def check(img):
out = pytesseract.image_to_string(img, lang='eng', config="-psm 5")
if out.__contains__('P'):
return 1
color = 255
for i in range(len_x):
for j in range(len_y):
if data[i, j] == 255:
color -= 1
floodfill(i, j, color)
print(color)
for c in range(color, 255):
(a, b) = np.where(data == c)
if a.shape[0] > 0:
x_min, x_max = np.min(a), np.max(a)
y_min, y_max = np.min(b), np.max(b)
if (x_max - x_min > 5) and (y_max - y_min > 5):
tmp = data[x_min:x_max, y_min:y_max]
out = pytesseract.image_to_string(tmp,
lang='eng',
config="-psm 5")
# rect=cv2.rectangle(img, (x_min, y_min), (x_max, y_max), (0,255,0), 2)
# #Store the small blocks, change the directions!!!
# cv2.imwrite('/Users/chenlingna/Desktop/output/'+str(x_min)+'.JPG', tmp)
if out.__contains__('P') or out.__contains__(
'F') or out.__contains__('E'):
return 1
return 0
def get_ocr_number(img_orig):
"""Return float value from image. -1.0f when OCR failed"""
img_resized = prepareImage(img_orig)
lst = []
config_ocr = '--psm 7 --oem 1 -c tessedit_char_whitelist=0123456789.$£B'
lst.append(
pytesseract.image_to_string(img_resized, 'eng', config=config_ocr).
strip().replace('$', '').replace('£', '').replace('B', ''))
try:
return float(lst[-1])
except ValueError:
images = [img_orig, img_resized] # , img_min, img_mod, img_med, img_sharp]
i = 0
while i < 2:
j = 0
while j < len(images):
lst.append(
pytesseract.image_to_string(images[j], 'eng', config=config_ocr).
strip().replace('$', '').replace('£', '').replace('B', ''))
j += 1
config_ocr = '--psm 8 --oem 1 -c tessedit_char_whitelist=0123456789.$£B'
i += 1
log.debug(lst)
for element in lst:
try:
return float(element)
except ValueError:
pass
return -1.0
def check_exam(request, course_id, task_id):
if request.method == 'POST':
path = 'media/diploma.pdf'
fname = os.path.splitext(os.path.basename(path))[0]
pdf = PdfFileReader(path)
#split pdf file to multiple pds
for page in range(pdf.getNumPages()):
pdf_writer = PdfFileWriter()
pdf_writer.addPage(pdf.getPage(page))
output_filename = 'media/diploma/{}_page_{}.pdf'.format(
fname, page + 1)
with open(output_filename, 'wb') as out:
pdf_writer.write(out)
#print('Created: {}'.format(output_filename))
#convert pdf to png, because tesseracts workd only with png file
for page in range(pdf.getNumPages()):
images = convert_from_path('media/diploma/{}_page_{}.pdf'.format(
fname, page + 1))
for image in images:
image.save(
'media/diploma/{}_page_{}.png'.format(fname, page + 1),
'PNG')
image_name2 = 'media/diploma/{}_page_{}.png'.format(
fname, page + 1)
#crop name or id
img = Image.open(image_name2)
student_id = img.crop((430, 0, 920, 276))
student_id.save('media/diploma/student_{}.png'.format(page +
1))
img_cv = cv2.imread(
'media/diploma/student_{}.png'.format(page + 1))
student_id_rgb = cv2.cvtColor(img_cv, cv2.COLOR_BGR2RGB)
student_id_text = pytesseract.image_to_string(student_id_rgb)
if User.objects.filter(groups__name='Student').filter(
last_name=student_id_text).exists():
#tesseract converts handwriting to txx file
img_cv = cv2.imread(image_name2)
img_rgb = cv2.cvtColor(img_cv, cv2.COLOR_BGR2RGB)
new = pytesseract.image_to_string(img_rgb)
true_answers = open('media/RightAnswers.txt').read()
m = SequenceMatcher(None, new, true_answers) #insert mark
mark = m.ratio() * 1388 #insert markz
Assignments.objects.create(
grade=mark,
task=Task.objects.get(id=task_id),
student=Student.objects.get(user=User.objects.filter(
groups__name='Student').filter(
last_name=student_id_text).first()))
assignments = Assignments.objects.filter(task_id=task_id)
return render(
