forked from D-Net/dnet-hadoop
improved workflow importing datacite
This commit is contained in:
parent
b5b7dc2104
commit
1dfda3624e
|
|
@ -57,10 +57,13 @@ abstract class AbstractRestClient extends Iterator[String]{
|
|||
|
||||
private def doHTTPRequest[A <: HttpUriRequest](r: A) :String ={
|
||||
val client = HttpClients.createDefault
|
||||
try {
|
||||
var tries = 4
|
||||
try {
|
||||
while (tries > 0) {
|
||||
|
||||
println(s"requesting ${r.getURI}")
|
||||
val response = client.execute(r)
|
||||
println(s"get response with status${response.getStatusLine.getStatusCode}")
|
||||
if (response.getStatusLine.getStatusCode > 400) {
|
||||
tries -= 1
|
||||
}
|
||||
|
|
|
|||
|
|
@ -3,7 +3,7 @@ package eu.dnetlib.dhp.actionmanager.datacite
|
|||
import org.json4s.{DefaultFormats, JValue}
|
||||
import org.json4s.jackson.JsonMethods.{compact, parse, render}
|
||||
|
||||
class DataciteAPIImporter(timestamp: Long = 0, blocks: Long = 10) extends AbstractRestClient {
|
||||
class DataciteAPIImporter(timestamp: Long = 0, blocks: Long = 10, until:Long = -1) extends AbstractRestClient {
|
||||
|
||||
override def extractInfo(input: String): Unit = {
|
||||
implicit lazy val formats: DefaultFormats.type = org.json4s.DefaultFormats
|
||||
|
|
@ -16,9 +16,15 @@ class DataciteAPIImporter(timestamp: Long = 0, blocks: Long = 10) extends Abstra
|
|||
current_index = 0
|
||||
}
|
||||
|
||||
def get_url():String ={
|
||||
val to = if (until> 0) s"$until" else "*"
|
||||
s"https://api.datacite.org/dois?page[cursor]=1&page[size]=$blocks&query=updated:[$timestamp%20TO%20$to]"
|
||||
|
||||
}
|
||||
|
||||
override def getBufferData(): Unit = {
|
||||
if (!complete) {
|
||||
val response = if (scroll_value.isDefined) doHTTPGETRequest(scroll_value.get) else doHTTPGETRequest(s"https://api.datacite.org/dois?page[cursor]=1&page[size]=$blocks&query=updated:[$timestamp%20TO%20*]")
|
||||
val response = if (scroll_value.isDefined) doHTTPGETRequest(scroll_value.get) else doHTTPGETRequest(get_url())
|
||||
extractInfo(response)
|
||||
}
|
||||
}
|
||||
|
|
|
|||
|
|
@ -164,12 +164,11 @@ object DataciteToOAFTransformation {
|
|||
case _: Throwable => try {
|
||||
return Some(LocalDate.parse(a_date, df_it).toString)
|
||||
} catch {
|
||||
case _: Throwable => try {
|
||||
case _: Throwable =>
|
||||
return None
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
d
|
||||
}
|
||||
|
||||
|
|
|
|||
|
|
@ -1,5 +1,6 @@
|
|||
package eu.dnetlib.dhp.actionmanager.datacite
|
||||
|
||||
import eu.dnetlib.dhp.actionmanager.datacite.DataciteToOAFTransformation.df_it
|
||||
import eu.dnetlib.dhp.application.ArgumentApplicationParser
|
||||
import org.apache.hadoop.conf.Configuration
|
||||
import org.apache.hadoop.fs.{FileSystem, LocalFileSystem, Path}
|
||||
|
|
@ -15,7 +16,7 @@ import org.apache.spark.sql.functions.max
|
|||
import org.slf4j.{Logger, LoggerFactory}
|
||||
|
||||
import java.time.format.DateTimeFormatter._
|
||||
import java.time.{LocalDateTime, ZoneOffset}
|
||||
import java.time.{LocalDate, LocalDateTime, ZoneOffset}
|
||||
import scala.io.Source
|
||||
|
||||
object ImportDatacite {
|
||||
|
|
@ -37,7 +38,6 @@ object ImportDatacite {
|
|||
}
|
||||
|
||||
|
||||
|
||||
def main(args: Array[String]): Unit = {
|
||||
|
||||
val parser = new ArgumentApplicationParser(Source.fromInputStream(getClass.getResourceAsStream("/eu/dnetlib/dhp/actionmanager/datacite/import_from_api.json")).mkString)
|
||||
|
|
@ -56,6 +56,10 @@ object ImportDatacite {
|
|||
val hdfsTargetPath = new Path(targetPath)
|
||||
log.info(s"hdfsTargetPath is $hdfsTargetPath")
|
||||
|
||||
|
||||
val spkipImport = parser.get("skipImport")
|
||||
log.info(s"skipImport is $spkipImport")
|
||||
|
||||
val spark: SparkSession = SparkSession.builder()
|
||||
.appName(ImportDatacite.getClass.getSimpleName)
|
||||
.master(master)
|
||||
|
|
@ -104,12 +108,11 @@ object ImportDatacite {
|
|||
val dump: Dataset[DataciteType] = spark.read.load(dataciteDump).as[DataciteType]
|
||||
val ts = dump.select(max("timestamp")).first().getLong(0)
|
||||
|
||||
log.info(s"last Timestamp is $ts")
|
||||
println(s"last Timestamp is $ts")
|
||||
|
||||
val cnt = writeSequenceFile(hdfsTargetPath, ts, conf)
|
||||
val cnt = if ("true".equalsIgnoreCase(spkipImport)) 1 else writeSequenceFile(hdfsTargetPath, ts, conf)
|
||||
|
||||
|
||||
log.info(s"Imported from Datacite API $cnt documents")
|
||||
println(s"Imported from Datacite API $cnt documents")
|
||||
|
||||
if (cnt > 0) {
|
||||
|
||||
|
|
@ -135,19 +138,21 @@ object ImportDatacite {
|
|||
}
|
||||
|
||||
private def writeSequenceFile(hdfsTargetPath: Path, timestamp: Long, conf: Configuration): Long = {
|
||||
val client = new DataciteAPIImporter(timestamp*1000, 1000)
|
||||
var from:Long = timestamp * 1000
|
||||
val delta:Long = 50000000L
|
||||
var client: DataciteAPIImporter = null
|
||||
val now :Long =System.currentTimeMillis()
|
||||
var i = 0
|
||||
try {
|
||||
val writer = SequenceFile.createWriter(conf, SequenceFile.Writer.file(hdfsTargetPath), SequenceFile.Writer.keyClass(classOf[IntWritable]), SequenceFile.Writer.valueClass(classOf[Text]))
|
||||
try {
|
||||
|
||||
var start: Long = System.currentTimeMillis
|
||||
while (from < now) {
|
||||
client = new DataciteAPIImporter(from, 1000, from + delta)
|
||||
var end: Long = 0
|
||||
val key: IntWritable = new IntWritable(i)
|
||||
val value: Text = new Text
|
||||
while ( {
|
||||
client.hasNext
|
||||
}) {
|
||||
while (client.hasNext) {
|
||||
key.set({
|
||||
i += 1;
|
||||
i - 1
|
||||
|
|
@ -162,8 +167,15 @@ object ImportDatacite {
|
|||
start = System.currentTimeMillis
|
||||
}
|
||||
}
|
||||
println(s"updating from value: $from -> ${from+delta}")
|
||||
from = from + delta
|
||||
}
|
||||
} catch {
|
||||
case e: Throwable =>
|
||||
println("Error", e)
|
||||
} finally if (writer != null) writer.close()
|
||||
}
|
||||
i
|
||||
}
|
||||
|
||||
}
|
||||
|
|
@ -12,6 +12,12 @@
|
|||
"paramDescription": "the path of the Datacite dump",
|
||||
"paramRequired": true
|
||||
},
|
||||
{
|
||||
"paramName": "s",
|
||||
"paramLongName": "skipImport",
|
||||
"paramDescription": "avoid to downlaod new items but apply the previous update",
|
||||
"paramRequired": false
|
||||
},
|
||||
{
|
||||
"paramName": "n",
|
||||
"paramLongName": "namenode",
|
||||
|
|
|
|||
|
|
@ -13,6 +13,11 @@
|
|||
<name>nativeInputPath</name>
|
||||
<description>the path of the input MDStore</description>
|
||||
</property>
|
||||
<property>
|
||||
<name>skipimport</name>
|
||||
<value>false</value>
|
||||
<description>the path of the input MDStore</description>
|
||||
</property>
|
||||
|
||||
|
||||
</parameters>
|
||||
|
|
@ -51,6 +56,7 @@
|
|||
<arg>-t</arg><arg>${nativeInputPath}</arg>
|
||||
<arg>-d</arg><arg>${mdstoreInputPath}</arg>
|
||||
<arg>-n</arg><arg>${nameNode}</arg>
|
||||
<arg>-s</arg><arg>${skipimport}</arg>
|
||||
<arg>--master</arg><arg>yarn-cluster</arg>
|
||||
</spark>
|
||||
<ok to="TransformJob"/>
|
||||
|
|
@ -81,7 +87,7 @@
|
|||
<arg>-tr</arg><arg>${isLookupUrl}</arg>
|
||||
<arg>--master</arg><arg>yarn-cluster</arg>
|
||||
</spark>
|
||||
<ok to="DeletePathIfExists"/>
|
||||
<ok to="End"/>
|
||||
<error to="Kill"/>
|
||||
</action>
|
||||
|
||||
|
|
|
|||
|
|
@ -57,7 +57,7 @@ object SparkGenerateDOIBoostActionSet {
|
|||
|
||||
|
||||
val asCRelation = spark.read.load(crossRefRelation).as[Relation]
|
||||
.filter(r => r!= null || (r.getSource != null && r.getTarget != null))
|
||||
.filter(r => r!= null && r.getSource != null && r.getTarget != null)
|
||||
.map(d=>DoiBoostMappingUtil.toActionSet(d))(Encoders.tuple(Encoders.STRING, Encoders.STRING))
|
||||
|
||||
|
||||
|
|
|
|||
|
|
@ -59,6 +59,17 @@ class CrossrefMappingTest {
|
|||
}
|
||||
|
||||
|
||||
@Test
|
||||
def testSum() :Unit = {
|
||||
val from:Long = 1613135645000L
|
||||
val delta:Long = 1000000L
|
||||
|
||||
|
||||
println(s"updating from value: $from -> ${from+delta}")
|
||||
|
||||
|
||||
}
|
||||
|
||||
@Test
|
||||
def testOrcidID() :Unit = {
|
||||
val json = Source.fromInputStream(getClass.getResourceAsStream("orcid_data.json")).mkString
|
||||
|
|
|
|||
Loading…
Reference in New Issue