d, cluster=cluster_som):
     #compute the 10 dimensional embeding vector
     embedding10d = reducer.transform(data)
 
+    # retrieve the prediction and get the animation
+    prediction = cluster_som.predict(embedding10d)
+    processed_data = process_som_data(data,prediction)
+
+    # Write the processed data to a CSV file
+    header = ['Gait', 'TS', 'State', 'Condition', 'Shape1', 'Shape2', 'Shape3', 'Shape4', 'Color1', 'Color2', 'Color3', 'Color4', 'Danger1', 'Danger2', 'Danger3', 'Danger4']
+    with open('animation_table.csv', 'w', newline='') as csvfile:
+        csv_writer = csv.writer(csvfile)
+        csv_writer.writerow(header)
+        csv_writer.writerows(processed_data)
+    
+    os.system('curl -X POST -F "csv_file=@animation_table.csv" https://metric-space.ngrok.io/generate --output animation.mp4')
+
     # prediction = cluster_som.predict(embedding10d)
     fig = cluster.plot_activation_v2(embedding10d, slice_select)
 
-    return fig
+    return fig, 'animation.mp4'
 
 def attach_label_to_json(json_file, label_text):
     # Read the JSON file
@@ -114,7 +202,7 @@ with gr.Blocks(title='Cabasus') as cabasus_sensor:
         slice_slider.change(plot_sensor_data_from_json, inputs=[json_file_box, leg_dropdown, slice_slider], 
                             outputs=[plot_box_leg, plot_slice_leg, get_all_slice, slice_json_box, plot_box_overlay])
         
-        som_create.click(get_som_mp4, inputs=[json_file_box, slice_slider], outputs=[som_figures])
+        som_create.click(get_som_mp4, inputs=[json_file_box, slice_slider], outputs=[som_figures, animation])
         button_label_Add.click(attach_label_to_json, inputs=[slice_json_box, label_name], outputs=[slice_json_label_box])
 
 cabasus_sensor.queue(concurrency_count=2).launch(debug=True)

funcs/convertors.py

@@ -17,11 +17,11 @@ def slice_csv_to_json(input_file, slice_size=64, min_slice_size=16, sample_rate=
 
     gz_columns = [col for col in data.columns if col.startswith("GZ")]
     all_peaks = []
-    upsample_factor = sample_rate
-    combined_smoothed_signals_upsampled = np.zeros(upsample_signal(data[gz_columns[0]].values, upsample_factor).size, dtype=float)
+    # upsample_factor = sample_rate
+    combined_smoothed_signals_upsampled = np.zeros(upsample_signal(data[gz_columns[0]].values, sample_rate).size, dtype=float)
     for gz_col in gz_columns:
         gz_signal = data[gz_col].values
-        upsampled_smoothed_signal, peaks = process_signals(gz_signal, upsample_factor, window_size=window_size)
+        upsampled_smoothed_signal, peaks = process_signals(gz_signal, sample_rate, window_size=window_size)
         all_peaks.append(peaks)
         combined_smoothed_signals_upsampled += upsampled_smoothed_signal
 
@@ -31,7 +31,7 @@ def slice_csv_to_json(input_file, slice_size=64, min_slice_size=16, sample_rate=
     slices = []
     start_index = 0
     for i, precise_slice_point in enumerate(precise_slice_points):
-        end_index = round(precise_slice_point / upsample_factor)
+        end_index = round(precise_slice_point / sample_rate)
         if i == 0:
             start_index = end_index
             continue
@@ -52,7 +52,6 @@ def slice_csv_to_json(input_file, slice_size=64, min_slice_size=16, sample_rate=
 
                 # Compute precise_timestamp and precise_time_diff for each GZ channel individually
                 for j, gz_col in enumerate(gz_columns):
-                    #slice_data[f"{gz_col}_precise_timestamp"] = slices[-1][f"{gz_col}_precise_timestamp"] + all_peaks[j][i] - all_peaks[j][i - 1]
                     slice_data[f"{gz_col}_precise_time_diff"] = all_peaks[j][i] - all_peaks[j][i - 1]
             else:
                 precise_timestamp = data.index.values[start_index]

funcs/processor.py

@@ -10,11 +10,11 @@ def process_data(input_file, slice_size=64, min_slice_size=16, sample_rate=20, w
     # Read the data from the file, including the CRC column
     try:
         if input_file.name is None:
-            return None, None, None, None, None, None, None, None
+            return None, None, None, None, None, None, None, None, None
         data = pd.read_csv(input_file.name, delimiter=";", index_col="NR", usecols=["NR", "TS", "LEG", "GX", "GY", "GZ", "AX", "AY", "AZ", "CRC"])        
     except:
         if input_file is None:
-            return None, None, None, None, None, None, None, None
+            return None, None, None, None, None, None, None, None, None
         data = pd.read_csv(input_file, delimiter=";", index_col="NR", usecols=["NR", "TS", "LEG", "GX", "GY", "GZ", "AX", "AY", "AZ", "CRC"])
 
 
@@ -70,7 +70,7 @@ def process_data(input_file, slice_size=64, min_slice_size=16, sample_rate=20, w
     if not no_significant_change_index.empty:
         # Save the data up to the point where no significant change appears in all channels
         data = data.loc[:no_significant_change_index[0]]
-        return None, None, f'Warning: Significantly shortened > check the recordings', None, None, None, None, None
+        return None, None, f'Warning: Significantly shortened > check the recordings', None, None, None, None, None, None
 
     # Save the resulting DataFrame to a new file
     data.to_csv('output.csv', sep=";", na_rep="NaN", float_format="%.0f")

wetransfer_files_2023-05-04_1807.zip

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:e555b7d207aecddbbd6f0a704801d9c05d53eb62218dc04a9aa9ea2995c9d565
-size 9167229