Ocular Ryla Evaluation Scanner and Tabulator for Python 3. Needs imagemagick installed on a system path and the python extension "wand"

rylaeval2014.py

# python eval.py test.pdf 'Top Person' 'Top topic' 'Bottom Person' 'Bottom topic'
import argparse
import os
import shlex
import subprocess
import glob

from wand.image import Image
from wand.display import display
from wand.color import Color

YEAR = '2014'
FOLDER = YEAR + '/*.pdf'

coords = ({'yes': (450, 70, 535, 95), 'no': (535, 70, 620, 95), 'na': (615, 70, 705, 95)},
        {'yes': (450, 95, 535, 120), 'no': (535, 95, 620, 120), 'na': (615, 95, 705, 120)},
        {'yes': (450, 120, 535, 140), 'no': (535, 120, 620, 140), 'na': (615, 120, 705, 140)},
        {'yes': (450, 140, 535, 160), 'no': (535, 140, 620, 160), 'na': (615, 140, 705, 160)},
        {'yes': (450, 160, 535, 180), 'no': (535, 160, 620, 180), 'na': (615, 160, 705, 180)},
        {'yes': (450, 180, 535, 205), 'no': (535, 180, 620, 205), 'na': (615, 180, 705, 205)},
        {'yes': (450, 205, 535, 225), 'no': (535, 205, 620, 225), 'na': (615, 205, 705, 225)},
        {'yes': (450, 225, 535, 250), 'no': (535, 225, 620, 250), 'na': (615, 225, 705, 250)},
        {'yes': (450, 250, 535, 270), 'no': (535, 250, 620, 270), 'na': (615, 250, 705, 270)},
        {'yes': (450, 270, 535, 290), 'no': (535, 270, 620, 290), 'na': (615, 270, 705, 290)})
written_coords = 300
questions = [
            'I learned something new or gained new insights',
            'I learned a new skill that will be useful outside of RYLA',
            'My perceptions were challenged by a new perspective',
            'I feel motivated to serve my community',
            'The speaker presented the topic clearly',
            'The speaker responded well to questions',
            'The presentation was well balanced and engaging',
            'My overall impressions of the speaker was positive',
            'This speaker should be at RYLA every year',
            'This topic should be at RYLA every year'
            ]


def evaluate(path):
    filename = path
    topname = 'blank'
    bottomname = 'blank'
    top_topic = 'blank'
    btm_topic = 'blank'
    
    temp_name = path.lstrip(YEAR + '/').rstrip('.pdf')
    split = temp_name.split(' - ')
    topname = split[0]
    top_topic = split[1]
    if len(split) == 4:
        bottomname = split[2]
        btm_topic = split[3]
    

    res = 50
    factor = res/100
    white = Color('#fff')
    save_path = '/home/croxis/Dropbox/Ryla Results/'

    default_white = []
    top_questions = []
    btm_questions = []


    def scan_a(img, coord):
        white_count = 0
        with img.clone() as a:
            a.crop(int(coord[0] * factor), int(coord[1] * factor), int(coord[2] * factor), int(coord[3] * factor))
            for row in a:
                for col in row:
                    if col.red == 1.0 and col.blue == 1.0 and col.green == 1.0:
                        white_count += 1
                    #if col.alpha_int8 != 0:
                    #    white_count += 1
        return white_count


    def split_page(page):
        top_img = page.clone()
        top_img.crop(0, int(145 * factor), int(page.width), int(590 * factor))
        #top_img.save(filename='a.png')
        btm_img = page.clone()
        #btm_img.crop(0, int(613 * factor), int(page.width), int(1057 * factor))
        btm_img.crop(0, int(613 * factor), int(page.width), page.height)
        #btm_img.save(filename='b.png')
        return top_img, btm_img


    def tabulate_question(page, i, results):
        result_yes = scan_a(page, coords[i]['yes'])
        result_no = scan_a(page, coords[i]['no'])
        result_na = scan_a(page, coords[i]['na'])
        
        changeyes = abs(result_yes - default_white[i]['yes'])/float(default_white[i]['yes'])
        changeno = abs(result_no - default_white[i]['no'])/float(default_white[i]['no'])
        changena = abs(result_na - default_white[i]['na'])/float(default_white[i]['na'])
        
