Repository navigation
Pass router as request_handler to crawler - is this intended? #1600
Description
Activity
It is intended, but it can be re-evaluated for sure. Originally I wanted to be as permissive as possible and accept even simple callables in the
request_handler.I understand, and we can keep the current behavior if it makes sense. But I believe the name should reflect the most common use case, which is registering a router. Registering a single handler can be done without passing it as an argument, and throughout our documentation, we consistently use the default router's request-handler decorator for that purpose. So I believe most people will do just that.
Well, I believe the most common case is that people just use the default router and never touch this parameter:
crawler = BasicCrawler() @crawler.router.default_handler async def handle(context: BasicCrawlingContext): ...
Yes, I agree - registering the handler through the router's default handler is the most common usage. And right after that would come using the router instance and passing it as an argument. And only as the least common would be passing the handler itself directly as an argument. So we should name the argument according to its most common usage pattern.
For example:
Shouldn't the arg be named
router?We would also need to update its type and this part of the BasicCrawler's init: