mirror of
https://github.com/j93es/browser-use-oauth.git
synced 2026-06-04 08:21:52 +09:00
[Update] agent 호출 구조 변경
This commit is contained in:
tv0924@icloud.com
parent
495b3a52da
commit
b7e6afb227
12 changed files with 267 additions and 343 deletions
|
|
@ -1,92 +0,0 @@
|
|||
import json
|
||||
from pydantic import BaseModel
|
||||
from browser_use import (
|
||||
Agent,
|
||||
Controller,
|
||||
)
|
||||
from lib.agents.run_agent import run_agent
|
||||
from lib.utils.logger import logger
|
||||
from lib.browser_use_utils.create_google_ai import create_google_ai
|
||||
from lib.config import GOOGLE_MODEL, GOOGLE_PLANNER_MODEL
|
||||
|
||||
NOT_FOUND_LOGIN_PAGE = 0
|
||||
FOUND_LOGIN_PAGE = 1
|
||||
|
||||
class FindLoginPageResponse(BaseModel):
|
||||
status: int = NOT_FOUND_LOGIN_PAGE # 0 if not found, 1 if found
|
||||
msg: str | None = None
|
||||
url: str | None = None
|
||||
|
||||
async def find_login_page(target_url, session) -> tuple[bool, str | None]:
|
||||
initial_actions = [{"open_tab": {"url": target_url}}]
|
||||
task = """
|
||||
You are an expert in finding login pages.
|
||||
|
||||
Your task is to navigate to the login page of the given URL. Follow the steps below strictly and return results only in the specified format.
|
||||
|
||||
※ You are NOT allowed to navigate to URLs that are not directly discoverable within the initial domain. Do NOT use search engines or guess external login URLs.
|
||||
|
||||
0. INITIAL BLOCK CHECK
|
||||
- If the browser is blocked when trying to access the page — due to firewall, CAPTCHA, regional restrictions, or other access denials — immediately terminate the process and return the following JSON:
|
||||
```json
|
||||
{
|
||||
"status": 0,
|
||||
"msg": "Blocked",
|
||||
"url": ""
|
||||
}
|
||||
```
|
||||
- Do NOT proceed to further steps in this case.
|
||||
|
||||
1. LOGIN PAGE NAVIGATION
|
||||
- Navigate only to a **client-side (non-enterprise)** login page within the provided domain.
|
||||
- Do NOT rely on external tools, search engines, or links not directly found on the site.
|
||||
- If a consent popup (e.g. for privacy/cookies) appears, you MUST dismiss or close it before proceeding.
|
||||
- Since step 0 confirmed access, assume the page now loads properly.
|
||||
|
||||
2. RETURN FORMAT
|
||||
- Once the login page is reached, return a JSON object matching the following schema:
|
||||
```json
|
||||
{
|
||||
"status": 1, // 1 if login page is found, 0 otherwise
|
||||
"msg": "Login page found", // Optional message
|
||||
"url": "https://example.com/login" // Full URL of the login page if found
|
||||
}
|
||||
```
|
||||
- If the login page cannot be found, return:
|
||||
```json
|
||||
{
|
||||
"status": 0,
|
||||
"msg": "Login page not found",
|
||||
"url": ""
|
||||
}
|
||||
```
|
||||
- Return ONLY the JSON object. Do NOT include any explanation, logging, or extra output.
