Rewrite FlareSolverr from scratch in Python + Selenium

This commit is contained in:
ngosang 2022-09-23 02:17:50 +02:00
parent 8d1ac09bf2
commit 1505595591
10 changed files with 929 additions and 0 deletions

5
requirements.txt Normal file
View File

@ -0,0 +1,5 @@
bottle==0.12.23
waitress==2.1.2
selenium==4.4.3
undetected-chromedriver==3.1.5.post4
func-timeout==4.3.5

View File

View File

@ -0,0 +1,22 @@
from bottle import response
import logging
def error_plugin(callback):
"""
Bottle plugin to handle exceptions
https://stackoverflow.com/a/32764250
"""
def wrapper(*args, **kwargs):
try:
actual_response = callback(*args, **kwargs)
except Exception as e:
logging.error(str(e))
actual_response = {
"error": str(e)
}
response.status = 500
return actual_response
return wrapper

View File

@ -0,0 +1,23 @@
from bottle import request, response
import logging
def logger_plugin(callback):
"""
Bottle plugin to use logging module
http://bottlepy.org/docs/dev/plugindev.html
Wrap a Bottle request so that a log line is emitted after it's handled.
(This decorator can be extended to take the desired logger as a param.)
"""
def wrapper(*args, **kwargs):
actual_response = callback(*args, **kwargs)
if not request.url.endswith("/health"):
logging.info('%s %s %s %s' % (request.remote_addr,
request.method,
request.url,
response.status))
return actual_response
return wrapper

83
src/dtos.py Normal file
View File

@ -0,0 +1,83 @@
STATUS_OK = "ok"
STATUS_ERROR = "error"
class ChallengeResolutionResultT:
url: str = None
status: int = None
headers: list = None
response: str = None
cookies: list = None
userAgent: str = None
def __init__(self, _dict):
self.__dict__.update(_dict)
class ChallengeResolutionT:
status: str = None
message: str = None
result: ChallengeResolutionResultT = None
def __init__(self, _dict):
self.__dict__.update(_dict)
if self.result is not None:
self.result = ChallengeResolutionResultT(self.result)
class V1RequestBase(object):
# V1RequestBase
cmd: str = None
cookies: list = None
maxTimeout: int = None
proxy: dict = None
session: str = None
headers: list = None # deprecated v2.0.0, not used
userAgent: str = None # deprecated v2.0.0, not used
# V1Request
url: str = None
postData: str = None
returnOnlyCookies: bool = None
download: bool = None # deprecated v2.0.0, not used
returnRawHtml: bool = None # deprecated v2.0.0, not used
def __init__(self, _dict):
self.__dict__.update(_dict)
class V1ResponseBase(object):
# V1ResponseBase
status: str = None
message: str = None
startTimestamp: int = None
endTimestamp: int = None
version: str = None
# V1ResponseSolution
solution: ChallengeResolutionResultT = None
# hidden vars
__error_500__: bool = False
def __init__(self, _dict):
self.__dict__.update(_dict)
if self.solution is not None:
self.solution = ChallengeResolutionResultT(self.solution)
class IndexResponse(object):
msg: str = None
version: str = None
userAgent: str = None
def __init__(self, _dict):
self.__dict__.update(_dict)
class HealthResponse(object):
status: str = None
def __init__(self, _dict):
self.__dict__.update(_dict)