request, 'qmain/check_exam.html', {
'title': 'Exam check',
'course_id': course_id,
'task_id': task_id,
'assignments': assignments
})
def parseDataFromImgs(nameImg, itemImg):
# Tess extraction
nameText = pytesseract.image_to_string(nameImg)
itemText = pytesseract.image_to_string(itemImg)
# Parsing
name = parseName(nameText)
itemName, itemQuantity, price, seenAs = parseItemData(itemText)
return name, itemName, itemQuantity, price, seenAs
def process_image():
image = Image.open("screenshots/1ttp39.jpg")
width, height = image.size
pixel_access = image.load()
# When searching the image, either we haven't found a band, or we're in a swb color band
# or we're in a recipient color band
# possible values are None, "swb" and "recipient"
search_mode = None
# it's a goddamned state machine
# y increases from top to bottom of image
lower_y = None
# list of tuples of (lower_y, upper y)
image_band_dimensions = []
for y in xrange(height):
if search_mode == None:
if is_swb_color_in_row(pixel_access, width, y):
lower_y = y
search_mode = 'swb'
continue
if is_recipient_color_in_row(pixel_access, width, y):
lower_y = y
search_mode = 'recipient'
continue
# I'm making the assumption that you're always going to have some space b/w the text bubbles
if search_mode == "swb":
if (not is_swb_color_in_row(pixel_access, width, y)):
search_mode = None
image_band_dimensions.append((lower_y, y))
continue
if search_mode == 'recipient':
if (not is_recipient_color_in_row(pixel_access, width, y)):
search_mode = None
image_band_dimensions.append((lower_y, y))
continue
image_bands = []
for band_dimension in image_band_dimensions:
box = (0, band_dimension[0], width, band_dimension[1])
region = image.crop(box)
# region.show()
image_bands.append(region)
for band in image_bands:
image_name = str(int(round(random.random() * 10000))) + '.jpg'
print image_name
band.save('screenshots/processed/' + image_name)
print pytesseract.image_to_string(band)
print pytesseract.image_to_string(image)
def process_image():
image = Image.open("screenshots/1ttp39.jpg")
width, height = image.size
pixel_access = image.load()
# When searching the image, either we haven't found a band, or we're in a swb color band
# or we're in a recipient color band
# possible values are None, "swb" and "recipient"
search_mode = None
# it's a goddamned state machine
# y increases from top to bottom of image
lower_y = None
# list of tuples of (lower_y, upper y)
image_band_dimensions = []
for y in xrange(height):
if search_mode == None:
if is_swb_color_in_row(pixel_access, width, y):
lower_y = y
search_mode = 'swb'
continue
if is_recipient_color_in_row(pixel_access, width, y):
lower_y = y
search_mode = 'recipient'
continue
# I'm making the assumption that you're always going to have some space b/w the text bubbles
if search_mode == "swb":
if (not is_swb_color_in_row(pixel_access, width, y)):
search_mode = None
image_band_dimensions.append((lower_y, y))
continue
if search_mode == 'recipient':
if (not is_recipient_color_in_row(pixel_access, width, y)):
search_mode = None
image_band_dimensions.append((lower_y, y))
continue
image_bands = []
for band_dimension in image_band_dimensions:
box = (0, band_dimension[0], width, band_dimension[1])
region = image.crop(box)
# region.show()
image_bands.append(region)
for band in image_bands:
image_name = str(int(round(random.random() * 10000))) + '.jpg'
print image_name
band.save('screenshots/processed/' + image_name)
print pytesseract.image_to_string(band)
print pytesseract.image_to_string(image)
def similar_image(ffc, frames, image):
