[PDF Ink Signatures] Re-work backend -> frontend API

Clicks will be handled by the PDF internal plugin and forwarded to
the PDF viewer. It is the PDF viewer's responsibility to initialize a
text annotation based on these clicks.

This CL:
- Removes handling for the 'updateTextBoxRect' message, which is now no
  longer planned.
- Adds click forwarding from the internal plugin to the viewer.
- Adds a method on Ink2Manager to initialize a text annotation based
  on the click position; if click is not on a page, click is ignored.
- To support this, at a new getPageAtPoint() method on the viewport
  and corresponding tests.
- Note: Ink2Manager sends full information (including styles,
  page number, etc) about the annotation to be edited in the
  initialization event, as opposed to only location information as
  in the earlier prototype which was based on the never-implemented
  backend updateTextBoxRect message.
- Note 2: Ink2Manager initializes the text box using screen
  coordinates instead of page coordinates.

This CL does not add:
- Logic for determining whether the click is on an existing or new
  annotation
- Page boundary enforcement for annotations.
- Updates to the frontend -> backend messages.

Bug: 402546155
Change-Id: Ic2a5cb9979db00bc75d6d41b5078bf154d219b14
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/6499128
Commit-Queue: Rebekah Potter <rbpotter@chromium.org>
Reviewed-by: Andy Phan <andyphan@chromium.org>
Cr-Commit-Position: refs/heads/main@{#1454553}