        #print "Percent change:", changeyes, changeno, changena
        
        if changeyes > changeno and changeyes > changena:
            #print "CONCLUSION: YES"
            results[i]['yes'] += 1
        elif changeno > changeyes and changeno > changena:
            #print "CONCLUSION: NO"
            results[i]['no'] += 1
        elif changena > changeno and changena > changeyes:
            #print "CONCLUSION: N/A"
            results[i]['na'] += 1
        else:
            print("Ok something screwed up. Contact David and yell at him.")
            print(result_yes, result_no, result_na)
            print(changeyes, changeno, changena)


    def generate_comment_page():
        #page = Image(width = 3400, height = 4400, background=white) # 400dpi
        page = Image(width=2250, height=3300, background=white) # 300 dpi
        page.format = 'pdf'
        #blank_img.background_color = white
        #blank_img.alpha_channel = False
        return page


    print('Initializing for ' + topname + ' and ' + bottomname)

    print('Debug for reference questions:', default_white)

    print("Loading survay")
    top_feedback = [generate_comment_page()]
    btm_feedback = [generate_comment_page()]
    top_page_count = 0
    btm_page_count = 0
    top_scan_height = 0
    btm_scan_height = 0
    scan_int = 100
    max_height = 1000
    with Image(filename=filename, resolution=res) as img:
        p = 0
        for image in img.sequence:
            top_img, btm_img = split_page(image)
            print("Scanning page", p, '/', len(img.sequence))
            if p is 0:
                print("Calibrating blank page")
                for i in range(10):
                    top_questions.append({'yes': 0, 'no': 0, 'na': 0})
                    btm_questions.append({'yes': 0, 'no': 0, 'na': 0})
                    default_white.append({'yes': 0, 'no': 0, 'na': 0})

                for i in range(10):
                    print("Calibrating question", i+1)
                    default_white[i]['yes'] = scan_a(top_img, coords[i]['yes'])
                    default_white[i]['no'] = scan_a(top_img, coords[i]['no'])
                    default_white[i]['na'] = scan_a(top_img, coords[i]['na'])
            else:
                for i in range(10):
                    #print("Scanning page", p, '/', len(img.sequence), "question", i+1)
                    tabulate_question(top_img, i, top_questions)
                    tabulate_question(btm_img, i, btm_questions)
            p += 1

    print('')
    print("RESULTS")
    top_final = {'yes': 0, 'no': 0, 'na': 0}
    bottom_final = {'yes': 0, 'no': 0, 'na': 0}
    for i in range(10):
        print("Top Question", i+1, top_questions[i])
        print("Btm Question", i+1, btm_questions[i])
        top_final['yes'] += top_questions[i]['yes']
        top_final['no'] += top_questions[i]['no']
        top_final['na'] += top_questions[i]['na']
        bottom_final['yes'] += btm_questions[i]['yes']
        bottom_final['no'] += btm_questions[i]['no']
        bottom_final['na'] += btm_questions[i]['na']


    print('')
    print('Compiling comments')
    print('Reloading surveys at full resolution')
    top_feedback = [generate_comment_page()]
    btm_feedback = [generate_comment_page()]
    top_page_count = 0
    btm_page_count = 0
    top_scan_height = 0
    btm_scan_height = 0
    scan_int = 100
    #scan_int = 200
    max_height = 1000
    res=300
    factor = res/100

    #TODO: Write tests that scans line below main righting and bottom of section to see
    #if anyoen wrote there
    #If so expand selection
    with Image(filename=filename, resolution=res) as img:
        p = 0
        top_extra = 0
        btm_extra = 0
        for image in img.sequence:
            print('Processing page', p)
            if p is 0:
                p += 1
                continue  # Skip calibration page
            top_img, btm_img = split_page(image)

