Files
dify/api/controllers/console/datasets/website.py

111 lines
4.4 KiB
Python

from typing import Literal
from flask_restx import Resource
from pydantic import BaseModel, Field, RootModel
from controllers.common.schema import query_params_from_model, register_response_schema_models, register_schema_models
from controllers.console import console_ns
from controllers.console.datasets.error import WebsiteCrawlError
from controllers.console.wraps import account_initialization_required, model_validate, setup_required
from extensions.ext_application_services import application_services
from fields.base import ResponseModel
from libs.external_api import error_responses
from libs.helper import dump_response
from libs.login import login_required
from services.data_source.entities.website import CrawledPage, ScrapedPage
from services.data_source.website_service import WebsiteCrawlApiRequest, WebsiteCrawlStatusApiRequest
class WebsiteCrawlPayload(BaseModel):
provider: Literal["firecrawl", "watercrawl", "jinareader"]
url: str
options: dict[str, object]
class WebsiteCrawlStatusQuery(BaseModel):
provider: Literal["firecrawl", "watercrawl", "jinareader"]
class WebsiteCrawlStartedResponse(ResponseModel):
status: str
job_id: str
class WebsitePageCrawledResponse(ResponseModel):
status: str
data: ScrapedPage
class WebsiteCrawlResponse(RootModel[WebsiteCrawlStartedResponse | WebsitePageCrawledResponse]):
"""A remote crawl job or the immediately available single-page content."""
class WebsiteCrawlProgressResponse(ResponseModel):
status: str
job_id: str
total: int
current: int
data: list[CrawledPage]
time_consuming: str | float | None = Field(default=None, exclude_if=lambda value: value is None)
register_schema_models(console_ns, WebsiteCrawlPayload, WebsiteCrawlStatusQuery)
register_response_schema_models(console_ns, WebsiteCrawlResponse, WebsiteCrawlProgressResponse)
@console_ns.route("/website/crawl")
class WebsiteCrawlApi(Resource):
@error_responses(WebsiteCrawlError)
@console_ns.doc("crawl_website")
@console_ns.doc(description="Crawl website content")
@console_ns.expect(console_ns.models[WebsiteCrawlPayload.__name__])
@console_ns.response(200, "Website crawl initiated successfully", console_ns.models[WebsiteCrawlResponse.__name__])
@console_ns.response(400, "Invalid crawl parameters")
@setup_required
@login_required
@account_initialization_required
@model_validate(WebsiteCrawlPayload)
def post(self, req_data: WebsiteCrawlPayload) -> tuple[dict[str, object], int]:
# Create typed request and validate
try:
api_request = WebsiteCrawlApiRequest.from_args(req_data.model_dump())
except ValueError as e:
raise WebsiteCrawlError(str(e)) from e
# Crawl URL using typed request
try:
result = application_services().data_sources.website.crawl_url(api_request)
except Exception as e:
raise WebsiteCrawlError(str(e)) from e
return dump_response(WebsiteCrawlResponse, result, exclude_unset=True), 200
@console_ns.route("/website/crawl/status/<string:job_id>")
class WebsiteCrawlStatusApi(Resource):
@error_responses(WebsiteCrawlError)
@console_ns.doc("get_crawl_status")
@console_ns.doc(description="Get website crawl status")
@console_ns.doc(params={"job_id": "Crawl job ID", "provider": "Crawl provider (firecrawl/watercrawl/jinareader)"})
@console_ns.doc(params=query_params_from_model(WebsiteCrawlStatusQuery))
@console_ns.response(
200, "Crawl status retrieved successfully", console_ns.models[WebsiteCrawlProgressResponse.__name__]
)
@console_ns.response(400, "Invalid provider")
@setup_required
@login_required
@account_initialization_required
@model_validate(WebsiteCrawlStatusQuery)
def get(self, req_data: WebsiteCrawlStatusQuery, job_id: str) -> tuple[dict[str, object], int]:
# Create typed request and validate
try:
api_request = WebsiteCrawlStatusApiRequest.from_args(req_data.model_dump(), job_id)
except ValueError as e:
raise WebsiteCrawlError(str(e)) from e
# Get crawl status using typed request
try:
result = application_services().data_sources.website.get_crawl_status_typed(api_request)
except Exception as e:
raise WebsiteCrawlError(str(e)) from e
return dump_response(WebsiteCrawlProgressResponse, result), 200