Merge pull request 'HostedByMap' (#201) from hostedByMap_update into beta

Reviewed-on: #201
This commit is contained in:
Claudio Atzori 2022-03-04 16:26:27 +01:00
commit 9f3036c847
29 changed files with 1182 additions and 234 deletions

View File

@ -0,0 +1,40 @@
package eu.dnetlib.dhp.common.collection;
import java.io.BufferedOutputStream;
import java.io.IOException;
import java.util.zip.GZIPOutputStream;
import org.apache.commons.compress.archivers.tar.TarArchiveEntry;
import org.apache.commons.compress.archivers.tar.TarArchiveInputStream;
import org.apache.commons.compress.compressors.gzip.GzipCompressorInputStream;
import org.apache.commons.io.IOUtils;
import org.apache.hadoop.fs.FSDataInputStream;
import org.apache.hadoop.fs.FSDataOutputStream;
import org.apache.hadoop.fs.FileSystem;
import org.apache.hadoop.fs.Path;
public class DecompressTarGz {
public static void doExtract(FileSystem fs, String outputPath, String tarGzPath) throws IOException {
FSDataInputStream inputFileStream = fs.open(new Path(tarGzPath));
try (TarArchiveInputStream tais = new TarArchiveInputStream(
new GzipCompressorInputStream(inputFileStream))) {
TarArchiveEntry entry = null;
while ((entry = tais.getNextTarEntry()) != null) {
if (!entry.isDirectory()) {
try (
FSDataOutputStream out = fs
.create(new Path(outputPath.concat(entry.getName()).concat(".gz")));
GZIPOutputStream gzipOs = new GZIPOutputStream(new BufferedOutputStream(out))) {
IOUtils.copy(tais, gzipOs);
}
}
}
}
}
}

View File

@ -1,19 +1,13 @@
package eu.dnetlib.doiboost.crossref;
import java.io.BufferedOutputStream;
import java.net.URI;
import java.util.zip.GZIPOutputStream;
import static eu.dnetlib.dhp.common.collection.DecompressTarGz.doExtract;
import java.net.URI;
import org.apache.commons.compress.archivers.tar.TarArchiveEntry;
import org.apache.commons.compress.archivers.tar.TarArchiveInputStream;
import org.apache.commons.compress.compressors.gzip.GzipCompressorInputStream;
import org.apache.commons.io.IOUtils;
import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.FSDataInputStream;
import org.apache.hadoop.fs.FSDataOutputStream;
import org.apache.hadoop.fs.FileSystem;
import org.apache.hadoop.fs.Path;
import org.mortbay.log.Log;
import eu.dnetlib.dhp.application.ArgumentApplicationParser;
@ -33,31 +27,16 @@ public class ExtractCrossrefRecords {
final String outputPath = parser.get("outputPath");
final String crossrefFileNameTarGz = parser.get("crossrefFileNameTarGz");
Path hdfsreadpath = new Path(workingPath.concat("/").concat(crossrefFileNameTarGz));
Configuration conf = new Configuration();
conf.set("fs.defaultFS", workingPath);
conf.set("fs.hdfs.impl", org.apache.hadoop.hdfs.DistributedFileSystem.class.getName());
conf.set("fs.file.impl", org.apache.hadoop.fs.LocalFileSystem.class.getName());
FileSystem fs = FileSystem.get(URI.create(workingPath), conf);
FSDataInputStream crossrefFileStream = fs.open(hdfsreadpath);
try (TarArchiveInputStream tais = new TarArchiveInputStream(
new GzipCompressorInputStream(crossrefFileStream))) {
TarArchiveEntry entry = null;
while ((entry = tais.getNextTarEntry()) != null) {
if (!entry.isDirectory()) {
try (
FSDataOutputStream out = fs
.create(new Path(outputPath.concat(entry.getName()).concat(".gz")));
GZIPOutputStream gzipOs = new GZIPOutputStream(new BufferedOutputStream(out))) {
IOUtils.copy(tais, gzipOs);
doExtract(fs, outputPath, workingPath.concat("/").concat(crossrefFileNameTarGz));
}
}
}
}
Log.info("Crossref dump reading completed");
}
}

