aboutsummaryrefslogtreecommitdiffstats
path: root/scalding/src
diff options
context:
space:
mode:
authorBryan Newbold <bnewbold@archive.org>2018-08-21 21:28:18 -0700
committerBryan Newbold <bnewbold@archive.org>2018-08-21 21:28:18 -0700
commit9c2324d1b32116dc3a8a0f43c08ee7dac1bdf5f4 (patch)
tree1f3b264118c437130c8a676d560152b0ff9abf07 /scalding/src
parente99c1d8485ee232d782d2b82c44dc9e1d0e01044 (diff)
parent2f8a538fb658b3de6bf7c9361bd13f2eb7968abb (diff)
downloadsandcrawler-9c2324d1b32116dc3a8a0f43c08ee7dac1bdf5f4.tar.gz
sandcrawler-9c2324d1b32116dc3a8a0f43c08ee7dac1bdf5f4.zip
Merge branch 'bnewbold-match-scale'
Diffstat (limited to 'scalding/src')
-rw-r--r--scalding/src/main/scala/sandcrawler/ScoreJob.scala13
-rw-r--r--scalding/src/test/scala/sandcrawler/ScoreJobTest.scala2
2 files changed, 10 insertions, 5 deletions
diff --git a/scalding/src/main/scala/sandcrawler/ScoreJob.scala b/scalding/src/main/scala/sandcrawler/ScoreJob.scala
index 75d45e9..28e9132 100644
--- a/scalding/src/main/scala/sandcrawler/ScoreJob.scala
+++ b/scalding/src/main/scala/sandcrawler/ScoreJob.scala
@@ -13,15 +13,18 @@ class ScoreJob(args: Args) extends JobBase(args) {
val pipe1 : TypedPipe[(String, ReduceFeatures)] = sc1.getInputPipe(args)
val pipe2 : TypedPipe[(String, ReduceFeatures)] = sc2.getInputPipe(args)
- pipe1.join(pipe2).map { entry =>
- val (slug : String, (features1 : ReduceFeatures, features2 : ReduceFeatures)) = entry
- new ReduceOutput(
+ pipe1
+ .addTrap(TypedTsv(args("output") + ".trapped"))
+ .join(pipe2)
+ .map { entry =>
+ val (slug : String, (features1 : ReduceFeatures, features2 : ReduceFeatures)) = entry
+ new ReduceOutput(
slug,
Scorable.computeSimilarity(features1, features2),
features1.json,
features2.json)
- }
- //TypedTsv doesn't work over case classes.
+ }
+ //TypedTsv doesn't work over case classes.
.map { entry => (entry.slug, entry.score, entry.json1, entry.json2) }
.write(TypedTsv[(String, Int, String, String)](args("output")))
}
diff --git a/scalding/src/test/scala/sandcrawler/ScoreJobTest.scala b/scalding/src/test/scala/sandcrawler/ScoreJobTest.scala
index 00e4659..35c31e5 100644
--- a/scalding/src/test/scala/sandcrawler/ScoreJobTest.scala
+++ b/scalding/src/test/scala/sandcrawler/ScoreJobTest.scala
@@ -182,6 +182,8 @@ class ScoreJobTest extends FlatSpec with Matchers {
1 -> CrossrefStrings(1),
2 -> CrossrefStrings(2),
3 -> CrossrefStrings(3)))
+ .sink[(String, ReduceFeatures)](TypedTsv[(String, ReduceFeatures)](output + ".trapped")) {
+ _ => () }
.sink[(String, Int, String, String)](TypedTsv[(String, Int, String, String)](output)) {
// Grobid titles and slugs (in parentheses):
// Title 1 (title1)