def scrape(self, **kwargs):
logger.info("Solving reCAPTCHA (~30s)")
captcha_solver = CaptchaSolver(api_key=config.ANTI_CAPTCHA_KEY)
# Site key obtained from: https://www.homedepot.com/mycheckout/assets/react/giftcard.bundle.1.2302.0.js
captcha_resp = captcha_solver.solve_recaptcha(
self.website_url, "6LfEHBkTAAAAAHX6YgeUw9x1Sutr7EzhMdpbIfWJ")
if captcha_resp["errorId"] != 0:
raise RuntimeError(
f"Unable to solve reCAPTCHA ({captcha_resp['errorDescription']})"
)
payload = {
"GiftCardsRequest": {
"cardNumber": kwargs["card_number"],
"pinNumber": kwargs["pin"],
"reCaptcha": captcha_resp["solution"]["gRecaptchaResponse"],
}
}
logger.info("Fetching balance from API")
try:
resp = requests.post(self.api_endpoint,
json=payload,
headers=HEADERS)
if resp.status_code != 200:
raise RuntimeError(
f"Invalid API response (status code {resp.status_code})")
result = deep_get(resp.json(), "giftCards.giftCard")
if result is None:
raise RuntimeError(
f"Invalid API response: unable to find giftCard key in JSON response"
)
err_code = deep_get(result, "errorCode")
if err_code:
err_desc = deep_get(result, "description")
raise RuntimeError(
f"Failed to retrieve balance from API: {err_desc} ({err_code})"
)
initial_balance = deep_get(result, "originalAmount")
avail_balance = deep_get(result, "availableAmount")
logger.info(f"Success! Card balance: {avail_balance}")
return {
"initial_balance": initial_balance,
"available_balance": avail_balance,
}
except requests.exceptions.RequestException as e:
raise RuntimeError(f"Error on API post: {e}")
except JSONDecodeError:
raise RuntimeError("Failed to parse API response as JSON")
def scrape(self, fields):
session = requests.Session()
session.headers.update({"User-Agent": config.USER_AGENT})
logger.info("Fetching balance check page")
resp = session.get(self.website_url)
if resp.status_code != 200:
raise RuntimeError(
"Failed to GET Spafinder website (status code {})".format(
resp.status_code))
page_html = BeautifulSoup(resp.content, features="html.parser")
inquiry = page_html.find(id="balance-inquiry")
form = inquiry.find("form")
if not form:
raise RuntimeError("Unable to find balance check form")
endpoint = "{}{}".format(self.base_url, form["action"])
# Page has bad HTML, need to search from top level
recaptcha_field = page_html.find("div", class_="g-recaptcha")
if not recaptcha_field:
raise RuntimeError("Unable to find reCAPTCHA")
site_key = recaptcha_field["data-sitekey"]
logger.info("Solving reCAPTCHA (~30s)")
captcha_solver = CaptchaSolver(api_key=config.ANTI_CAPTCHA_KEY)
captcha_resp = captcha_solver.solve_recaptcha(self.website_url,
site_key)
if captcha_resp["errorId"] != 0:
raise RuntimeError("Unable to solve reCAPTCHA ({})".format(
captcha_resp["errorDescription"]))
fields["g-recaptcha-response"] = captcha_resp["solution"][
"gRecaptchaResponse"]
logger.info("Fetching card balance")
form_resp = session.post(endpoint, data=fields)
if form_resp.status_code != 200:
raise RuntimeError(
"Failed to retrieve card balance (status code {})".format(
form_resp.status_code))
resp_html = BeautifulSoup(form_resp.content, features="html.parser")
error_html = resp_html.find("div", class_="alert-danger")
if error_html:
raise RuntimeError("Got error while checking balance: {}".format(
error_html.text))
balance_container = resp_html.find("div", class_="alert-success")
if not balance_container:
raise RuntimeError("Unable to find balance container")
balance_match = re.search("(\d{1,2}\.\d{2})",
balance_container.get_text())
if not balance_match:
raise RuntimeError("Unable to find balance text")
balance = "${}".format(balance_match.group(0))
logger.info("Success! Card balance: {}".format(balance))
return {"balance": balance}
def scrape(self, fields):
# Open Selenium browser
browser = webdriver.Chrome()
browser.set_window_size(600, 800)
logger.info("Fetching balance check page")
browser.get(self.website_url)
try:
form = browser.find_elements_by_tag_name("form")[1]
except (NoSuchElementException, IndexError):
raise RuntimeError("Unable to find login form on page")
logger.info("Filling login form 1/2")
try:
form.find_element_by_name("accountNumber").send_keys(
fields["accountNumber"])
time.sleep(1)
except NoSuchElementException:
browser.close()
raise RuntimeError(f"Unable to find 'accountNumber' field on page")
# Click continue button
form.find_element_by_css_selector("input[type='submit']").click()
# Wait for page to load
try:
WebDriverWait(browser, 3).until(
EC.presence_of_element_located((By.ID, "login-form")))
except TimeoutException:
browser.close()