View File

@ -23,7 +23,7 @@ public class DownloadCSV {
private static final Logger log = LoggerFactory.getLogger(DownloadCSV.class);
public static final char DEFAULT_DELIMITER = ';';
public static final char DEFAULT_DELIMITER = ',';
public static void main(final String[] args) throws Exception {
final ArgumentApplicationParser parser = new ArgumentApplicationParser(
@ -40,9 +40,6 @@ public class DownloadCSV {
final String fileURL = parser.get("fileURL");
log.info("fileURL {}", fileURL);
final String workingPath = parser.get("workingPath");
log.info("workingPath {}", workingPath);
final String outputFile = parser.get("outputFile");
log.info("outputFile {}", outputFile);
@ -63,31 +60,15 @@ public class DownloadCSV {
FileSystem fileSystem = FileSystem.get(conf);
new DownloadCSV().doDownload(fileURL, workingPath, outputFile, classForName, delimiter, fileSystem);
new DownloadCSV().doDownload(fileURL, outputFile, classForName, delimiter, fileSystem);
}
protected void doDownload(String fileURL, String workingPath, String outputFile, String classForName,
protected void doDownload(String fileURL, String outputFile, String classForName,
char delimiter, FileSystem fs)
throws IOException, ClassNotFoundException, CollectorException {
final HttpConnector2 connector2 = new HttpConnector2();
final Path path = new Path(workingPath + "/replaced.csv");
try (BufferedReader in = new BufferedReader(
new InputStreamReader(connector2.getInputSourceAsStream(fileURL)))) {
try (PrintWriter writer = new PrintWriter(
new OutputStreamWriter(fs.create(path, true), StandardCharsets.UTF_8))) {
String line;
while ((line = in.readLine()) != null) {
writer.println(line.replace("\\\"", "\""));
}
}
}
try (InputStreamReader reader = new InputStreamReader(fs.open(path))) {
try (InputStreamReader reader = new InputStreamReader(new HttpConnector2().getInputSourceAsStream(fileURL))) {
GetCSV.getCsv(fs, reader, outputFile, classForName, delimiter);
}
}

View File

@ -1,84 +0,0 @@
package eu.dnetlib.dhp.oa.graph.hostedbymap;
import java.io.*;
import java.util.Objects;
import java.util.Optional;
import org.apache.commons.io.FileUtils;
import org.apache.commons.io.IOUtils;
import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.FileSystem;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import eu.dnetlib.dhp.application.ArgumentApplicationParser;
import eu.dnetlib.dhp.common.collection.GetCSV;
import eu.dnetlib.dhp.common.collection.HttpConnector2;
public class DownloadCSV2 {
private static final Logger log = LoggerFactory.getLogger(DownloadCSV2.class);
public static final char DEFAULT_DELIMITER = ';';
public static void main(final String[] args) throws Exception {
final ArgumentApplicationParser parser = new ArgumentApplicationParser(
IOUtils
.toString(
Objects
.requireNonNull(
DownloadCSV2.class
.getResourceAsStream(
"/eu/dnetlib/dhp/oa/graph/hostedbymap/download_csv_parameters.json"))));
parser.parseArgument(args);
final String fileURL = parser.get("fileURL");
log.info("fileURL {}", fileURL);
final String tmpFile = parser.get("tmpFile");
log.info("tmpFile {}", tmpFile);
final String outputFile = parser.get("outputFile");
log.info("outputFile {}", outputFile);
final String hdfsNameNode = parser.get("hdfsNameNode");
log.info("hdfsNameNode {}", hdfsNameNode);
final String classForName = parser.get("classForName");
log.info("classForName {}", classForName);
final char delimiter = Optional
.ofNullable(parser.get("delimiter"))
.map(s -> s.charAt(0))
.orElse(DEFAULT_DELIMITER);
log.info("delimiter {}", delimiter);
HttpConnector2 connector2 = new HttpConnector2();
try (BufferedReader in = new BufferedReader(
new InputStreamReader(connector2.getInputSourceAsStream(fileURL)))) {
try (PrintWriter writer = new PrintWriter(new BufferedWriter(new FileWriter(tmpFile)))) {
String line;
while ((line = in.readLine()) != null) {
writer.println(line.replace("\\\"", "\""));
}
}
}
try (BufferedReader in = new BufferedReader(new FileReader(tmpFile))) {
Configuration conf = new Configuration();
conf.set("fs.defaultFS", hdfsNameNode);
FileSystem fileSystem = FileSystem.get(conf);
GetCSV.getCsv(fileSystem, in, outputFile, classForName, delimiter);
} finally {
FileUtils.deleteQuietly(new File(tmpFile));
}
}
}