            nonwhite = 0
            for col in top_img[int(395*factor)]:
                if col.red < 1.0 and col.blue < 1.0 and col.green < 1.0:
                    nonwhite += 1
            if nonwhite/top_img.width > 0.2:
                top_extra = int(8 * factor)
            nonwhite = 0
            for col in top_img[int(432*factor)]:
                if col.red < 1.0 and col.blue < 1.0 and col.green < 1.0:
                    nonwhite += 1
            if nonwhite/top_img.width > 0.2:
                top_extra = int(44 * factor)
            nonwhite = 0
            for col in btm_img[int(395*factor)]:
                if col.red < 1.0 and col.blue < 1.0 and col.green < 1.0:
                    nonwhite += 1
            if nonwhite/btm_img.width > 0.2:
                btm_extra = int(8 * factor)
            nonwhite = 0
            for col in btm_img[int(432*factor)]:
                if col.red < 1.0 and col.blue < 1.0 and col.green < 1.0:
                    nonwhite += 1
            if nonwhite/btm_img.width > 0.2:
                btm_extra = int(44 * factor)

            top_img.crop(0, int(written_coords * factor), top_img.width, int(((written_coords + scan_int) * factor + top_extra)))
            btm_img.crop(0, int(written_coords * factor), btm_img.width, int(((written_coords + scan_int) * factor + btm_extra)))
            
            
            if top_scan_height + (scan_int * factor) + top_extra > max_height * factor:
                print('Generating top review page', top_page_count + 2)
                top_scan_height = 0
                top_feedback.append(generate_comment_page())
                top_page_count += 1
            if btm_scan_height + (scan_int * factor) + btm_extra > max_height * factor:
                print('Generating btm review page', btm_page_count + 2)
                btm_scan_height = 0
                btm_feedback.append(generate_comment_page())
                btm_page_count += 1
            top_feedback[top_page_count].composite(top_img, 0, top_scan_height + top_extra)
            btm_feedback[btm_page_count].composite(btm_img, 0, btm_scan_height + btm_extra)
            top_img.close()
            btm_img.close()
            top_scan_height += int(scan_int * factor) + top_extra + 10
            btm_scan_height += int(scan_int * factor) + btm_extra + 10
            top_extra = 0
            btm_extra = 0
            p += 1


    top_cmt = top_feedback[0]
    for i in range(1, len(top_feedback)):
        top_cmt.sequence.append(top_feedback[i].sequence[0])
    btm_cmt = btm_feedback[0]
    for i in range(1, len(btm_feedback)):
        btm_cmt.sequence.append(btm_feedback[i].sequence[0])

    '''for i in range(1, len(top_feedback)):
        top_cmt.sequence.append(top_feedback.sequence[0])
    for i in range(1, len(btm_feedback)):
        btm_cmt.sequence.append(btm_feedback.sequence[0])'''

    # Prep feedback sheet
    print('')
    print('Saving Results')

    with open(save_path + 'results.txt', 'a') as f:
        f.write('----------------------------------------------------------\n')
        f.write(topname + ' - ' + top_topic + '\n')
        for i in range(10):
            f.write(questions[i] + ' - \tYes: ' + str(top_questions[i]['yes']) + '\tNo: ' + str(top_questions[i]['no']) + '\tN/A: ' + str(top_questions[i]['na']) + '\n')
        f.write('----------------------------------------------------------\n')
        f.write(bottomname + ' - ' + btm_topic + '\n')
        for i in range(10):
            f.write(questions[i] + ' - \tYes: ' + str(btm_questions[i]['yes']) + '\tNo: ' + str(btm_questions[i]['no']) + '\tN/A: ' + str(btm_questions[i]['na']) + '\n')



    print('')
    print('Compiling final documents')
    top_cmt.save(filename=save_path + topname + '-' + top_topic + '.pdf')
    btm_cmt.save(filename=save_path + bottomname + '-' + btm_topic + '.pdf')

    print('Finished as ' + topname + '-' + top_topic + '.pdf' + ' and ' + bottomname + '-' + btm_topic + '.pdf')
    print('')
    del top_feedback
    del btm_feedback
    del top_cmt
    del btm_cmt
    import gc
    gc.collect()


if __name__ == '__main__':
    from multiprocessing import Process
    for path in glob.glob(FOLDER):
        p = Process(target=evaluate, args=(path,))
        p.start()
        p.join()
        p.terminate()