91
src/flaresolverr.py Normal file
View File

@ -0,0 +1,91 @@
import json
import logging
import os
import sys
from bottle import run, response, Bottle, request
from bottle_plugins.error_plugin import error_plugin
from bottle_plugins.logger_plugin import logger_plugin
from dtos import IndexResponse, V1RequestBase
import flaresolverr_service
import utils
class JSONErrorBottle(Bottle):
"""
Handle 404 errors
"""
def default_error_handler(self, res):
response.content_type = 'application/json'
return json.dumps(dict(error=res.body, status_code=res.status_code))
app = JSONErrorBottle()
# plugin order is important
app.install(logger_plugin)
app.install(error_plugin)
@app.route('/')
def index():
"""
Show welcome message
"""
res = flaresolverr_service.index_endpoint()
return utils.object_to_dict(res)
@app.route('/health')
def health():
"""
Healthcheck endpoint.
This endpoint is special because it doesn't print traces
"""
res = flaresolverr_service.health_endpoint()
return utils.object_to_dict(res)
@app.post('/v1')
def controller_v1():
"""
Controller v1
"""
req = V1RequestBase(request.json)
res = flaresolverr_service.controller_v1_endpoint(req)
if res.__error_500__:
response.status = 500
return utils.object_to_dict(res)
if __name__ == "__main__":
# validate configuration
log_level = os.environ.get('LOG_LEVEL', 'info').upper()
log_html = utils.get_config_log_html()
server_host = os.environ.get('HOST', '0.0.0.0')
server_port = int(os.environ.get('PORT', 8191))
# configure logger
logging.basicConfig(
format='%(asctime)s %(levelname)-8s ReqId %(thread)s %(message)s',
level=log_level,
datefmt='%Y-%m-%d %H:%M:%S',
handlers=[
logging.StreamHandler(sys.stdout)
]
)
# disable warning traces from urllib3
logging.getLogger('urllib3').setLevel(logging.ERROR)
logging.getLogger('selenium.webdriver.remote.remote_connection').setLevel(logging.WARNING)
logging.getLogger('undetected_chromedriver').setLevel(logging.WARNING)
logging.info(f'FlareSolverr {utils.get_flaresolverr_version()}')
logging.debug('Debug log enabled')
# test browser installation
flaresolverr_service.test_browser_installation()
# start webserver
# default server 'wsgiref' does not support concurrent requests
run(app, host=server_host, port=server_port, quiet=True, server='waitress')

247
src/flaresolverr_service.py Normal file
View File

@ -0,0 +1,247 @@
import logging
import time
from urllib.parse import unquote
from func_timeout import func_timeout, FunctionTimedOut
from selenium.common import TimeoutException
from selenium.webdriver.chrome.webdriver import WebDriver
from selenium.webdriver.common.by import By
from selenium.webdriver.support.wait import WebDriverWait
from selenium.webdriver.support.expected_conditions import presence_of_element_located, staleness_of
from dtos import V1RequestBase, V1ResponseBase, ChallengeResolutionT, ChallengeResolutionResultT, IndexResponse, \
HealthResponse, STATUS_OK, STATUS_ERROR
import utils
CHALLENGE_SELECTORS = [
# Cloudflare
'#cf-challenge-running', '.ray_id', '.attack-box', '#cf-please-wait', '#trk_jschal_js',
# DDoS-GUARD
'#link-ddg',
# Custom CloudFlare for EbookParadijs, Film-Paleis, MuziekFabriek and Puur-Hollands
'td.info #js_info'
]
SHORT_TIMEOUT = 5
def test_browser_installation():
logging.info("Testing web browser installation...")
user_agent = utils.get_user_agent()
logging.info("FlareSolverr User-Agent: " + user_agent)
logging.info("Test successful")
def index_endpoint() -> IndexResponse:
res = IndexResponse({})
res.msg = "FlareSolverr is ready!"
res.version = utils.get_flaresolverr_version()
res.userAgent = utils.get_user_agent()
return res
def health_endpoint() -> HealthResponse:
res = HealthResponse({})
res.status = STATUS_OK
return res
def controller_v1_endpoint(req: V1RequestBase) -> V1ResponseBase:
start_ts = int(time.time() * 1000)
logging.info(f"Incoming request => POST /v1 body: {utils.object_to_dict(req)}")
res: V1ResponseBase
try:
res = _controller_v1_handler(req)
except Exception as e:
res = V1ResponseBase({})
res.__error_500__ = True
res.status = STATUS_ERROR
res.message = "Error: " + str(e)
logging.error(res.message)
res.startTimestamp = start_ts
res.endTimestamp = int(time.time() * 1000)
res.version = utils.get_flaresolverr_version()
logging.debug(f"Response => POST /v1 body: {utils.object_to_dict(res.solution)}")
logging.info(f"Response in {(res.endTimestamp - res.startTimestamp) / 1000} s")
return res
def _controller_v1_handler(req: V1RequestBase) -> V1ResponseBase:
# do some validations
if req.cmd is None:
raise Exception("Request parameter 'cmd' is mandatory.")
if req.headers is not None:
logging.warning("Request parameter 'headers' was removed in FlareSolverr v2.")
if req.userAgent is not None:
logging.warning("Request parameter 'userAgent' was removed in FlareSolverr v2.")
# set default values
if req.maxTimeout is None or req.maxTimeout < 1:
req.maxTimeout = 60000
# execute the command
res: V1ResponseBase
if req.cmd == 'sessions.create':
raise Exception("Not implemented yet.")
elif req.cmd == 'sessions.list':
raise Exception("Not implemented yet.")
elif req.cmd == 'sessions.destroy':
raise Exception("Not implemented yet.")
elif req.cmd == 'request.get':
res = _cmd_request_get(req)
elif req.cmd == 'request.post':
res = _cmd_request_post(req)
else:
raise Exception(f"Request parameter 'cmd' = '{req.cmd}' is invalid.")
return res
def _cmd_request_get(req: V1RequestBase) -> V1ResponseBase:
# do some validations
if req.url is None:
raise Exception("Request parameter 'url' is mandatory in 'request.get' command.")
if req.postData is not None:
raise Exception("Cannot use 'postBody' when sending a GET request.")
if req.returnRawHtml is not None:
logging.warning("Request parameter 'returnRawHtml' was removed in FlareSolverr v2.")
if req.download is not None:
logging.warning("Request parameter 'download' was removed in FlareSolverr v2.")
challenge_res = _resolve_challenge(req, 'GET')
res = V1ResponseBase({})
res.status = challenge_res.status
res.message = challenge_res.message
res.solution = challenge_res.result
return res
def _cmd_request_post(req: V1RequestBase) -> V1ResponseBase:
# do some validations
if req.postData is None:
raise Exception("Request parameter 'postData' is mandatory in 'request.post' command.")
if req.returnRawHtml is not None:
logging.warning("Request parameter 'returnRawHtml' was removed in FlareSolverr v2.")
if req.download is not None:
logging.warning("Request parameter 'download' was removed in FlareSolverr v2.")
challenge_res = _resolve_challenge(req, 'POST')
res = V1ResponseBase({})
res.status = challenge_res.status
res.message = challenge_res.message
res.solution = challenge_res.result
return res
def _resolve_challenge(req: V1RequestBase, method: str) -> ChallengeResolutionT:
timeout = req.maxTimeout / 1000
driver = None
try:
driver = utils.get_webdriver()
return func_timeout(timeout, _evil_logic, (req, driver, method))
except FunctionTimedOut:
raise Exception(f'Error solving the challenge. Timeout after {timeout} seconds.')
except Exception as e:
raise Exception('Error solving the challenge. ' + str(e))
finally:
if driver is not None:
driver.quit()
def _evil_logic(req: V1RequestBase, driver: WebDriver, method: str) -> ChallengeResolutionT:
res = ChallengeResolutionT({})
res.status = STATUS_OK
res.message = ""
# navigate to the page
logging.debug(f'Navigating to... {req.url}')
if method == 'POST':
_post_request(req, driver)
else:
driver.get(req.url)
if utils.get_config_log_html():
logging.debug(f"Response HTML:\n{driver.page_source}")
# find challenge selectors
html_element = driver.find_element(By.TAG_NAME, "html")
challenge_found = False
for selector in CHALLENGE_SELECTORS:
found_elements = driver.find_elements(By.CSS_SELECTOR, selector)
if len(found_elements) > 0:
challenge_found = True
logging.info("Challenge detected. Selector found: " + selector)
break
if challenge_found:
while True:
try:
# then wait until all the selectors disappear
for selector in CHALLENGE_SELECTORS:
logging.debug("Waiting for selector: " + selector)
WebDriverWait(driver, SHORT_TIMEOUT).until_not(
presence_of_element_located((By.CSS_SELECTOR, selector)))
# all elements not found
break
except TimeoutException:
logging.debug("Timeout waiting for selector")
# update the html (cloudflare reloads the page every 5 s)
html_element = driver.find_element(By.TAG_NAME, "html")
# waits until cloudflare redirection ends
logging.debug("Waiting for redirect")
# noinspection PyBroadException
try:
WebDriverWait(driver, SHORT_TIMEOUT).until(staleness_of(html_element))
except Exception:
logging.debug("Timeout waiting for redirect")
logging.info("Challenge solved!")
else:
logging.info("Challenge not detected!")
challenge_res = ChallengeResolutionResultT({})
challenge_res.url = driver.current_url
challenge_res.status = 200 # todo: fix, selenium not provides this info
challenge_res.cookies = driver.get_cookies()
if not req.returnOnlyCookies:
challenge_res.headers = {} # todo: fix, selenium not provides this info
challenge_res.response = driver.page_source
challenge_res.userAgent = utils.get_user_agent(driver)
res.result = challenge_res
return res
def _post_request(req: V1RequestBase, driver: WebDriver):
post_form = f'<form id="hackForm" action="{req.url}" method="POST">'
query_string = req.postData if req.postData[0] != '?' else req.postData[1:]
pairs = query_string.split('&')
for pair in pairs:
parts = pair.split('=')
# noinspection PyBroadException
try:
name = unquote(parts[0])
except Exception:
name = parts[0]
if name == 'submit':
continue
# noinspection PyBroadException
try:
value = unquote(parts[1])
except Exception:
value = parts[1]
post_form += f'<input type="text" name="{name}" value="{value}"><br>'
post_form += '</form>'
html_content = f"""
<!DOCTYPE html>
<html>
<body>
{post_form}
<script>document.getElementById('hackForm').submit();</script>
</body>
</html>"""
driver.get("data:text/html;charset=utf-8," + html_content)