|
||||
"""
|
||||
|
||||
|
||||
controller = Controller(output_model=FindLoginPageResponse, exclude_actions=['search_google'])
|
||||
agent = Agent(
|
||||
browser_session=session,
|
||||
initial_actions=initial_actions,
|
||||
task=task,
|
||||
llm=create_google_ai(GOOGLE_MODEL),
|
||||
controller=controller,
|
||||
)
|
||||
|
||||
is_failed, final_result = await run_agent(agent)
|
||||
if is_failed:
|
||||
logger(f"⚠️ 스캔 실패: {target_url} | {final_result}")
|
||||
print(f"⚠️ 스캔 실패: {target_url} | {final_result}")
|
||||
return False, None;
|
||||
|
||||
data = json.loads(final_result)
|
||||
try:
|
||||
resp = FindLoginPageResponse(**data)
|
||||
if resp.status == FOUND_LOGIN_PAGE and len(resp.url) > 0:
|
||||
return True, resp.url
|
||||
else:
|
||||
return False, resp.msg
|
||||
except Exception as e:
|
||||
logger(f"⚠️ 결과 파싱 실패: {target_url} | {e}\n원본 결과: {data.msg}")
|
||||
print(f"⚠️ 결과 파싱 실패: {target_url} | {e}\n원본 결과: {data.msg}")
|
||||
return False, data.msg
|
||||
|
|
@ -1,66 +0,0 @@
|
|||
import json
|
||||
from pydantic import BaseModel
|
||||
from browser_use import (
|
||||
Agent,
|
||||
Controller,
|
||||
)
|
||||
from lib.agents.run_agent import run_agent
|
||||
from lib.utils.logger import logger
|
||||
from lib.browser_use_utils.create_google_ai import create_google_ai
|
||||
from lib.config import GOOGLE_MODEL, GOOGLE_PLANNER_MODEL
|
||||
|
||||
NOT_FOUND_SSO_LIST = 0
|
||||
FOUND_SSO_LIST = 1
|
||||
|
||||
class EachSSOProvider(BaseModel):
|
||||
provider: str
|
||||
oauth_uri: str | None = None
|
||||
|
||||
class FindLoginPageResponse(BaseModel):
|
||||
EachSSOProviders: list[EachSSOProvider] | None = None
|
||||
status: int = NOT_FOUND_SSO_LIST # 0 if not found,
|
||||
msg: str | None = None
|
||||
|
||||
async def get_sso_list(target_url, session) -> tuple[bool, str | None]:
|
||||
initial_actions = [{"open_tab": {"url": target_url}}]
|
||||
task = "Navigate to the login page, and return the result in the specified format."
|
||||
extend_planner_system_message = """
|
||||
You are an expert in finding login pages.
|
||||
Your task is to navigate to the login page of the given URL.
|
||||
Once you reach the login page, stop and return a JSON object that matches the following schema:
|
||||
```json
|
||||
{
|
||||
"status": 1, # 1 if login page found, 0 otherwise
|
||||
"url": "https://example.com/login" # Full URL of the login page if found
|
||||
}
|
||||
Return only this JSON object. Do not include any explanation or additional text.
|
||||
"""
|
||||
|
||||
controller = Controller(output_model=FindLoginPageResponse, exclude_actions=['search_google'])
|
||||
agent = Agent(
|
||||
browser_session=session,
|
||||
initial_actions=initial_actions,
|
||||
task=task,
|
||||
llm=create_google_ai(GOOGLE_MODEL),
|
||||
planner_llm=create_google_ai(GOOGLE_PLANNER_MODEL),
|
||||
controller=controller,
|
||||
extend_planner_system_message=extend_planner_system_message,
|
||||
)
|
||||
|
||||
is_failed, final_result = await run_agent(agent)
|
||||
if is_failed:
|
||||
logger(f"⚠️ 스캔 실패: {target_url} | {final_result}")
|
||||
print(f"⚠️ 스캔 실패: {target_url} | {final_result}")
|
||||
return False, None;
|
||||
|
||||
data = json.loads(final_result)
|
||||
try:
|
||||
resp = FindLoginPageResponse(**data)
|
||||
if resp.status == FOUND_SSO_LIST:
|
||||
return True, resp
|
||||
else:
|
||||
return False, None
|
||||
except Exception as e:
|
||||