# accumulate lists of frames with similar text
try:
if not frames: # If this is the first frame in the potential bottom third
return pytesseract.image_to_string(image), True
else:
if len(pytesseract.image_to_string(image)) < 3:
return ffc, False
return ffc, SequenceMatcher(
None, ffc,
pytesseract.image_to_string(image)).ratio() > .5
except TesseractError as e:
print(e)
return ffc, True
def validateNumber(cell, found, could_be):
k = getKernel(cell)
eroded = cv2.erode(cell, k)
dilated = cv2.dilate(cell, k)
eroded_d = pytesseract.image_to_string(eroded,
lang='eng',
config='--psm 6 --oem 3')[0]
dilated_d = pytesseract.image_to_string(dilated,
lang='eng',
config='--psm 6 --oem 3')[0]
potential_numbers = [eroded_d, dilated_d, found]
could_be_count = potential_numbers.count(could_be)
return could_be if could_be_count > 1 else found
def extract_data(image):
# extract text
text = pytesseract.image_to_string(image)
# convert to PDF
pdf = pytesseract.image_to_pdf_or_hocr(image)
# done
return text, pdf
def run_tests():
imgs = []
fails = 0
tests = 0
for filename in os.listdir("img/tests"):
t = (Image.open("img/tests/" + filename), filename)
tests += 1
imgs.append(t)
for e in imgs:
dprint(e)
print()
for ele in imgs:
img = ele[0]
img = enhance(img, "enhance-" + ele[1][0])
text = image_to_string(img, lang='eng')
c = solve_check(text) # Gets character for afk check
ans = ele[1][0] # Answer is in the file name
if c == ans:
print("Test passed")
else:
if c:
print("[FAILED] Test failed. Got '%s' instead of '%s'." % (c, ans))
else:
print("[FAILED] Test failed. Regex did not return any result.")
fails += 1
print("Failed %d out of %d times" % (fails, tests))
def captcha_to_text():
"captures the captcha image"
elem = driver.find_element_by_id("captcha_image")
loc = elem.location
size = elem.size
left = loc['x']
top = loc['y']
width = size['width']
height = size['height']
box = (int(left), int(top), int(left + width), int(top + height))
screenshot = driver.get_screenshot_as_base64()
img = Image.open(BytesIO(base64.b64decode(screenshot)))
area = img.crop(box)
full_path = r'/home/sangharshmanuski/Documents/e_courts/captcha'
area.save(os.path.join(full_path, 'file_trial.png'), 'PNG')
img = cv2.imread(os.path.join(full_path, 'file_trial.png'), 0)
ret, thresh1 = cv2.threshold(img, 111, 255, cv2.THRESH_BINARY)
cv2.imwrite(
'/home/sangharshmanuski/Documents/e_courts/editC/oneDisNoLoop.png',
thresh1)
# know the text with pytesseract
captcha_text = pytesseract.image_to_string(
Image.open(
'/home/sangharshmanuski/Documents/e_courts/editC/oneDisNoLoop.png')
)
return captcha_text
def getDigit(self, image, iter=0):
if iter > 0:
image = noise_removal(image)
dig = remove_specials(
pytesseract.image_to_string(image,
lang='eng',
config='--psm 6 --oem 3'))
dig = fixCommonErrors(dig)
digitsOnly = [d for d in dig if d.isdigit()]
if not digitsOnly:
if iter > 3:
return "0"
else:
return self.getDigit(image, iter + 1)
if digitsOnly[0] == "1":
digitsOnly[0] = validateNumber(image, "1", "7")
if digitsOnly[0] == "2":
digitsOnly[0] = validateNumber(image, "2", "7")
if digitsOnly[0] == "4":
digitsOnly[0] = validateNumber(image, "4", "1")
return digitsOnly[0]
def imgtotxt():
elem = driver.find_element_by_id("captcha_image")
loc = elem.location
size = elem.size
left = loc['x']
top = loc['y']
width = size['width']
height = size['height']
box = (int(left), int(top), int(left + width),
int(top + height))
screenshot = driver.get_screenshot_as_base64()
img = Image.open(BytesIO(base64.b64decode(screenshot)))
area = img.crop(box)
area.save(
'/home/sangharshmanuski/Documents/e_courts/captcha/file_trial.png',
'PNG')
fullPath = r'/home/sangharshmanuski/Documents/e_courts/captcha'
f = os.listdir(fullPath)
desPath = r"/home/sangharshmanuski/Documents/e_courts/editC"
img = cv2.imread(os.path.join(fullPath, 'file_trial.png'), 0)
ret, thresh1 = cv2.threshold(img, 111, 255, cv2.THRESH_BINARY)
cv2.imwrite(
'/home/sangharshmanuski/Documents/e_courts/editC/oneDisNoLoop.png',
thresh1)
# know the text with pytesseract
captchaText = pytesseract.image_to_string(
Image.open(
'/home/sangharshmanuski/Documents/e_courts/editC/oneDisNoLoop.png'
))
captcha = driver.find_element_by_id('captcha')
captcha.send_keys(captchaText)
driver.find_element_by_css_selector(
'input.button:nth-child(1)').click()
time.sleep(1)
def convert_to_text():
while True:
time.sleep(1)
image = Image.open('img.png')
txt = pytesseract.image_to_string(image)
pyperclip.copy(txt)
print(txt)
def get_mark(path, path_answers):
img_cv = cv2.imread(path)
img_rgb = cv2.cvtColor(img_cv, cv2.COLOR_BGR2RGB)
new = pytesseract.image_to_string(img_rgb)
true_answers = open(path_answers).read()
mark = SequenceMatcher(None, new, true_answers).ratio() * 1388
return mark
def generateProcessedFiles(splittedName):
# Get string from Image
fullName = splittedName[0] + '.' + splittedName[1]
imgText = pt.image_to_string(Image.open(fullName), lang='ron')
with open('../TextIntermediar/' + splittedName[0] + "text.txt", 'w') as f:
f.write(imgText)
# Get bounding boxes
pt.run_tesseract(fullName,
splittedName[0] + 'output',
lang='ron',
boxes=True,
config="hocr")
#Remove non-alphanumeric characters
with open(splittedName[0] + 'output.box', 'r+') as f:
buf = ''
for line in f:
if line[0].isalnum() or line[0] == '(' or line[0] == ')':
buf += line
f.seek(0)
f.write(buf)
# To read the coordinates
boxes = []
with open(splittedName[0] + 'output.box', 'r') as f:
reader = csv.reader(f, delimiter=' ')
for row in reader:
if (len(row) == 6):
boxes.append(row)
def validate_code_recognition():
flag = request.form.get("flag", True)
image = request.files.get("image", None)
lang = request.form.get("lang", "eng")
oem = request.form.get("oem", "1")
psm = request.form.get("psm", "3")
url = request.form.get("url", None)
config = f"-l {lang} --oem {oem} --psm {psm}"
if not image and not url:
return jsonify({"msg": "image file don`t exist!", "code": 0}), 400
if image:
filename = image.filename
else:
filename = download_image(url)
image = filename
if not allowed_file(filename):
return jsonify({
"msg": f"Allowed file types are: {ALLOWED_EXTENSIONS}",
"code": 0
}), 400
if flag:
image = get_blackwhite_image(image)
else:
image = Image.open(image)
text = pytesseract.image_to_string(image, config=config)
return jsonify({"code": 1, "msg": "success", "data": {"text": text}}), 200
def recogit(imgobject=None):
'''
# 程序入口，识别模块
:param imgobject: 图片对象
:return: nearest:最可能的值，result : [48, '4', '4_0.txt', 48, '4', '4_0.txt']
'''
if not imgobject:
im = Image.open('D:/workspace/captcha/40.png')
else:
im = imgobject
im = im.convert("L")
size = im.size
img_array = im.load()
for i in range(size[0]):
for j in range(size[1]):
if i != 0 and j != 0 and i != 44 and j != 59: # 把边界变为纯白
if (img_array[i, j] > 250):
# print('像素点：%d,%d,%d'%(i,j,img_array[i,j]))
img_array[i, j] = 255
else:
img_array[i, j] = 0
else:
img_array[i, j] = 255
# 去除噪点
clearnoise(img_array, size, 3)
clearnoise(img_array, size, 3)
clearnoise(img_array, size, 3)
clearnoise2(im, 140, 3, 1)
clearnoise(img_array, size, 3)
s = pt.image_to_string(im, lang="dt", config="-psm 7")
return s
def get_ocr_string(img_orig):
"""Return string value from image."""