360
src/tests.py Normal file
View File

@ -0,0 +1,360 @@
import unittest
from datetime import datetime, timezone
from webtest import TestApp
from dtos import IndexResponse, HealthResponse, V1ResponseBase, STATUS_OK, STATUS_ERROR
import flaresolverr
import utils
def _find_obj_by_key(key: str, value: str, _list: list) -> dict | None:
for obj in _list:
if obj[key] == value:
return obj
return None
class TestFlareSolverr(unittest.TestCase):
proxy_url = "http://127.0.0.1:8888"
proxy_socks_url = "socks5://127.0.0.1:1080"
google_url = "https://www.google.com"
post_url = "https://ptsv2.com/t/qv4j3-1634496523"
cloudflare_url = "https://nowsecure.nl"
cloudflare_url_2 = "https://idope.se/torrent-list/harry/"
ddos_guard_url = "https://anidex.info/"
custom_cloudflare_url = "https://www.muziekfabriek.org"
app = TestApp(flaresolverr.app)
def test_wrong_endpoint(self):
res = self.app.get('/wrong', status=404)
self.assertEqual(res.status_code, 404)
body = res.json
self.assertEqual("Not found: '/wrong'", body['error'])
self.assertEqual(404, body['status_code'])
def test_index_endpoint(self):
res = self.app.get('/')
self.assertEqual(res.status_code, 200)
body = IndexResponse(res.json)
self.assertEqual("FlareSolverr is ready!", body.msg)
self.assertEqual(utils.get_flaresolverr_version(), body.version)
self.assertIn("Chrome/", body.userAgent)
def test_health_endpoint(self):
res = self.app.get('/health')
self.assertEqual(res.status_code, 200)
body = HealthResponse(res.json)
self.assertEqual(STATUS_OK, body.status)
def test_v1_endpoint_wrong_cmd(self):
res = self.app.post_json('/v1', {
"cmd": "request.bad",
"url": self.google_url
}, status=500)
self.assertEqual(res.status_code, 500)
body = V1ResponseBase(res.json)
self.assertEqual(STATUS_ERROR, body.status)
self.assertEqual("Error: Request parameter 'cmd' = 'request.bad' is invalid.", body.message)
self.assertGreater(body.startTimestamp, 10000)
self.assertGreaterEqual(body.endTimestamp, body.startTimestamp)
self.assertEqual(utils.get_flaresolverr_version(), body.version)
def test_v1_endpoint_request_get_no_cloudflare(self):
res = self.app.post_json('/v1', {
"cmd": "request.get",
"url": self.google_url
})
self.assertEqual(res.status_code, 200)
body = V1ResponseBase(res.json)
self.assertEqual(STATUS_OK, body.status)
self.assertEqual("", body.message)
self.assertGreater(body.startTimestamp, 10000)
self.assertGreaterEqual(body.endTimestamp, body.startTimestamp)
self.assertEqual(utils.get_flaresolverr_version(), body.version)
solution = body.solution
self.assertIn(self.google_url, solution.url)
self.assertEqual(solution.status, 200)
self.assertIs(len(solution.headers), 0)
self.assertIn("<title>Google</title>", solution.response)
self.assertGreater(len(solution.cookies), 0)
self.assertIn("Chrome/", solution.userAgent)
def test_v1_endpoint_request_get_cloudflare_js_1(self):
res = self.app.post_json('/v1', {
"cmd": "request.get",
"url": self.cloudflare_url
})
self.assertEqual(res.status_code, 200)
body = V1ResponseBase(res.json)
self.assertEqual(STATUS_OK, body.status)
self.assertEqual("", body.message)
self.assertGreater(body.startTimestamp, 10000)
self.assertGreaterEqual(body.endTimestamp, body.startTimestamp)
self.assertEqual(utils.get_flaresolverr_version(), body.version)
solution = body.solution
self.assertIn(self.cloudflare_url, solution.url)
self.assertEqual(solution.status, 200)
self.assertIs(len(solution.headers), 0)
self.assertIn("<title>nowSecure</title>", solution.response)
self.assertGreater(len(solution.cookies), 0)
self.assertIn("Chrome/", solution.userAgent)
cf_cookie = _find_obj_by_key("name", "cf_clearance", solution.cookies)
self.assertIsNotNone(cf_cookie, "Cloudflare cookie not found")
self.assertGreater(len(cf_cookie["value"]), 30)
def test_v1_endpoint_request_get_cloudflare_js_2(self):
res = self.app.post_json('/v1', {
"cmd": "request.get",
"url": self.cloudflare_url_2
})
self.assertEqual(res.status_code, 200)
body = V1ResponseBase(res.json)
self.assertEqual(STATUS_OK, body.status)
self.assertEqual("", body.message)
self.assertGreater(body.startTimestamp, 10000)
self.assertGreaterEqual(body.endTimestamp, body.startTimestamp)
