Add more logic on  duplicate pdf detection

The duplicate pdf detection currently relies on the file path,
this CL adds more logic to it:
1. the http method has to be GET.
2. There are no past download with the same file path and different
  URL.
3. There is a past download with the same URL and file path.

Bug: 415572411, 416023683
Change-Id: I48d840dc7e4dc6ccc2002fe39792e99cf69e9572
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6514053
Reviewed-by: Shu Yang <shuyng@google.com>
Commit-Queue: Shu Yang <shuyng@google.com>
Cr-Commit-Position: refs/heads/main@{#1456696}