img_resized = prepareImage(img_orig)
config_ocr = '--psm 7 --oem 1'
return pytesseract.image_to_string(img_resized, 'eng',
config=config_ocr).strip()
def exchange(user, index):
_session = requests.session()
resp = _session.get(login_bypass_url)
headers = {
"Content-Type": "application/x-www-form-urlencoded; charset=UTF-8",
"Origin": "http://h5.nty.tv189.com",
"Referer": "http://h5.nty.tv189.com/csite/tysx/uc/login-by-pass?goBackUrl=",
"User-Agent": "Mozilla/5.0 (Linux; Android 6.0.1; MuMu Build/V417IR; wv) AppleWebKit/537.36 (KHTML, like Gecko) Version/4.0 Chrome/52.0.2743.100 Mobile Safari/537.36/newtysx-android-ua-5.5.9.39",
"X-Requested-With": "XMLHttpRequest"}
login_resp = _session.post(login_url,
data={'uname': user, 'upass': users['upass']},
headers=headers)
login_json = json.loads(login_resp.text)
if login_json.get('code') == 0:
for i in range(1, 100):
png = _session.get(validate_url)
if png.content[:2].hex() == '0d0a':
bys = png.content[2:]
else:
bys = png.content
image = Image.open(io.BytesIO(bys)).convert('L')
b_img = binarizing(image, 126)
img = depoint(b_img)
code = pytesseract.image_to_string(img, lang='tv189', config=tessdate_dir)[:-2].replace(" ", "")
print(code)
resp = _session.get(exchange_url % (str(phones[index]), code))
resp_json = json.loads(resp.text)
print(resp.json())
if resp_json['code'] == 0 or resp_json['code'] == -3:
return
else:
print("login failed")
def get_verification_code_by_pytesseract(driver, selector):
"""
使用pytesseract通过webdriver和验证码的定位器selector获取验证码
:param driver: webdriver对象
:param selector: 验证码的定位器selector
:return: 验证码
"""
# 获取screenshot文件夹路径
path = os.path.dirname(os.path.dirname(__file__)) + '\\screenshot'
# 截屏保存并保存图片名称为当前时间戳.png
raw_picture_name = path + '\\' + str(time.time()) + ".png"
driver.save_screenshot(raw_picture_name)
# 定位到验证码图片并获取到图片的左上右下的xy坐标
element = driver.find_element_by_css_selector(selector)
left_top_x = element.location['x']
left__top_y = element.location['y']
right_down_x = element.size['width'] + left_top_x
right_down_y = element.size['height'] + left__top_y
# 获取屏幕的缩放比例
dpr = driver.execute_script('return window.devicePixeRatio')
# 在截屏中抠出验证码图片并保存图片名称为当前时间戳.png
raw_image = Image.open(raw_picture_name)
target_image = raw_image.crop((left_top_x * dpr, left__top_y * dpr, right_down_x * dpr, right_down_y * dpr))
code_picture_name = path + '\\' + str(time.time()) + ".png"
target_image.save(code_picture_name)
# 使用pytesseract获取图片中的验证码
code_image = Image.open(code_picture_name)
verification_code = pytesseract.image_to_string(code_image)
return verification_code
def captcha_processor():
hp_image_name = 'home_page.png'
driver.get_screenshot_as_file(hp_image_name)
img = Image.open(hp_image_name)
box = (285, 380, 345, 415)
region = img.crop(box)
captcha_img_name = 'captcha.png'
region.save(captcha_img_name)
# get img source
# captcha_img = driver.find_element_by_id('captcha')
# captcha_img_src = captcha_img.get_attribute('src')
# urllib.urlretrieve(captcha_img_src, captcha_img_name)
im = Image.open(captcha_img_name)
imgry = im.convert('L')
sharpness = ImageEnhance.Contrast(imgry)
sharp_img = sharpness.enhance(2.0)
sharp_img.save(captcha_img_name)
captcha_code = pytesseract.image_to_string(sharp_img)
if os.path.exists(hp_image_name):
os.remove(hp_image_name)
if os.path.exists(captcha_img_name):
os.remove(captcha_img_name)
return captcha_code.replace(' ', '') # replace all space
def get_code_text():
r = session.get(url=img_url)
img = Image.open(BytesIO(r.content))
if is_save:
img.save("D:\\data\\code{}.PNG".format(mobile_no))
