Closed JulienPeloton closed 5 years ago
Merging #110 into master will increase coverage by
2.8%
. The diff coverage is92.85%
.
@@ Coverage Diff @@
## master #110 +/- ##
=========================================
+ Coverage 86.84% 89.64% +2.8%
=========================================
Files 29 32 +3
Lines 1140 1178 +38
Branches 201 194 -7
=========================================
+ Hits 990 1056 +66
+ Misses 150 122 -28
Flag | Coverage Δ | |
---|---|---|
#python | 91.24% <ø> (-2.4%) |
:arrow_down: |
#scala | 88.95% <92.85%> (+4.88%) |
:arrow_up: |
Impacted Files | Coverage Δ | |
---|---|---|
src/main/scala/com/spark3d/Repartitioning.scala | 80.43% <ø> (ø) |
:arrow_up: |
...main/scala/com/spark3d/python/PythonClassTag.scala | 100% <100%> (ø) |
:arrow_up: |
...com/spark3d/utils/BoundedUniquePriorityQueue.scala | 90.47% <100%> (+90.47%) |
:arrow_up: |
...c/main/scala/com/spark3d/spatialOperator/KNN.scala | 100% <100%> (ø) |
|
src/main/scala/com/spark3d/utils/Utils.scala | 94.11% <88.88%> (+38.85%) |
:arrow_up: |
src/main/scala/com/spark3d/Queries.scala | 93.33% <93.33%> (ø) |
|
checkers.py | 100% <0%> (ø) |
:arrow_up: |
queries.py | 41.66% <0%> (ø) |
|
... and 4 more |
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 19873df...155392a. Read the comment docs.
This PR re-introduces the KNN method, but with the DataFrame API.
The core of the routine remains untouched (developed during GSoC 2018 by Mayur, and using RDD), but the I/O are different. It takes a DataFrame as input, and return a DataFrame containing the coordinates of the K nearest neighbours.
Note: the option
unique
seems bugged (i.e. using KNN withunique=true
produces wrong results). I know thatordering.leastOf
recently changes its interface (from Scala iterable to Java Iterable) - so I won't be surprised if the meaning changed as well.