s0urceBOT/dataLoad.py

#!/usr/bin/env python3

from PIL import Image
from pprint import pprint

import sys, time, os, requests, random, json, argparse

parser = argparse.ArgumentParser(description='Unknown program that breaks in strange ways.')
parser.add_argument("--download", help="Download Images", action="store_true")
parser.add_argument("JSON", type=str, nargs="?", help="Filename to save results", default="test.js")
args = parser.parse_args()
pprint(args)

# Should we spend the time to download image, and process it? (True = Yes, False = No)
# DOWNLOAD = False
DOWNLOAD = args.download

# Should we add the JSON in a file? (True is filename, False = do not do)
# JSONME = 'test.js'
JSONME = args.JSON

# NOTE: To begin the insert of the JSONIFIED image and word its
# // T
# A JS comment with a uppercase T
# To stop its
# // t
# A JS comment with a lowercase t

# httpbin.org/headers
sess = requests.Session()
head = {'User-Agent': 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_10_1) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/39.0.2171.95 Safari/537.36'}
sess.headers.update(head)

ON = 'X'
OFF = '.'
DIR = 'data'

INTENSITY = 75

im = ''
pix = [0, 0]
size = [0, 0]
pal = ''

if not os.path.exists('in'):
    os.mkdir('in')


def download(howhard, index):
    global sess
    """
    Download an image based upon how hard it is.

    Note:  This sucks, because it overwrites images.  
    Why not save the images and how hard?
    Quit running your script on floppy disks!
    """
    r = sess.get(f'http://s0urce.io/client/img/word/{howhard}/{index}')
    if(r.status_code == 200):
        with open(f'in/{index}.png', 'wb') as f:
            f.write(r.content)
        return f'{index}.png ' + str(r.status_code)
    return r.status_code

def img_point(x, y):
        global pix
        return pix[x, y]

def img_avg(x, y):
    global im, pal
    rgb = img_point(x,y)
    if(im.mode == 'P'):
        rgb = pal[rgb*3:(rgb+1)*3]

    if(im.mode == 'I'):
        return rgb >> 8

    return int( ( rgb[0] + rgb[1] + rgb[2] ) / 3 )

def is_set(x, y):
    global INTENSITY
    avg = img_avg(x,y)
    return (avg < INTENSITY)

def scan_img():
    global sx, sy, ex, ey, total
    total = 0
    sx = size[0]; ex = 0
    sy = size[1]; ey = 0

    for y in range( 0,size[1] ):
        for x in range( 0,size[0] ):
            pnt_is = is_set(x,y)
            if (pnt_is):
                total += 1
                if x < sx:
                    sx = x
                if x > ex:
                    ex = x
                if y < sy:
                    sy = y
                if y > ey:
                    ey = y

    #print (sx,ex,sy,ey)
    # give us a little border to work with
    if sx > 0:
        sx -= 1
    if ex < size[0]:
        ex += 1

    if sy > 0:
        sy -= 1
    if ey < size[1]:
        ey += 1

    #print (sx,ex,sy,ey)
    return(sx,sy,ex,ey)

def save_image():
    result = []
    for y in range(sy,ey):
        s = ''
        for x in range(sx,ex):
            if is_set(x,y):
                s += ON
            else:
                s += OFF
        result.append(s)

    return result

def run(VAL):
    for x in range(0, 70):
        fname = f'in/{x}.png'

        try:
            with open(fname, 'r') as f:
                f = f
        except FileNotFoundError:
            print("File does not exist")
            continue

        print(f"Loading: {fname}")
        im = Image.open(fname)
        pix = im.load()
        size = im.size
        print(f"Size: {size[0]} x {size[1]}")

        if(im.mode == 'P'):
            pal = im.getpalette()

        sx = 0
        ex = size[0]
        sy = 0
        ey = size[1]
        total = 0

        scan_img()

        print(f"Chars within ({sx}, {sy}) - ({ex}, {ey}) total {total} pixels")

        img_s = save_image()
        for l in img_s:
            print(l)
        img_s.append(input('Word: '))
        y = x + VAL
        with open(f'{DIR}/{y}.txt', 'w') as f:
            for i_s in img_s:
                f.write(f'{i_s}\n')

        print(f"Image saved to '{DIR}/{y}.txt' in byte string")
        os.remove(f'{fname}')
        print(f"File '{fname}' automatically removed")

if (DOWNLOAD == True):
    print("Downloading s0urce.io Words")
    print("EASY")
    time.sleep(5)
    for e in range(0, 62):
        print(download('e', e))
        # time.sleep(random.randint(10, 15))

    run(0)

    print("MEDIUM")
    # time.sleep(5)
    for m in range(0, 66):
        print(download('m', m))
        # time.sleep(random.randint(10, 15))

    run(62)

    print("HARD")
    # time.sleep(5)
    for h in range(0, 55):
        print(download('h', h))
        # time.sleep(random.randint(10, 15))

    run(128)

# ----------------------------------------------------------------------------------------
# All below was in a seperate dataJS.py file... but now I have fixed it so it's 1 script!

if (JSONME.lower() != 'false'):
    print("Now exporting to JSON")
    print(f"Targeting file: '{JSONME}'")
    time.sleep(5)

    def test(t):
        global DIR
        fname = f'{DIR}/{t}.txt'
        r = []
        try:
            with open(fname, 'r') as f:
                for l in f:
                    r.append(l.strip())
            return r
        except FileNotFoundError:
            return None

    def insertJS(item):
        global JSON
        item = json.dumps(item)
        item = f'{item},'
        r = []
        try:
            with open(f'{JSONME}', 'r') as f:
                for l in f:
                    if l != '':
                        r.append(l.strip('\n'))
                    else:
                        r.append('')
        except FileNotFoundError:
            print(f"File {JSONME} Not Found!")
            sys.exit()
        
        c = 0
        for e in r:
            if('// T' == e):
                temp = r[c+1]
                del r[c+1]
                r.insert(c+1, item)
                r.insert(c+2, temp)
            elif('// t' == e):
                break
            c += 1
        
        with open(f'{JSONME}', 'w') as f:
            for e in r:
                f.write(f'{e}\n')

    for x in range(0, 183):
        te = test(x)
        if(te != None):
            word = te
            insertJS(word)

print("Complete")