df.append depreciated in newer version of pandas--switch to list?

[audreydrotos]

Describe the bug
In the Jupyter notebook, the cell that predicts pupil size for the whole video has these lines:

        # Append to list instead of DataFrame
        data.append({
            'frameN': int(i+1),
            'pupilSize': np.sum(morphedMask)/255,
            'pupCntr_x': centroid[1],
            'pupCntr_y': centroid[0],
            'eyeProb': eyeProbability,
            'blinkProb': blinkProbability
        })

However, df.append is depreciated in newer versions of pandas so this throws an error. Instead, this information could be added to a list, then a dataframe can be created from the list? Or the docker/readme could be updated to reflect that the older version of pandas is needed.

To Reproduce
Steps to reproduce the behavior: run the pre-trained model on the jupyter notebook.

Expected behavior
An error that append does not work with data frames.

Fix
The following will fix the cell:

cap = cv.VideoCapture(VIDEOPATH)
numFrames = int(cap.get(cv.CAP_PROP_FRAME_COUNT))

# df = pd.DataFrame(columns=['frameN','pupilSize','pupCntr_x','pupCntr_y','eyeProb','blinkProb'])

# Create empty list to collect data
data = []

try:
    for i in range(numFrames):
        _, frame = cap.read()
        frame = cv.cvtColor(frame, cv.COLOR_BGR2GRAY)
        
        croppedFrame = cropResize(frame)
        if INVERTIMAGE:
            croppedFrame = cv.bitwise_not(croppedFrame)
        
        networkInput = croppedFrame.astype(np.float32) / 255.0
        networkInput = networkInput[None, :, :, None]
        mask, info = model(networkInput)
        prediction = mask[0,:,:,0]
        morphedMask, centroid = morphProcessing(prediction)
        
        # Extract infos
        eyeProbability = info[0,0]
        blinkProbability = info[0,1]
        
        # Append to list instead of DataFrame
        data.append({
            'frameN': int(i+1),
            'pupilSize': np.sum(morphedMask)/255,
            'pupCntr_x': centroid[1],
            'pupCntr_y': centroid[0],
            'eyeProb': eyeProbability,
            'blinkProb': blinkProbability
        })
        
        if (i!=0) & (i%400==0) :
            print('Processing frames... ({}/{})'.format(i,numFrames))
finally:
    cap.release()
    # winsound.Beep(BEEP_FREQ, BEEP_DURATION_MS)

# Create DataFrame from list (after loop completes)
df = pd.DataFrame(data)