self.assertEqual(utils.get_flaresolverr_version(), body.version)
solution = body.solution
self.assertIn(self.cloudflare_url_2, solution.url)
self.assertEqual(solution.status, 200)
self.assertIs(len(solution.headers), 0)
self.assertIn("<title>harry - idope torrent search</title>", solution.response)
self.assertGreater(len(solution.cookies), 0)
self.assertIn("Chrome/", solution.userAgent)
cf_cookie = _find_obj_by_key("name", "cf_clearance", solution.cookies)
self.assertIsNotNone(cf_cookie, "Cloudflare cookie not found")
self.assertGreater(len(cf_cookie["value"]), 30)
def test_v1_endpoint_request_get_ddos_guard_js(self):
res = self.app.post_json('/v1', {
"cmd": "request.get",
"url": self.ddos_guard_url
})
self.assertEqual(res.status_code, 200)
body = V1ResponseBase(res.json)
self.assertEqual(STATUS_OK, body.status)
self.assertEqual("", body.message)
self.assertGreater(body.startTimestamp, 10000)
self.assertGreaterEqual(body.endTimestamp, body.startTimestamp)
self.assertEqual(utils.get_flaresolverr_version(), body.version)
solution = body.solution
self.assertIn(self.ddos_guard_url, solution.url)
self.assertEqual(solution.status, 200)
self.assertIs(len(solution.headers), 0)
self.assertIn("<title>AniDex</title>", solution.response)
self.assertGreater(len(solution.cookies), 0)
self.assertIn("Chrome/", solution.userAgent)
cf_cookie = _find_obj_by_key("name", "__ddg1_", solution.cookies)
self.assertIsNotNone(cf_cookie, "DDOS-Guard cookie not found")
self.assertGreater(len(cf_cookie["value"]), 10)
def test_v1_endpoint_request_get_custom_cloudflare_js(self):
res = self.app.post_json('/v1', {
"cmd": "request.get",
"url": self.custom_cloudflare_url
})
self.assertEqual(res.status_code, 200)
body = V1ResponseBase(res.json)
self.assertEqual(STATUS_OK, body.status)
self.assertEqual("", body.message)
self.assertGreater(body.startTimestamp, 10000)
self.assertGreaterEqual(body.endTimestamp, body.startTimestamp)
self.assertEqual(utils.get_flaresolverr_version(), body.version)
solution = body.solution
self.assertIn(self.custom_cloudflare_url, solution.url)
self.assertEqual(solution.status, 200)
self.assertIs(len(solution.headers), 0)
self.assertIn("<title>MuziekFabriek : Aanmelden</title>", solution.response)
self.assertGreater(len(solution.cookies), 0)
self.assertIn("Chrome/", solution.userAgent)
cf_cookie = _find_obj_by_key("name", "ct_anti_ddos_key", solution.cookies)
self.assertIsNotNone(cf_cookie, "Custom Cloudflare cookie not found")
self.assertGreater(len(cf_cookie["value"]), 10)
# todo: test Cmd 'request.get' should return fail with Cloudflare CAPTCHA
# todo: test Cmd 'request.get' should return fail with Cloudflare Blocked
# todo: test Cmd 'request.get' should return OK with 'cookies' param
def test_v1_endpoint_request_get_returnOnlyCookies_param(self):
res = self.app.post_json('/v1', {
"cmd": "request.get",
"url": self.google_url,
"returnOnlyCookies": True
})
self.assertEqual(res.status_code, 200)
body = V1ResponseBase(res.json)
self.assertEqual(STATUS_OK, body.status)
self.assertEqual("", body.message)
self.assertGreater(body.startTimestamp, 10000)
self.assertGreaterEqual(body.endTimestamp, body.startTimestamp)
self.assertEqual(utils.get_flaresolverr_version(), body.version)
solution = body.solution
self.assertIn(self.google_url, solution.url)
self.assertEqual(solution.status, 200)
self.assertIsNone(solution.headers)
self.assertIsNone(solution.response)
self.assertGreater(len(solution.cookies), 0)
self.assertIsNone(solution.userAgent)
# todo: test Cmd 'request.get' should return OK with HTTP 'proxy' param
# todo: test Cmd 'request.get' should return OK with HTTP 'proxy' param with credentials
# todo: test Cmd 'request.get' should return OK with SOCKSv5 'proxy' param
# todo: test Cmd 'request.get' should fail with wrong 'proxy' param
def test_v1_endpoint_request_get_fail_timeout(self):
res = self.app.post_json('/v1', {
"cmd": "request.get",
"url": self.google_url,
"maxTimeout": 10
}, status=500)
self.assertEqual(res.status_code, 500)
body = V1ResponseBase(res.json)
self.assertEqual(STATUS_ERROR, body.status)
self.assertEqual("Error: Error solving the challenge. Timeout after 0.01 seconds.", body.message)
self.assertGreater(body.startTimestamp, 10000)
self.assertGreaterEqual(body.endTimestamp, body.startTimestamp)
self.assertEqual(utils.get_flaresolverr_version(), body.version)
def test_v1_endpoint_request_get_fail_bad_domain(self):
res = self.app.post_json('/v1', {
"cmd": "request.get",
"url": "https://www.google.combad"
}, status=500)
self.assertEqual(res.status_code, 500)
body = V1ResponseBase(res.json)
self.assertEqual(STATUS_ERROR, body.status)
self.assertIn("Message: unknown error: net::ERR_NAME_NOT_RESOLVED", body.message)
def test_v1_endpoint_request_get_deprecated_param(self):
res = self.app.post_json('/v1', {
"cmd": "request.get",
"url": self.google_url,
"userAgent": "Test User-Agent" # was removed in v2, not used
})
self.assertEqual(res.status_code, 200)
body = V1ResponseBase(res.json)
self.assertEqual(STATUS_OK, body.status)
self.assertEqual("", body.message)
def test_v1_endpoint_request_post_no_cloudflare(self):
res = self.app.post_json('/v1', {
"cmd": "request.post",
"url": self.post_url + '/post',
"postData": "param1=value1&param2=value2"
})
self.assertEqual(res.status_code, 200)
body = V1ResponseBase(res.json)
self.assertEqual(STATUS_OK, body.status)
self.assertEqual("", body.message)
self.assertGreater(body.startTimestamp, 10000)
self.assertGreaterEqual(body.endTimestamp, body.startTimestamp)
self.assertEqual(utils.get_flaresolverr_version(), body.version)
solution = body.solution
self.assertIn(self.post_url, solution.url)
self.assertEqual(solution.status, 200)
self.assertIs(len(solution.headers), 0)
self.assertIn("I hope you have a lovely day!", solution.response)
self.assertEqual(len(solution.cookies), 0)
self.assertIn("Chrome/", solution.userAgent)
# check that we sent the post data
res2 = self.app.post_json('/v1', {
"cmd": "request.get",
"url": self.post_url
})
self.assertEqual(res2.status_code, 200)
body2 = V1ResponseBase(res2.json)
self.assertEqual(STATUS_OK, body2.status)
date_hour = datetime.now(timezone.utc).isoformat().split(':')[0].replace('T', ' ')
self.assertIn(date_hour, body2.solution.response)
def test_v1_endpoint_request_post_cloudflare(self):
res = self.app.post_json('/v1', {
"cmd": "request.post",
"url": self.cloudflare_url,
"postData": "param1=value1&param2=value2"
})
self.assertEqual(res.status_code, 200)
body = V1ResponseBase(res.json)
self.assertEqual(STATUS_OK, body.status)
self.assertEqual("", body.message)
self.assertGreater(body.startTimestamp, 10000)
self.assertGreaterEqual(body.endTimestamp, body.startTimestamp)
self.assertEqual(utils.get_flaresolverr_version(), body.version)
solution = body.solution
self.assertIn(self.cloudflare_url, solution.url)
self.assertEqual(solution.status, 200)
self.assertIs(len(solution.headers), 0)
self.assertIn("<title>405 Not Allowed</title>", solution.response)
self.assertGreater(len(solution.cookies), 0)
self.assertIn("Chrome/", solution.userAgent)
cf_cookie = _find_obj_by_key("name", "cf_clearance", solution.cookies)
self.assertIsNotNone(cf_cookie, "Cloudflare cookie not found")
self.assertGreater(len(cf_cookie["value"]), 30)
def test_v1_endpoint_request_post_fail_no_post_data(self):
res = self.app.post_json('/v1', {
"cmd": "request.post",
"url": self.google_url
}, status=500)
self.assertEqual(res.status_code, 500)
body = V1ResponseBase(res.json)
self.assertEqual(STATUS_ERROR, body.status)
self.assertIn("Request parameter 'postData' is mandatory in 'request.post' command", body.message)
def test_v1_endpoint_request_post_deprecated_param(self):
res = self.app.post_json('/v1', {
"cmd": "request.post",
"url": self.google_url,
"postData": "param1=value1&param2=value2",
"userAgent": "Test User-Agent" # was removed in v2, not used
})
self.assertEqual(res.status_code, 200)
body = V1ResponseBase(res.json)
self.assertEqual(STATUS_OK, body.status)
self.assertEqual("", body.message)
# todo: test Cmd 'sessions.create' should return OK
# todo: test Cmd 'sessions.create' should return OK with session
# todo: test Cmd 'sessions.list' should return OK
# todo: test Cmd 'sessions.destroy' should return OK
# todo: test Cmd 'sessions.destroy' should fail
# todo: test Cmd 'request.get' should use session
if __name__ == '__main__':
unittest.main()

97
src/utils.py Normal file
View File

@ -0,0 +1,97 @@
import json
import logging
import os
from selenium.webdriver.chrome.webdriver import WebDriver
import undetected_chromedriver as uc
FLARESOLVERR_VERSION = None
CHROME_MAJOR_VERSION = None
USER_AGENT = None
def get_config_log_html() -> bool:
return os.environ.get('LOG_HTML', 'false').lower() == 'true'
def get_flaresolverr_version() -> str:
global FLARESOLVERR_VERSION
if FLARESOLVERR_VERSION is not None:
return FLARESOLVERR_VERSION
package_path = os.path.join(os.path.dirname(os.path.abspath(__file__)), os.pardir, 'package.json')
with open(package_path) as f:
FLARESOLVERR_VERSION = json.loads(f.read())['version']
return FLARESOLVERR_VERSION
def get_webdriver() -> WebDriver:
logging.debug('Launching web browser...')
# undetected_chromedriver
options = uc.ChromeOptions()
options.add_argument('--no-sandbox')
options.add_argument('--window-size=1920,1080')
# todo: this param shows a warning in chrome headfull
options.add_argument('--disable-setuid-sandbox')
options.add_argument('--disable-dev-shm-usage')
# note: headless mode is detected
# options.headless = True
# if we are inside the Docker container, we avoid downloading the driver
driver_exe_path = None
version_main = None
if os.path.exists("/app/chromedriver"):
driver_exe_path = "/app/chromedriver"
else:
version_main = get_chrome_major_version()
# downloads and patches the chromedriver
# todo: if we don't set driver_executable_path it downloads, patches, and deletes the driver each time
driver = uc.Chrome(options=options, driver_executable_path=driver_exe_path, version_main=version_main)
# selenium vanilla
# options = webdriver.ChromeOptions()
# options.add_argument('--no-sandbox')
# options.add_argument('--window-size=1920,1080')
# options.add_argument('--disable-setuid-sandbox')
# options.add_argument('--disable-dev-shm-usage')
# driver = webdriver.Chrome(options=options)
return driver
def get_chrome_major_version() -> str:
global CHROME_MAJOR_VERSION
if CHROME_MAJOR_VERSION is not None:
return CHROME_MAJOR_VERSION
chrome_path = uc.find_chrome_executable()
# Example 1: 'Chromium 104.0.5112.79 Arch Linux\n'
# Example 2: 'Google Chrome 104.0.5112.79 Arch Linux\n'
process = os.popen(f"{chrome_path} --version")
complete_version = process.read()
process.close()
CHROME_MAJOR_VERSION = complete_version.split('.')[0].split(' ')[-1]
return CHROME_MAJOR_VERSION
def get_user_agent(driver=None) -> str:
global USER_AGENT
if USER_AGENT is not None:
return USER_AGENT
try:
if driver is None:
driver = get_webdriver()
USER_AGENT = driver.execute_script("return navigator.userAgent")
return USER_AGENT
except Exception as e:
raise Exception("Error getting browser User-Agent. " + str(e))
finally:
if driver is not None:
driver.quit()
def object_to_dict(_object):
return json.loads(json.dumps(_object, default=lambda o: o.__dict__))

1
test-requirements.txt Normal file
View File

@ -0,0 +1 @@
WebTest==3.0.0