Merge branch 'spertus-abstract' into 'master'

Refactoring to enable easy creation of groupby jobs.

See merge request webgroup/sandcrawler!6

7 files changed, 60 insertions, 70 deletions

diff --git a/scalding/src/main/scala/sandcrawler/HBaseBuilder.scala b/scalding/src/main/scala/sandcrawler/HBaseBuilder.scala
index 87cc0cb..85766d6 100644
--- a/scalding/src/main/scala/sandcrawler/HBaseBuilder.scala
+++ b/scalding/src/main/scala/sandcrawler/HBaseBuilder.scala
@@ -13,20 +13,25 @@ object HBaseBuilder {
     "match0" -> List("status", "doi", "info"))
   val inverseSchema = {for ((k,vs) <- schema; v <-vs) yield (k + ":" + v)}.toList
 
+  // The argument should be of the form family:column, such as "file:size".
+  @throws(classOf[IllegalArgumentException])
+  def parseColSpec(colSpec: String) {
+    if (!(inverseSchema contains colSpec)) {
+      throw new IllegalArgumentException("No such column: " + colSpec)
+    }
+    val pair = colSpec split(":")
+    if (pair.length != 2) {
+      throw new IllegalArgumentException("Bad column specifier " + colSpec +
+        " (specifiers should be family:name)")
+    }
+    (pair(0), pair(1))
+  }
+
   // The argument should be a comma-separated list of family:column, such as "f:c, file:size".
   @throws(classOf[IllegalArgumentException])
-  def parseColSpec(colSpecs: List[String]) : (List[String], List[Fields]) = {
+  def parseColSpecs(colSpecs: List[String]) : (List[String], List[Fields]) = {
     // Verify that all column specifiers are legal.
-    for (colSpec <- colSpecs) {
-      if (!(inverseSchema contains colSpec)) {
-        throw new IllegalArgumentException("No such column: " + colSpec)
-      }
-      val pair = colSpec split(":")
-      if (colSpec.split(":").length != 2) {
-        throw new IllegalArgumentException("Bad column specifier " + colSpec + 
-          " (specifiers should be family:name)")
-      }
-    }
+    for (colSpec <- colSpecs) parseColSpec(colSpec)
 
     // Produce and return a tuple containing:
     // 1. A list of column families.
@@ -39,8 +44,8 @@ object HBaseBuilder {
     (families, groupedColNames.map({fields => new Fields(fields : _*)}))
   }
 
-  def build(table: String, server: String, colSpec: List[String], sourceMode: SourceMode, keyList: List[String] = List("key")) = {
-    val (families, fields) = parseColSpec(colSpec)
+  def build(table: String, server: String, colSpecs: List[String], sourceMode: SourceMode, keyList: List[String] = List("key")) = {
+    val (families, fields) = parseColSpecs(colSpecs)
     new HBaseSource(table, server, new Fields("key"), families, fields, sourceMode = sourceMode, keyList = keyList)
   }
 }
diff --git a/scalding/src/main/scala/sandcrawler/HBaseCountJob.scala b/scalding/src/main/scala/sandcrawler/HBaseCountJob.scala
new file mode 100644
index 0000000..815478b
--- /dev/null
+++ b/scalding/src/main/scala/sandcrawler/HBaseCountJob.scala
@@ -0,0 +1,30 @@
+package sandcrawler
+
+import cascading.property.AppProps
+import cascading.tuple.Fields
+import com.twitter.scalding._
+import java.util.Properties
+import parallelai.spyglass.base.JobBase
+import parallelai.spyglass.hbase.{HBaseSource, HBasePipeConversions}
+import parallelai.spyglass.hbase.HBaseConstants.SourceMode
+
+class HBaseCountJob(args: Args, colSpec: String) extends JobBase(args) with HBasePipeConversions {
+  val output = args("output")
+  HBaseBuilder.parseColSpec(colSpec)
+  val Col: String = colSpec.split(":")(1)
+
+  HBaseCountJob.getHBaseSource(colSpec)
+    .read
+    .fromBytesWritable(Symbol(Col))
+    .debug
+    .groupBy(Col){group => group.size('count)}
+    .write(Tsv(output))
+}
+
+object HBaseCountJob {
+  def getHBaseSource(colSpec: String) = HBaseBuilder.build(
+    "wbgrp-journal-extract-0-qa",     // HBase Table Name
+    "mtrcs-zk1.us.archive.org:2181",  // HBase Zookeeper server (to get runtime config info; can be array?)
+    List(colSpec),
+    SourceMode.SCAN_ALL)
+}
diff --git a/scalding/src/main/scala/sandcrawler/HBaseMimeCountJob.scala b/scalding/src/main/scala/sandcrawler/HBaseMimeCountJob.scala
index 819a652..250a12d 100644
--- a/scalding/src/main/scala/sandcrawler/HBaseMimeCountJob.scala
+++ b/scalding/src/main/scala/sandcrawler/HBaseMimeCountJob.scala
@@ -1,28 +1,6 @@
 package sandcrawler
 
-import cascading.property.AppProps
-import cascading.tuple.Fields
-import com.twitter.scalding._
-import java.util.Properties
-import parallelai.spyglass.base.JobBase
-import parallelai.spyglass.hbase.{HBaseSource, HBasePipeConversions}
-import parallelai.spyglass.hbase.HBaseConstants.SourceMode
+import com.twitter.scalding.Args
 
-class HBaseMimeCountJob(args: Args) extends JobBase(args) with HBasePipeConversions {
-  val output = args("output")
+class HBaseMimeCountJob(args: Args) extends HBaseCountJob(args, "file:mime") {}
 
-  HBaseMimeCountJob.getHBaseSource
-    .read
-    .fromBytesWritable(List('mime))
-    .debug
-    .groupBy('mime){group => group.size('count)}
-    .write(Tsv(output))
-}
-
-object HBaseMimeCountJob {
-  def getHBaseSource = HBaseBuilder.build(
-    "wbgrp-journal-extract-0-qa",     // HBase Table Name
-    "mtrcs-zk1.us.archive.org:2181",  // HBase Zookeeper server (to get runtime config info; can be array?)
-    List("file:mime"),
-    SourceMode.SCAN_ALL)
-}
diff --git a/scalding/src/main/scala/sandcrawler/HBaseStatusCountJob.scala b/scalding/src/main/scala/sandcrawler/HBaseStatusCountJob.scala
index 0675efc..27b3177 100644
--- a/scalding/src/main/scala/sandcrawler/HBaseStatusCountJob.scala
+++ b/scalding/src/main/scala/sandcrawler/HBaseStatusCountJob.scala
@@ -1,28 +1,5 @@
 package sandcrawler
 
-import cascading.property.AppProps
-import cascading.tuple.Fields
-import com.twitter.scalding._
-import java.util.Properties
-import parallelai.spyglass.base.JobBase
-import parallelai.spyglass.hbase.{HBaseSource, HBasePipeConversions}
-import parallelai.spyglass.hbase.HBaseConstants.SourceMode
+import com.twitter.scalding.Args
 
-class HBaseStatusCountJob(args: Args) extends JobBase(args) with HBasePipeConversions {
-  val output = args("output")
-
-  HBaseStatusCountJob.getHBaseSource
-    .read
-    .fromBytesWritable(List('status))
-//    .debug
-    .groupBy('status){group => group.size('count)}
-    .write(Tsv(output))
-}
-
-object HBaseStatusCountJob {
-  def getHBaseSource = HBaseBuilder.build(
-    "wbgrp-journal-extract-0-qa",     // HBase Table Name
-    "mtrcs-zk1.us.archive.org:2181",  // HBase Zookeeper server (to get runtime config info; can be array?)
-    List("grobid0:status"),
-    SourceMode.SCAN_ALL)
-}
+class HBaseStatusCountJob(args: Args) extends HBaseCountJob(args, "grobid0:status")
diff --git a/scalding/src/test/scala/sandcrawler/HBaseBuilderTest.scala b/scalding/src/test/scala/sandcrawler/HBaseBuilderTest.scala
index 4697f56..603a4c7 100644
--- a/scalding/src/test/scala/sandcrawler/HBaseBuilderTest.scala
+++ b/scalding/src/test/scala/sandcrawler/HBaseBuilderTest.scala
@@ -4,8 +4,8 @@ import cascading.tuple.Fields
 import org.scalatest._
 
 class HBaseBuilderTest extends FlatSpec with Matchers {
-  "parseColSpec()" should "work on legal nontrivial input" in {
-    val (fams, fields) = HBaseBuilder.parseColSpec(List("file:size", "file:cdx", "match0:status"))
+  "parseColSpecs()" should "work on legal nontrivial input" in {
+    val (fams, fields) = HBaseBuilder.parseColSpecs(List("file:size", "file:cdx", "match0:status"))
     fams should have length 2
     fields should have length 2
     val fileIndex = fams.indexOf("file")
@@ -17,26 +17,26 @@ class HBaseBuilderTest extends FlatSpec with Matchers {
   }
 
   it should "work on empty input" in {
-    val (fams, fields) = HBaseBuilder.parseColSpec(List())
+    val (fams, fields) = HBaseBuilder.parseColSpecs(List())
     fams should have length 0
     fields should have length 0
   }
 
   it should "throw IllegalArgumentException on malformed input" in {
     a [IllegalArgumentException] should be thrownBy {
-      HBaseBuilder.parseColSpec(List("file_size"))
+      HBaseBuilder.parseColSpecs(List("file_size"))
     }
   }
 
   it should "throw IllegalArgumentException on nonexistent family" in {
     a [IllegalArgumentException] should be thrownBy {
-      HBaseBuilder.parseColSpec(List("foo:bar"))
+      HBaseBuilder.parseColSpecs(List("foo:bar"))
     }
   }
 
   it should "throw IllegalArgumentException on nonexistent column" in {
     a [IllegalArgumentException] should be thrownBy {
-      HBaseBuilder.parseColSpec(List("file:bar"))
+      HBaseBuilder.parseColSpecs(List("file:bar"))
     }
   }
 }
diff --git a/scalding/src/test/scala/sandcrawler/HBaseMimeCountTest.scala b/scalding/src/test/scala/sandcrawler/HBaseMimeCountTest.scala
index eb6f4ff..cc400c5 100644
--- a/scalding/src/test/scala/sandcrawler/HBaseMimeCountTest.scala
+++ b/scalding/src/test/scala/sandcrawler/HBaseMimeCountTest.scala
@@ -41,7 +41,7 @@ class HBaseMimeCountTest extends FunSpec with TupleConversions {
     .arg("app.conf.path", "app.conf")
     .arg("output", output)
     .arg("debug", "true")
-    .source[Tuple](HBaseMimeCountJob.getHBaseSource,
+    .source[Tuple](HBaseCountJob.getHBaseSource("file:mime"),
       sampleData.map(l => new Tuple(l.map(s => {new ImmutableBytesWritable(Bytes.toBytes(s))}):_*)))
       .sink[Tuple](Tsv(output)) {
         outputBuffer =>
diff --git a/scalding/src/test/scala/sandcrawler/HBaseStatusCountTest.scala b/scalding/src/test/scala/sandcrawler/HBaseStatusCountTest.scala
index 8b5c3d6..a0b313e 100644
--- a/scalding/src/test/scala/sandcrawler/HBaseStatusCountTest.scala
+++ b/scalding/src/test/scala/sandcrawler/HBaseStatusCountTest.scala
@@ -41,7 +41,7 @@ class HBaseStatusCountTest extends FunSpec with TupleConversions {
     .arg("app.conf.path", "app.conf")
     .arg("output", output)
     .arg("debug", "true")
-    .source[Tuple](HBaseStatusCountJob.getHBaseSource,
+    .source[Tuple](HBaseCountJob.getHBaseSource("grobid0:status"),
       sampleData.map(l => new Tuple(l.map(s => {new ImmutableBytesWritable(Bytes.toBytes(s))}):_*)))
       .sink[Tuple](Tsv(output)) {
         outputBuffer =>