image = img.convert('L')
if is_save:
image.save("D:\\data\\code_black{}.PNG".format(mobile_no))
pixels = image.load()
# 【二值化】阀值：standard
standard1, standard2 = (100, 170)
# 【描边邻域降噪】阀值：standard
for x in range(image.width):
for y in range(image.height):
if x >= image.width - 1 or y >= image.height - 1:
# 边缘过滤
pixels[x, y] = 255
elif pixels[x, y] < standard2 and pixels[x + 1, y] < standard2 and pixels[x, y + 1] < standard2:
# 深色并且粗线保留
# 浅色加深
pixels[x, y] = 0
else:
# 细线过滤
pixels[x, y] = 255
if is_save:
image.save("D:\\data\\code_scan{}.PNG".format(mobile_no))
testdata_dir_config = '--tessdata-dir "D:/Program Files (x86)/Tesseract-OCR/tessdata"'
text_code = pytesseract.image_to_string(image, lang='eng', config=testdata_dir_config)
# 去掉非法字符，只保留字母数字
return re.sub("\W", "", text_code)
def sampleA(filepath):
# 1枠の縦横ピクセル
w = 80
h = 65
# 枠線を除去するマージンピクセル。縦横共通。
margin_w = 20
margin_h = 15
img = Image.open(filepath, "r")
for pos_h in range(0, 3):
# 読み取るマスの左上位置
leftpos_y = margin_h + (h * pos_h) + (margin_h * pos_h)
for pos_w in range(0, 3):
# 読み取るマスの左上位置
leftpos_x = margin_w + (w * pos_w) + (margin_w * pos_w)
# 該当のマスをトリミング
crop = img.crop((leftpos_x, leftpos_y, leftpos_x + w, leftpos_y + h))
# crop.save("/opt/crop" + str(imgid) + "-" + str(pos_h) + "-" + str(pos_w) + ".png")
# インストールしたtesseractコマンドのパス
pytesseract.tesseract_cmd = "/usr/bin/tesseract"
# -psm 10は1文字判定のフラグ
result = pytesseract.image_to_string(crop, config="-psm 10 -c tessedit_char_whitelist='0123456789-.'", lang="eng+jpn")
print(result)
def to_text(img: Image.Image) -> str:
"""Performs OCR on an image.
Performs :abbr:`OCR (optical character recognition)` on the
:class:`~PIL.Image.Image` :any:`img` and returns the resulting string.
Notes
-----
OCR is performed using :mod:`pytesseract`.
Parameters
----------
img: Image.Image
The image on which to perform OCR.
Returns
-------
str
The resulting text from OCR.
"""
log.debug("Performing OCR.")
config = f'--tessdata-dir "{g.config["tessdata_dir"]}"' if g.config[
"tessdata_dir"] else None
return pytesseract.image_to_string(img, config=config)
def extract_text(frames,
tess_psm="",
tess_char_whitelist=default_whitelist,
oem=""):
config = "-c tessedit_char_whitelist=" + tess_char_whitelist + " --oem " + oem + " --psm " + tess_psm
res = pytesseract.image_to_string(frames[0], config=config)
return res
def read_small_and_save(self, img_path, single_pic_doc, only_save=False):
"""
识别小图片并且保存
:param img_path: 图片路径
:param single_pic_doc: 单个文件进行保存的路径
:return:
"""
# 旧的切割图片方式
# small_pics = self.cutting_pic(img_path)
small_img_objs = self.cutting_pic_update(img_path)
if only_save:
name_list = []
for small_img_obj in small_img_objs:
time.sleep(0.1)
single_pic_path = '{0}/{1}.png'.format(single_pic_doc,
int(time.time() * 10))
resized = small_img_obj.resize((20, 30), Image.ANTIALIAS)
resized.save(single_pic_path)
name_list.append(single_pic_path)
return name_list
else:
for small_img_obj in small_img_objs:
text = pytesseract.image_to_string(small_img_obj,
config='-psm 10')
if not is_value_name(text):
text = 'temp'
small_img_path = '{0}/{1}_{2}.png'.format(
single_pic_doc, int(time.time() * 10), ''.join(text))
resized = small_img_obj.resize((20, 30), Image.ANTIALIAS)
resized.save(small_img_path)
print("单个图识别结果：", text)
def parse_code():
driver.save_screenshot('code.png')
img = cv2.imread('code.png')
cro = img[1042:1106, 2282:2397]
cv2.imwrite('real_code.png', cro)
text = pytesseract.image_to_string("real_code.png")
return text
def input():
a1 = pyautogui.position().x
a2 = pyautogui.position().y
