fix: Spark loader Task not serializable

[haohao0103]

fix #467

[imbajin]

Is it the root problem? And how to ensure it

[haohao0103]

thanks, i think the root cause of the problem is that we later used the specific properties of hugegraphsparkloader when building the vertex and edge data, such as: loadoptions,builders; spark needs to transfer the entire hugegraphsparkloader object over the network, so it requires both the hugegraphsparkloader itself and its properties to implement a serialization interface. However, the ExecutorService type does not implement serialization, so it throws an error that the task cannot be serialized. Using transient allows us to avoid serializing the ExecutorService executor when serializing hugegraphsparkloader, because we will not need the ExecutorService executor during the vertexs/edges construction process after submitting the job. So I think we can solve this problem by adding transient.

[simon824]

make sense

[codecov]

Codecov Report

Merging #471 (345ef1c) into master (55f4282) will not change coverage.
The diff coverage is n/a.

@@            Coverage Diff            @@
##             master     #471   +/-   ##
=========================================
  Coverage     62.51%   62.51%           
  Complexity      894      894           
=========================================
  Files            91       91           
  Lines          4396     4396           
  Branches        516      516           
=========================================
  Hits           2748     2748           
  Misses         1445     1445           
  Partials        203      203           

Impacted Files	Coverage Δ
...e/hugegraph/loader/spark/HugeGraphSparkLoader.java	0.00% <ø> (ø)

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more