[mahi] content extraction service.

This CL creates the utility process for mahi to extract content from
A11y tree snapshot. As the snapshot can come from any web page, it is
untrustworthy and should be processed in the utility process.

Currently it provides two functions:
1. Returns the content of the snapshot.
2. Returns the content word count of the snapshot.
The content extraction service distills the important content and
returns based on the requirement.

Mahi design doc: http://go/mahi-dd
Content extraction design doc: go/mahi-content-extraction

and will be covered when the other pipe end is implemented. Other files
has been well covered.

Low-Coverage-Reason: TESTS_IN_SEPARATE_CL The service change is trivial
Bug: b/318565573
Change-Id: I8baf8b2b9cb9251d8a94cc5fc13dcb32bd6641a7
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5256918
Auto-Submit: CJ Huang <chenjih@google.com>
Reviewed-by: Abigail Klein <abigailbklein@google.com>
Reviewed-by: Ramin Halavati <rhalavati@chromium.org>
Reviewed-by: Tony Yeoman <tby@chromium.org>
Reviewed-by: Sam McNally <sammc@chromium.org>
Reviewed-by: Erik Chen <erikchen@chromium.org>
Commit-Queue: CJ Huang <chenjih@google.com>
Cr-Commit-Position: refs/heads/main@{#1261401}