sleep(0.2)
while True:
b1 = pyautogui.position().x
b2 = pyautogui.position().y
if mouse.is_pressed(button='middle'):
break
temp1 = a1
temp2 = a2
if b2 - a2 < 0 and b1 - a1 < 0:
a1 = b1
a2 = b2
b1 = temp1
b2 = temp2
elif b2 - a2 > 0 and b1 - a1 < 0:
a1 = b1
b1 = temp1
elif b2 - a2 < 0 and b1 - a1 > 0:
a2 = b2
b2 = temp2
grab(bbox=(a1, a2, b1, b2)).save('screen.png')
screen = Image.open('screen.png')
"""
display = pygame.display.set_mode(screen.size)
showimage = pygame.image.load('screen.png')
display.blit(showimage, (0, 0))
pygame.display.update()
"""
vocab = pytesseract.image_to_string(screen)
sent = '\n' + vocab + "\n" + translator.translate(vocab, dest='th').text
lineNotify(sent)
def extracteddetails(filename):
c_id = ar[0]
mycursor.execute("SELECT * from vacancies where company_id='%s'" % c_id)
values1 = mycursor.fetchall()
# vid = datetime.datetime.now().strftime("%I:%M%p%B%d%Y").replace(":/ ", "") + filename
pytesseract.tesseract_cmd = r'C:\Program Files\Tesseract-OCR\tesseract.exe'
text = pytesseract.image_to_string(Image.open((os.path.join(app.config['UPLOAD_FOLDER'], filename))), lang='eng')
txt = []
for line in text.split('\n' or ' ' or '. ' or ','):
line2 = line.strip()
if line2 != '':
txt.append(line2)
print(txt)
otherskills = []
# education = []
# skills = []
experience = []
for lineindex, line in enumerate(txt):
if extract_otherskills(line) is not None:
otherskills.append(line)
# if extract_education(line) is not None:
# education.append(line)
# if extract_skills(line) is not None:
# skills.append(line)
experience = []
if extract_experience(line) is not None:
experience.append(line)
job = extract_job(text)
skills = extract_skills2(text)
education = extract_education(text)
vjob = json.dumps(job)
ved = json.dumps(education)
vsk = json.dumps(skills)
vos = json.dumps(otherskills)
vex = json.dumps(experience)
if request.method == 'POST':
jobv = request.form['vjob']
eduv = request.form['vedu']
expv = request.form['vexp']
eyv = request.form['vey']
vskv = request.form['vski']
osv = request.form['voski']
sql_vacancies = "INSERT INTO vacancies (company_id, job, education, experience, ex_year,skills,other_skills) VALUES(%s,%s,%s,%s,%s,%s,%s) "
val_vacancies = (c_id, jobv, eduv, expv, eyv, vskv, osv)
mycursor.execute(sql_vacancies, val_vacancies)
mydb.commit()
vid = mycursor.lastrowid
print("VID: " + str(vid))
vMatcher.VacancyMatching().matchingByVacencyId(vid)
flash("successfully saved")
return render_template("vacancy.html", data=values1, vvid=c_id)
return render_template('vacancyextract.html', vvjob=vjob, vved=ved, vvsk=vsk, vvos=vos, vvex=vex)
def run_tesseract(input_file, output_file, language="deu"):
"""
:param input_file: str
Path to image to OCR
:param output_file: str
Path to output file
:return: void
Runs tesseract on image and saves result
"""
print(ORANGE + '\t~: ' + RESET + 'Parse image using pytesseract' + RESET)
print(ORANGE + '\t~: ' + RESET + 'Parse image at: ' + input_file + RESET)
print(ORANGE + '\t~: ' + RESET + 'Write result to: ' + output_file + RESET)
with io.BytesIO() as transfer:
with WandImage(filename=input_file) as img:
img.auto_level()
img.sharpen(radius=0, sigma=4.0)
img.contrast()
img.save(transfer)
with Image.open(transfer) as img:
image_data = pytesseract.image_to_string(img, lang=language, timeout=60)
out = open(output_file, "w")
out.write(image_data)
out.close()
def code_veryfy(self,link):
local = 'verifycode.jpg'
try:
urllib.request.urlretrieve(link,local)
image1=Image.open(local)
text=image_to_string(image1)
except Exception :