logger(f"⚠️ 결과 파싱 실패: {target_url} | {e}\n원본 결과: {data.msg}")
|
||||
print(f"⚠️ 결과 파싱 실패: {target_url} | {e}\n원본 결과: {data.msg}")
|
||||
return False, data.msg
|
||||
|
|
@ -1,19 +0,0 @@
|
|||
from lib.browser_use_utils.clean_resources import clean_agent_resources
|
||||
|
||||
async def run_agent(agent) -> tuple[int, str]:
|
||||
try:
|
||||
response = await agent.run()
|
||||
final_result = response.final_result()
|
||||
|
||||
if final_result is None:
|
||||
return -1, "최종 결과가 없습니다. 에이전트 실행 실패"
|
||||
return 0, final_result
|
||||
except Exception as e:
|
||||
# API 쿼터 문제인지 확인
|
||||
if "ResourceExhausted" in str(e) or "429" in str(e):
|
||||
return 1, "API 쿼터 에러로 인한 실패"
|
||||
# 일반 에러 처리
|
||||
else:
|
||||
return 2, "일반 에러로 인한 실패"
|
||||
finally:
|
||||
await clean_agent_resources(agent)
|
||||
67
lib/agents/run_task.py
Normal file
67
lib/agents/run_task.py
Normal file
|
|
@ -0,0 +1,67 @@
|
|||
import json
|
||||
from typing import Any
|
||||
from pydantic import BaseModel
|
||||
from browser_use import (
|
||||
Agent,
|
||||
Controller,
|
||||
BrowserSession
|
||||
)
|
||||
from patchright.async_api import async_playwright as async_patchright
|
||||
from lib.utils.logger import logger
|
||||
from lib.prompt.get_sso_list import get_sso_list_task
|
||||
from lib.browser_use_utils.create_google_ai import create_google_ai
|
||||
from lib.browser_use_utils.get_profile import get_profile
|
||||
from lib.browser_use_utils.clean_resources import clean_session_resources, clean_agent_resources
|
||||
from lib.config import GOOGLE_MODEL
|
||||
|
||||
|
||||
async def run_task(target_url: str, ReturnModel: type[BaseModel], task: str) -> tuple[bool, str | Any | None]:
|
||||
session = BrowserSession(
|
||||
playwright=(await async_patchright().start()),
|
||||
browser_profile=await get_profile(),
|
||||
)
|
||||
|
||||
initial_actions = [{"open_tab": {"url": target_url}}]
|
||||
|
||||
controller = Controller(output_model=ReturnModel, exclude_actions=['search_google'])
|
||||
agent = Agent(
|
||||
browser_session=session,
|
||||
initial_actions=initial_actions,
|
||||
task=task,
|
||||
llm=create_google_ai(GOOGLE_MODEL),
|
||||
controller=controller,
|
||||
)
|
||||
|
||||
try:
|
||||
response = await agent.run()
|
||||
final_result = response.final_result()
|
||||
|
||||
if final_result is None:
|
||||
logger(f"⚠️ 최종 결과가 없습니다. 에이전트 실행 실패: {target_url}")
|
||||
print(f"⚠️ 최종 결과가 없습니다. 에이전트 실행 실패: {target_url}")
|
||||
return False, "최종 결과가 없습니다. 에이전트 실행 실패"
|
||||
except Exception as e:
|
||||
# API 쿼터 문제인지 확인
|
||||
if "ResourceExhausted" in str(e) or "429" in str(e):
|
||||
logger(f"⚠️ API 쿼터 에러로 인한 실패: {target_url} | {e}")
|
||||
print(f"⚠️ API 쿼터 에러로 인한 실패: {target_url} | {e}")
|
||||
return False, "API 쿼터 에러로 인한 실패"
|
||||
# 일반 에러 처리
|
||||
else:
|
||||
logger(f"⚠️ 일반 에러로 인한 실패: {target_url} | {e}")
|
||||
print(f"⚠️ 일반 에러로 인한 실패: {target_url} | {e}")
|
||||
return False, "일반 에러로 인한 실패"
|
||||
finally:
|
||||
await clean_agent_resources(agent)
|
||||
|
||||
try:
|
||||
data = json.loads(final_result)
|
||||
resp = ReturnModel(**data)
|
||||
return True, resp
|
||||
except Exception as e:
|
||||
logger(f"⚠️ LLM 응답 결과 파싱 실패: {target_url} | {e}\n원본 결과: {data.msg}")
|
||||
print(f"⚠️ LLM 응답 결과 파싱 실패: {target_url} | {e}\n원본 결과: {data.msg}")
|
||||
return False, "LLM 응답 결과 파싱 실패"
|
||||
finally:
|
||||
await clean_session_resources(session)
|
||||
|
||||
Loading…
Add table
Add a link
Reference in a new issue