Files
robotex/src/main.py
T

164 lines
4.4 KiB
Python

import json
import logging
import traceback
import uuid
from pathlib import Path
from typing import Annotated
from urllib.parse import urlsplit
from camoufox.async_api import AsyncCamoufox
from camoufox.fingerprints import generate_context_fingerprint
from camoufox.utils import determine_ua_os
from fastapi import Cookie
from fastapi import FastAPI
from fastapi import Request
from fastapi import Response
from playwright.async_api import ProxySettings
from ua_parser import user_agent_parser
from engine import actions
from engine.schemas import SolveRequest
logger = logging.getLogger(__name__)
logger.setLevel(logging.DEBUG)
logger.addHandler(logging.StreamHandler())
BROWSER_SLOTS = 10
IDENTITY_FILE = Path("identity.json")
def parse_proxy(proxy_url: str) -> ProxySettings:
parts = urlsplit(proxy_url)
server = f"{parts.scheme}://{parts.hostname}:{parts.port}"
settings: ProxySettings = {"server": server}
if parts.username:
settings["username"] = parts.username
if parts.password:
settings["password"] = parts.password
return settings
def load_or_create_identity(user_data_dir, os=None, ff_version=None, locale=None, time_zone=None, user_agent=None) -> dict:
identity_file = user_data_dir / IDENTITY_FILE
if identity_file.exists():
return json.loads(identity_file.read_text())
kwargs = {}
if user_agent:
os, ff_version = parse_os_and_ff(user_agent)
kwargs["config_overrides"] = {"navigator.UserAgent": user_agent}
fp = generate_context_fingerprint(os=os, ff_version=ff_version, locale=locale, timezone=time_zone)
identity_file.write_text(json.dumps(fp))
return fp
app = FastAPI()
@app.get("/")
async def root():
return {"message": "Hello World"}
_OS_MAP = {'win': 'windows', 'mac': 'macos', 'lin': 'linux'}
def parse_os_and_ff(user_agent) -> tuple[str, str]:
os_name = _OS_MAP[determine_ua_os(user_agent)]
parsed = user_agent_parser.ParseUserAgent(user_agent)
ff_version = int(parsed["major"]) if (
parsed.get("family") == "Firefox" and parsed.get("major")) else 150
return os_name, str(ff_version)
@app.post("/v1/solve")
async def solve(
request: Request,
payload: SolveRequest,
response: Response,
session_id: Annotated[str | None, Cookie()] = None
):
api_key = request.headers.get('X-API-Key')
logger.info("Incoming payload: [%s] %s", session_id, payload)
global BROWSER_SLOTS
if BROWSER_SLOTS <= 0:
return {
"ok": False,
"error": "No slots available",
}
else:
BROWSER_SLOTS -= 1
if not session_id:
session_id = str(uuid.uuid4())
response.set_cookie("session_id", session_id)
logger.info("Setup new session with session_id: %s", session_id)
scr_w, scr_h = payload.screen.split("x")
scr_size = int(scr_w), int(scr_h)
proxy = parse_proxy(payload.proxy)
user_data_dir = Path(__file__).parent / f"../extra/user_data_dir/{session_id}"
if not user_data_dir.exists():
user_data_dir.mkdir(parents=True)
config = load_or_create_identity(user_data_dir, user_agent=payload.user_agent)
ctx_opts = {**config["context_options"], "config": config["config"]}
async with AsyncCamoufox(
persistent_context=True,
user_data_dir=user_data_dir,
proxy=proxy,
window=scr_size,
i_know_what_im_doing=True,
geoip=True,
humanize=True,
main_world_eval=True,
**ctx_opts,
) as camoufo:
await camoufo.add_init_script(config["init_script"])
page = camoufo.pages[0] if camoufo.pages else await camoufo.new_page()
if payload.cookies:
await camoufo.add_cookies(payload.cookies)
page.set_default_timeout(30_000)
page.set_default_navigation_timeout(60_000)
await page.goto(url=payload.url)
actions_result = await actions.execute(
page, payload.actions, payload.captcha,
antibot_name=payload.captcha_type,
success_locator=payload.captcha_success_locator,
detect_locator=payload.captcha_detect_locator,
)
status = not "error" in actions_result[-1]
error = ["Something error", None][status]
cookies = await camoufo.cookies()
html = await page.inner_html('html')
# TODO: добавить как опцию. передачу скрина в base64.
# await page.screenshot(
# path=(
# f"/home/sokol/PycharmProjects/WebRoboApi/extra/screenshots/"
# f"{datetime.timestamp(datetime.now())}.png"
# )
# )
BROWSER_SLOTS += 1
result = {
"status": ["err", "ok"][status],
"user_agent": ctx_opts["user_agent"],
"cookies": cookies,
"actions": actions_result,
}
if payload.html:
result["html"] = html
if error:
result["error"] = error
logger.info("Outgoing payload: %s", result)
return result