print("Error occured in code verifing... ")
return text
def get_random_code():
resp = requests.get("http://10.3.254.23:8080/dangwebx/randCodeImage?a=1444784364985", stream=True)
with open("tmp.png", "wb") as f:
f.write(resp.raw.read())
# or
shutil.copyfileobj(resp.raw, f)
image = Image.open("tmp.png")
return pytesseract.image_to_string(image)
def perform_ocr_scoreboard(image: Image.Image, column: str, start=START, end=END, diff=DIFF)->(str, int, None):
"""Perform OCR on a part of an image"""
result = None
is_number = column in DIGITS
for treshold in range(start, end, -diff): # Continue until result is valid
template = high_pass_invert(image, treshold)
result = tesseract.image_to_string(template, config="" if column in DIGITS else WHITELIST)
if is_number and not result.isdigit(): # Try again for numbers
result = tesseract.image_to_string(template, config="-psm 10")
if is_number and not result.isdigit():
template = template.filter(ImageFilter.GaussianBlur())
result = tesseract.image_to_string(template, config="-psm 10")
if result == "" or (is_number and not result.isdigit()):
continue
break
if is_number and not result.isdigit():
result = match_digit(image)
if result == "" or (is_number and not result.isdigit()):
return 0 if is_number else None
if is_number:
return int(result)
return result.replace("\n", "").replace(" ", "")
def recognize_captcha(self,r=None, data=None):
if data is None:
return '0000'
file = 'd:/code_{}.png'.format(r)
with open(file, 'wb') as f:
f.write(data)
mem = BytesIO()
mem.write(data)
mem.seek(0)
print(type(Image.open(mem)))
captcha = pytesseract.image_to_string(Image.open(mem), config='digits')
print(captcha)
return captcha
def perform_ocr(image: Image.Image) -> (None, str):
"""Perform OCR on an Image"""
if not is_installed():
print("[Tesseract] Critical error: Tesseract is not installed!")
return None
result = None
for threshold in range(START, END, -DIFF):
template: Image.Image = high_pass_invert(image, threshold)
result: str = tesseract.image_to_string(template)
if result == "":
continue
break
if result == "":
return None
return result
# <h2>6/500</h2>
m = re.search("<h2>(\d+)/\d+</h2>", data)
if m is not None:
s = m.group(1)
print(s)
if int(s) < lastScore:
lastScore = int(s)
im.show() # debug, display failed captcha to improve blacklist
else:
lastScore = int(s)
f = opener.open("http://ctfquest.trendmicro.co.jp:8080/acf2e28903f698eda9bdefc043b6edc3/image")
image_file = io.BytesIO(f.read())
im = Image.open(image_file)
im = im.convert('RGB')
filling = fix_colors(im)
black_and_white(im, filling)
text = pytesseract.image_to_string(im,
config="-psm 8 --user-patterns /cygdrive/d/tess/pattern.txt /cygdrive/d/tess/conf.txt")
text = text.replace(" ", "")
if not on_blacklist(text):
print(text)
params = {"captcha": text}
params.update({c.name: c.value for c in cookies.cookiejar})
encoded_params = urllib.urlencode(params)
opener.open("http://ctfquest.trendmicro.co.jp:8080/acf2e28903f698eda9bdefc043b6edc3/challenge",
encoded_params)
else:
print "Possibly wrong decode " + text + " skipping"
except:
pass
import shutil
from PIL import Image
from pytesseract import pytesseract
import requests
__author__ = 'bida'
resp = requests.get("http://10.3.254.23:8080/dangwebx/randCodeImage?a=1444784364985", stream=True)
with open("tmp.png", "wb") as f:
f.write(resp.raw.read())
shutil.copyfileobj(resp.raw, f)
image = Image.open("tmp.png")
print pytesseract.image_to_string(image)
def print_text(cls, img):
text = pytesseract.image_to_string(img, lang="eng")
if text:
print(text)
