support custom partitioner for nebula when generate sst files

[Nicole00]

1. whether to use custom partitioner is configurable.
2. use custom partitioner make sure the keys in different sst files does not overlap.
3. When ingest sst files generated with custom partitioner, all most sst files lies on L6 (space is empty before ingest).
   

add config for each tag or edge in the config file:
repartitionWithNebula:false/true ， default is false.

[Nicole00]

close #46

[codecov-commenter]

Codecov Report

Merging #49 (ebec49c) into master (d31546b) will increase coverage by 4.40%.
The diff coverage is 69.44%.

@@             Coverage Diff              @@
##             master      #49      +/-   ##
============================================
+ Coverage     50.19%   54.60%   +4.40%     
- Complexity       74       76       +2     
============================================
  Files            16       17       +1     
  Lines          1291     1315      +24     
  Branches        246      249       +3     
============================================
+ Hits            648      718      +70     
+ Misses          525      472      -53     
- Partials        118      125       +7     

Impacted Files	Coverage Δ
...soft/exchange/common/utils/NebulaPartitioner.scala	11.11% <11.11%> (ø)
...om/vesoft/exchange/common/config/SinkConfigs.scala	73.33% <66.66%> (-3.59%)	⬇️
...m/vesoft/exchange/common/processor/Processor.scala	67.42% <71.42%> (+0.22%)	⬆️
...la/com/vesoft/exchange/common/config/Configs.scala	66.15% <100.00%> (+0.45%)	⬆️
.../vesoft/exchange/common/config/SchemaConfigs.scala	71.87% <100.00%> (+0.90%)	⬆️
...vesoft/exchange/common/writer/FileBaseWriter.scala	85.71% <100.00%> (+85.71%)	⬆️
...a/com/vesoft/exchange/common/utils/HDFSUtils.scala	27.58% <0.00%> (+27.58%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update d31546b...ebec49c. Read the comment docs.

[critical27]

Well done

[darionyaphet]

why don't use repartition directly?

[Nicole00]

why don't use repartition directly?

Dataframe doesn't have customed repartition function, it's RDD's function.