Optimize purview search logic

[enya-yx]

Description

• Improves 'get_project' and similar methods' query logic to reduce time cost

• The main reason that causes the slowness is we called purview APIs ( such as 'AtlasClient.get_entity') many times, especially in some recursion. It leads to a lot of HTTP requests so that we need to wait for relatively long time for their responses. I found by calling the API 'AtlasClient.get_entity_lineage' we can get all information we need. Then we just need to build edges and entities including their relationships and attributes based on the returned results which costs less time comparing with waiting for many responses from purview.

• Test data: (tested from backend)
  registry = PurviewRegistry(azure_purview_name = "feathrazuretest3-purview1")
  entity_id_by_name = registry.get_entity_id("enya_test_registry")
  start = time.time()
  project_pre = registry.get_project_origin(entity_id) // previous 'get_project' API
  print("duration1: ", time.time()-start)
  start = time.time()
  project_curr = registry.get_project(entity_id) // current 'get_project' API
  print("duration2: ", time.time()-start)

  duration1: 33.04559826850891
  duration2: 2.3997888565063477

How was this patch tested?

Does this PR introduce any user-facing changes?

• No. You can skip the rest of this section.
• Yes. Make sure to clarify your proposed changes.

Dependencies

• No. You can skip the rest of this section.

• Yes. Make sure to list all the dependencies and licenses.

[xiaoyongzhu]

Thanks @enya0405 . can you update the PR description to reflect:

What was the previous issue that caused it to be slow and what's the improved way?

Also can you put some numbers on the updated logic (like how much time does it take to load the projects)?