I have a python script that uses OpenCV, code works in python 2.7 but not python 3.7 and I am not sure why

Here is one way to avoid the issue of the properly accessing the contour item in the return values in findContours() in Python/OpenCV.

There are two different possible number of return values from findContours, depending upon which version of OpenCV you are using. So to access the position where contours can be found in the list of return values:

Replace

contours,hierarchy = cv2.findContours(tmp_im, cv2.RETR_LIST, cv2.CHAIN_APPROX_SIMPLE)

contours,hierarchy = cv2.findContours(tmp_im, cv2.RETR_LIST, cv2.CHAIN_APPROX_SIMPLE)
contours = contours[0] if len(contours) == 2 else contours[1]

When you were calling cv2.boundingRect(cnts) the value of cnts was wrong. I read around and this is because the version you were using is outdated. It appears more information is now returned from cv2.findContours(...), so you simply need to extract it.

def process_component(): Changes

contours,hierarchy = cv2.findContours(tmp_im, cv2.RETR_LIST, cv2.CHAIN_APPROX_SIMPLE)
cnts = contours[0]

Second, when you were drawing the circle using cv2.circle(...) your were passing it a tuple of floats. cv2.circle(...) cannot use floats apparently, so convert each value in the tuple to an int first. You may want to round these floats before converting to an int.

def process_component(): Changes

center = ( int(x+w/2), int(y+h/2) )
# Or use the rounded version below.
# center = ( int(round(x+w/2)), int(round(y+h/2)) )
cv2.circle(tmp_im, center, 3, (255,0,0), 2, 8, 0)

Here is your script with those implementations.

import os
import cv2
import numpy as np
from matplotlib import pyplot as plt


def distance_to_camera(knownWidth, focalLength, perWidth):
    """ 
    Find distance between camera and object with OpenCV.
    Return the distance from the maker (object?) to the camera.
    """ 
    return (knownWidth * focalLength) / perWidth



def imshow_components(labels):
    """ Display labels: For debugging purposes"""
    # Map component labels to hue val
    label_hue = np.uint8(179*labels/np.max(labels))
    blank_ch = 255*np.ones_like(label_hue)
    labeled_img = cv2.merge([label_hue, blank_ch, blank_ch])

    # cvt to BGR for display
    labeled_img = cv2.cvtColor(labeled_img, cv2.COLOR_HSV2BGR)

    # set bg label to black
    labeled_img[label_hue==0] = 0
    #labeled_img[labels==0] = 0

    cv2.imshow('labeled.png', labeled_img)
    cv2.waitKey(1000)
    cv2.imwrite('labeled_img.png',labeled_img)
    #cv2.imwrite('label_hue.png',label_hue)



def process_component(labels):
    """ 
    Describe what happens here.
    width of ____ is ___.
    Find contours of ___ for ___.
    etc. Write what happens in this __docstr__
    return the ______.
    """
    
    width = np.zeros(np.max(labels)) 
    
    for i in range(1,np.max(labels)+1):
        
        tmp_im= labels.copy()
        tmp_im[:] = 0
        tmp_im[labels==i] = 255
        
        file="imlabel_%d.png"%(i, )
        cv2.imwrite(file,tmp_im)
        tmp_im = tmp_im.astype(np.uint8)

        ##########################
        # Here was the first issue. https://docs.opencv.org/3.1.0/dd/d49/tutorial_py_contour_features.html
        contours,hierarchy = cv2.findContours(tmp_im, cv2.RETR_LIST, cv2.CHAIN_APPROX_SIMPLE)
        cnts = contours[0]

        x,y,w,h = cv2.boundingRect(cnts)
        ##########################

        width[i-1] = w
        tmp_im=cv2.rectangle(tmp_im,(x,y),(x+w,y+h),(255,0,0),2)
        

        ##########################
        # HERE IS THE NEXT ISSUE. 
        # Fixed it. The center tuple (x, y) cannot be of type floats apparently. Convert each value to an int after division.
        # center = center of the bounding box
        center=(int(x+w/2),int(y+h/2))

        cv2.circle(tmp_im, center, 3, (255,0,0), 2, 8, 0)
        ##########################

        cv2.imshow(file, tmp_im)
        cv2.waitKey(1000)
        
        
    return width




def main():
    """ Call this function to run the program. Put some detail in here. """
    
    # Get the path to the filename. (Im on windows right now)
    filename =  "many_blob.png"
    path = os.path.join(os.getcwd(), filename)
    
    
    # What are these?
    KNOWN_WIDTH = 8
    focalLength = 545
    
    # Read the path as an cv2 image and show it. Wait 1 sec.  
    img = cv2.imread(path)
    cv2.imshow("img", img)
    cv2.waitKey(1000)
    
    # Do something else here. I dont know what, sorry idk about cv2
    image = cv2.threshold(img, 127, 255, cv2.THRESH_BINARY)[1]  # ensure binary 127
    cv2.imshow("image", image)
    cv2.waitKey(1000)
    
    # Do something with parameters that will affect the image.
    kernel = np.ones((5,5),np.uint8)
    erosion = cv2.erode(image,kernel,iterations = 5)
    dilate=cv2.dilate(erosion,kernel,iterations = 5)
    edged = cv2.Canny(dilate, 0, 128)
    
    # Show the affected image.
    cv2.imshow("edged", edged)
    cv2.waitKey(1000)
    
    # idk.
    connectivity=8
    num_labels,labels_im = cv2.connectedComponents(edged,connectivity)
    
    # Do something here.
    width_array=process_component(labels_im)
    imshow_components(labels_im)
    cv2.imwrite('labels_img.png',labels_im)
    
    
    
    # do some more things here.
    for i in range(1,np.max(labels_im)+1):
        w=width_array[i-1]
        #marker = find_marker(image)
        dist_cm = distance_to_camera(KNOWN_WIDTH, focalLength, w)
        print("distance en cm = %d",dist_cm)




        
        
        
    
        
if __name__ == '__main__':
    """ If this file is executed, run the main function. """
    
    main()