feat: add async scan

[Fengzdadi]

What type of PR is this?

• bug
• feature
• enhancement

What problem(s) does this PR solve?

Issue(s) number:

N/A

Description:

• Added two new asynchronous scan APIs for Python clients: scan_vertex_async() and scan_edge_async()
• These methods enable multi‑partition concurrency and streaming batch retrieval of vertices and edges in an asyncio‑friendly way, filling the gap where only blocking, single‑threaded scans existed.

How do you solve it?

• Use concurrent.futures.ThreadPoolExecutor to submit one scan_*_with_part job per partition (via meta_cache.get_part_leaders), dispatching them in parallel.
• As each future completes, iterate its ScanResult in a while scan_result.has_next(): yield part, batch loop to stream (partition_id, batch) pairs.

Special notes for your reviewer, ex. impact of this fix, design document, etc:

• Verified on NebulaGraph 3.8 with manual testing.

[Malone-AI]

In the scan_vertex_async function,

scan_result = future.result()  

may throw an exception but is not handled.

In the scan_edge_async function,

while scan_result and scan_result.has_next():  
    batch = scan_result.next()  

If batch = scan_result.next() throws an exception, the scan_result may not be closed, leading to improper resource cleanup.

The comment may be incorrect:
"""
:type prop_names: if given empty, return all property # Should be :param
"""

Perhaps it should be:
"""
:param prop_names: if given empty, return all property
"""

[Copilot]

Pull Request Overview

This PR adds asynchronous scanning capabilities to the GraphStorageClient by introducing two new methods: scan_vertex_async() and scan_edge_async(). These methods enable concurrent, multi-partition scanning of vertices and edges using thread pools, addressing the limitation of only having blocking, single-threaded scan operations.

• Added concurrent.futures import for thread pool execution
• Implemented scan_vertex_async() method for parallel vertex scanning across partitions
• Implemented scan_edge_async() method for parallel edge scanning across partitions

[Copilot]

The future.result() call can raise exceptions from the background thread, but there's no exception handling. Consider wrapping this in a try-except block to handle potential errors gracefully, especially since the TODO comment mentions error handling as a known issue.

[Copilot]

Same as in scan_vertex_async: the future.result() call can raise exceptions from the background thread, but there's no exception handling. Consider wrapping this in a try-except block to handle potential errors gracefully.

[Fengzdadi]

Tks to @Malone-AI @wey-gu, I will fix these errors in the next few days.