bobseares-2/seleniumbase-website-scraper
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

write agent.py

Browse Source
This commit is contained in:
a2a-cloud
2026-06-08 22:49:08 +00:00
parent 8fd488af36
commit c1e9063165
1 changed files with 8 additions and 8 deletions
Download Patch File Download Diff File Expand all files Collapse all files

16
agent.py
View File

@@ -248,15 +248,16 @@ class SeleniumbaseWebsiteScraper(A2AAgent[SeleniumbaseWebsiteScraperConfig, NoAu
grant_write_prefixes=(OUTPUT_DIR,), grant_write_prefixes=(OUTPUT_DIR,),
grant_run_timeout_seconds=900, grant_run_timeout_seconds=900,
) )
async def iterate_scrape_goal( async def iterative_scrape_goal(
self, self,
ctx: RunContext[NoAuth], ctx: RunContext[NoAuth],
goal: str, goal: str,
url: str, start_url: str,
selectors: dict[str, str] | None = None, allowed_domains: list[str] | None = None,
min_items: int = 10, seed_selectors: dict[str, str] | None = None,
max_iterations: int = 5, max_steps: int = 2,
per_pass_max_pages: int = 1, max_pages: int = 1,
max_runtime_seconds: int = 600,
wait_seconds: float = 3.0, wait_seconds: float = 3.0,
output_format: Literal["json", "csv"] = "json", output_format: Literal["json", "csv"] = "json",
headless: bool = True, headless: bool = True,
@@ -264,10 +265,9 @@ class SeleniumbaseWebsiteScraper(A2AAgent[SeleniumbaseWebsiteScraperConfig, NoAu
respect_robots: bool = True, respect_robots: bool = True,
rate_limit_seconds: float = 1.0, rate_limit_seconds: float = 1.0,
screenshot: bool = False, screenshot: bool = False,
allow_selector_updates: bool = True,
) -> dict[str, Any]: ) -> dict[str, Any]:
"""Use an LLM planner to refine selectors/pagination until the goal is met or a stop condition occurs.""" """Use an LLM planner to refine selectors/pagination until the goal is met or a stop condition occurs."""
start_url = str(url or "").strip() start_url = str(start_url or "").strip()
if not start_url: if not start_url:
return {"status": "invalid_input", "items": [], "saved_paths": [], "warnings": ["url is required"], "screenshots": []} return {"status": "invalid_input", "items": [], "saved_paths": [], "warnings": ["url is required"], "screenshots": []}