raise RuntimeError("Login page took too long to load")
try:
form = browser.find_element_by_id("login-form")
except NoSuchElementException:
browser.close()
raise RuntimeError("Unable to find login form on page")
logger.info("Solving CAPTCHA (~10s)")
# Extract CAPTCHA image from page
captcha_b64 = get_image_b64_by_id(browser, "captchaImg")
captcha_solver = CaptchaSolver(api_key=config.ANTI_CAPTCHA_KEY)
captcha = captcha_solver.solve_image_b64(captcha_b64)
if captcha["errorId"] != 0:
browser.close()
raise RuntimeError("Unable to solve CAPTCHA ({})".format(
captcha["errorDescription"]))
logger.info("Filling login form 2/2")
try:
form.find_element_by_name("cv2").send_keys(fields["cv2"])
time.sleep(1)
except NoSuchElementException:
browser.close()
raise RuntimeError("Unable to find 'cv2' field on page")
try:
form.find_element_by_id(
"_MultiStageFSVpasswordloginresponsive_WAR_cardportalresponsive_captchaText"
).send_keys(captcha["solution"]["text"])
time.sleep(1)
except NoSuchElementException:
browser.close()
raise RuntimeError("Unable to find CAPTCHA field on page")
# Click continue button
form.find_element_by_css_selector("input[type='submit']").click()
# Wait for page to load
try:
WebDriverWait(browser, 3).until(
EC.presence_of_element_located((By.ID, "cardBalanceInfo")))
except TimeoutException:
browser.close()
raise RuntimeError("Balance page took too long to load")
logger.info("Obtaining card information")
try:
avail_balance = browser.find_element_by_class_name(
"cardBalanceText").text
except NoSuchElementException:
browser.close()
raise RuntimeError("Could not find available card balance")
browser.close()
logger.info(f"Success! Card balance: {avail_balance}")
return {"initial_balance": None, "available_balance": avail_balance}
def scrape(self, fields):
session = requests.Session()
session.headers.update({"User-Agent": config.USER_AGENT})
fields["X-Requested-With"] = "XMLHttpRequest"
logger.info("Fetching balance check page")
resp = session.get(self.website_url)
if resp.status_code != 200:
raise RuntimeError(
f"Failed to GET Blackhawk website (status code {resp.status_code})"
)
page_html = BeautifulSoup(resp.content, features="html.parser")
transactions = page_html.find(id="CheckBalanceTransactions")
form = transactions.find("form")
if not form:
raise RuntimeError("Unable to find balance check form")
endpoint = "{}{}".format(self.website_url, form["action"])
token_field = transactions.find(
"input", attrs={"name": "__RequestVerificationToken"})
if not token_field:
raise RuntimeError("Failed to retrieve verification token")
fields["__RequestVerificationToken"] = token_field["value"]
arkose_key = extract_arkose_key(resp.text)
if not arkose_key:
raise RuntimeError("Failed to extract Arkose Labs public key")
logger.info("Solving FunCaptcha (~30s)")
captcha_solver = CaptchaSolver(api_key=config.ANTI_CAPTCHA_KEY)
captcha = captcha_solver.solve_funcaptcha(self.website_url, arkose_key)
if captcha["errorId"] != 0:
raise RuntimeError("Unable to solve FunCaptcha ({})".format(
captcha["errorDescription"]))
fields["captchaToken"] = captcha["solution"]["token"]
logger.info("Fetching card balance")
session.headers.update({
"Accept": "*/*",
"Accept-Encoding": "gzip, deflate, br",
"Accept-Language": "en-US,en;q=0.5",
"Cache-Control": "no-cache",
"Connection": "keep-alive",
"Content-Type": "application/x-www-form-urlencoded; charset=UTF-8",
"Pragma": "no-cache",
"Referer": "https://mygift.giftcardmall.com/",
"X-Requested-With": "XMLHttpRequest",
})
form_resp = session.post(endpoint, data=fields)
if form_resp.status_code != 200:
raise RuntimeError(
"Failed to retrieve card balance (status code {})".format(
form_resp.status_code))
balance_html = BeautifulSoup(form_resp.content, features="html.parser")
avail_balance = (balance_html.find(
"div", text="Available Balance").parent.find("div",
class_="value").text)
initial_balance = (balance_html.find(
"div", text="Initial Balance").parent.find("div",
class_="value").text)
logger.info("Success! Card balance: {}".format(avail_balance))
return {
"initial_balance": initial_balance,
"available_balance": avail_balance
}
def scrape(self, fields):
session = requests.Session()
session.headers.update({"User-Agent": config.USER_AGENT})
logger.info("Fetching balance check page")
resp = session.get(self.website_url)
if resp.status_code != 200:
logger.critical(
f"Failed to GET Blackhawk website (status code {resp.status_code})"
)
sys.exit(1)
page_html = BeautifulSoup(resp.content, features="html.parser")
form = page_html.find("form")
if not form:
logger.critical("Unable to find balance check form")
