Use Shx Index to split large shapefiles

[halfabrane]

This PR allows the Shapefile Reader to split large shapefiles.
The Shapefile Reader looks for a Shx Index File for each (.shp) file. If the index file exists, we use it to split the shape file roughly at block boundaries as follows:

• If you want to control the splitSize, use Hadoop Configuration parameters "mapreduce.input.fileinputformat.split.minsize"(defaults to 1) and "mapreduce.input.fileinputformat.split.maxsize" (defaults to Long.MaxValue)
• We control the splitSize as follows:
  splitSize = Math.max(minSplitSize, Math.min(maxSplitSize, blockSize))
• By default, splitSize = blockSize

[codecov-io]

Codecov Report

Merging #146 into master will increase coverage by 0.46%.
The diff coverage is 93.54%.

@@            Coverage Diff             @@
##           master     #146      +/-   ##
==========================================
+ Coverage   85.56%   86.03%   +0.46%     
==========================================
  Files          48       49       +1     
  Lines        1372     1446      +74     
  Branches       96      100       +4     
==========================================
+ Hits         1174     1244      +70     
- Misses        198      202       +4

Impacted Files	Coverage Δ
src/main/scala/magellan/io/ShapeWritable.scala	55.55% <ø> (-4.45%)	⬇️
src/main/scala/magellan/ShapefileRelation.scala	95.83% <100%> (+0.59%)	⬆️
...ain/scala/magellan/mapreduce/ShapefileReader.scala	96.15% <88.88%> (-0.91%)	⬇️
...in/scala/magellan/mapreduce/ShapeInputFormat.scala	90.62% <90.32%> (-9.38%)	⬇️
...main/scala/magellan/mapreduce/ShxInputFormat.scala	96% <96%> (ø)
src/main/scala/magellan/Polygon.scala	80.64% <0%> (+0.8%)	⬆️

---

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 d13db84...c8f97a7. Read the comment docs.