sys.exit(1)
endpoint = "{}{}".format(self.website_url, form["action"])
token_field = page_html.find(
"input", attrs={"name": "__RequestVerificationToken"})
if not token_field:
logger.critical("Failed to retrieve verification token")
sys.exit(1)
fields["__RequestVerificationToken"] = token_field["value"]
recaptcha_field = page_html.find("div", class_="g-recaptcha")
if not recaptcha_field:
logger.critical("Unable to find reCAPTCHA")
sys.exit(1)
site_key = recaptcha_field["data-sitekey"]
logger.info("Solving reCAPTCHA (~30s)")
captcha_solver = CaptchaSolver(api_key=config.ANTI_CAPTCHA_KEY)
captcha = captcha_solver.solve_recaptcha(self.website_url, site_key)
if captcha["errorId"] != 0:
logger.critical(
f"Unable to solve reCAPTCHA ({captcha['errorDescription']})")
sys.exit(1)
fields["g-recaptcha-response"] = captcha["solution"][
"gRecaptchaResponse"]
logger.info("Fetching card balance")
session.headers.update({
"Accept": "*/*",
"Accept-Encoding": "gzip, deflate, br",
"Accept-Language": "en-US,en;q=0.5",
"Cache-Control": "no-cache",
"Connection": "keep-alive",
"Content-Type": "application/x-www-form-urlencoded; charset=UTF-8",
"Pragma": "no-cache",
"Referer": self.website_url,
"X-Requested-With": "XMLHttpRequest",
"Host": "cardholder.happycards.com",
"Origin": "https://cardholder.happycards.com",
})
form_resp = session.post(endpoint, data=fields)
if form_resp.status_code != 200:
logger.critical(
f"Failed to retrieve card balance (status code {form_resp.status_code})"
)
sys.exit(1)
balance_html = BeautifulSoup(form_resp.content, features="html.parser")
try:
avail_balance = (balance_html.find(
"div",
text="Available Balance").parent.find("div",
class_="value").text)
except:
print("Couldnt read available balance from page.")
print(form_resp.content)
initial_balance = (balance_html.find(
"div", text="Initial Balance").parent.find("div",
class_="value").text)
logger.info("Success! Card balance: {}".format(avail_balance))
return {
"initial_balance": initial_balance,
"available_balance": avail_balance
}
def scrape(self, **kwargs):
cookies = CookieJar()
opener = request.build_opener(request.HTTPCookieProcessor(cookies))
logger.info("Fetching balance check page")
# Use urllib directly as requests gets blocked
req = request.Request(self.website_url, headers=HEADERS)
resp = opener.open(req)
if resp.status != 200:
raise RuntimeError(
f"Failed to get GameStop website (status code {resp.status})")
page_html = BeautifulSoup(resp.read(), features="html.parser")
recaptcha_el = page_html.find("div", {"data-sitekey": True})
if not recaptcha_el:
raise RuntimeError("Unable to find reCAPTCHA on page")
csrf_el = page_html.find("input", {"name": "csrf_token"})
if not csrf_el:
raise RuntimeError("Unable to find CSRF on page")
site_key = recaptcha_el["data-sitekey"]
csrf_token = csrf_el["value"]
logger.info("Solving reCAPTCHA (~30s)")
captcha_solver = CaptchaSolver(api_key=config.ANTI_CAPTCHA_KEY)
captcha_resp = captcha_solver.solve_recaptcha(self.website_url,
site_key)
if captcha_resp["errorId"] != 0:
raise RuntimeError(
f"Unable to solve reCAPTCHA ({captcha_resp['errorDescription']})"
)
payload = {
"dwfrm_giftCard_balance_accountNumber": kwargs["card_number"],
"dwfrm_giftCard_balance_pinNumber": kwargs["pin"],
"g-recaptcha-response":
captcha_resp["solution"]["gRecaptchaResponse"],
"csrf_token": csrf_token,
}
logger.info("Fetching balance from API")
try:
req = request.Request(self.api_endpoint, headers=HEADERS)
data = urlencode(payload).encode("utf-8")
req.add_header("Content-Type",
"application/x-www-form-urlencoded; charset=UTF-8")
req.add_header("Content-Length", len(data))
resp = opener.open(req, data)
if resp.status != 200:
raise RuntimeError(
f"Invalid API response (status code {resp.status})")
result = json.loads(resp.read().decode(
resp.info().get_param("charset") or "utf-8"))
errors = deep_get(result, "error")
if errors:
err = errors[0]
raise RuntimeError(
f"Failed to retrieve balance from API: {err}")
balance = deep_get(result, "balance")
if balance is None:
raise RuntimeError(
f"Invalid API response: unable to find required key in JSON response"
)
logger.info(f"Success! Card balance: ${balance}")
return {
"balance": f"${balance}",
}
except json.JSONDecodeError:
raise RuntimeError("Failed to parse API response as JSON")
import colorlog from balance_check import config from balance_check.utils.logging import configure_logger from balance_check.utils.captcha import CaptchaSolver logger = colorlog.getLogger() configure_logger(logger) captcha_solver = CaptchaSolver(api_key=config.ANTI_CAPTCHA_KEY)