View File

@ -0,0 +1,117 @@
package eu.dnetlib.dhp.oa.graph.hostedbymap;
import static eu.dnetlib.dhp.common.collection.DecompressTarGz.doExtract;
import java.io.BufferedOutputStream;
import java.io.IOException;
import java.io.PrintWriter;
import java.util.Arrays;
import java.util.Objects;
import org.apache.commons.io.IOUtils;
import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.*;
import org.apache.hadoop.io.compress.CompressionCodec;
import org.apache.hadoop.io.compress.CompressionCodecFactory;
import org.apache.hadoop.io.compress.CompressionInputStream;
import org.jetbrains.annotations.NotNull;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import com.fasterxml.jackson.core.JsonProcessingException;
import com.fasterxml.jackson.databind.ObjectMapper;
import eu.dnetlib.dhp.application.ArgumentApplicationParser;
import eu.dnetlib.dhp.oa.graph.hostedbymap.model.DOAJModel;
import eu.dnetlib.dhp.oa.graph.hostedbymap.model.doaj.DOAJEntry;
public class ExtractAndMapDoajJson {
private static final Logger log = LoggerFactory.getLogger(ExtractAndMapDoajJson.class);
public static void main(final String[] args) throws Exception {
final ArgumentApplicationParser parser = new ArgumentApplicationParser(
IOUtils
.toString(
Objects
.requireNonNull(
ExtractAndMapDoajJson.class
.getResourceAsStream(
"/eu/dnetlib/dhp/oa/graph/hostedbymap/download_json_parameters.json"))));
parser.parseArgument(args);
final String compressedInput = parser.get("compressedFile");
log.info("compressedInput {}", compressedInput);
final String hdfsNameNode = parser.get("hdfsNameNode");
log.info("hdfsNameNode {}", hdfsNameNode);
final String outputPath = parser.get("outputPath");
log.info("outputPath {}", outputPath);
final String workingPath = parser.get("workingPath");
log.info("workingPath {}", workingPath);
Configuration conf = new Configuration();
conf.set("fs.defaultFS", hdfsNameNode);
FileSystem fs = FileSystem.get(conf);
CompressionCodecFactory factory = new CompressionCodecFactory(conf);
CompressionCodec codec = factory.getCodecByClassName("org.apache.hadoop.io.compress.GzipCodec");
doExtract(fs, workingPath, compressedInput);
doMap(fs, workingPath, outputPath, codec);
}
private static void doMap(FileSystem fs, String workingPath, String outputPath, CompressionCodec codec)
throws IOException {
RemoteIterator<LocatedFileStatus> fileStatusListIterator = fs
.listFiles(
new Path(workingPath), true);
Path hdfsWritePath = new Path(outputPath);
if (fs.exists(hdfsWritePath)) {
fs.delete(hdfsWritePath, true);
}
try (
FSDataOutputStream out = fs
.create(hdfsWritePath);
PrintWriter writer = new PrintWriter(new BufferedOutputStream(out))) {
while (fileStatusListIterator.hasNext()) {
Path path = fileStatusListIterator.next().getPath();
if (!fs.isDirectory(path)) {
FSDataInputStream is = fs.open(path);
CompressionInputStream compressionInputStream = codec.createInputStream(is);
DOAJEntry[] doajEntries = new ObjectMapper().readValue(compressionInputStream, DOAJEntry[].class);
Arrays.stream(doajEntries).forEach(doaj -> {
try {
writer.println(new ObjectMapper().writeValueAsString(getDoajModel(doaj)));
} catch (JsonProcessingException e) {
e.printStackTrace();
}
});
}
}
}
}
@NotNull
public static DOAJModel getDoajModel(DOAJEntry doaj) {
DOAJModel doajModel = new DOAJModel();
doajModel.setOaStart(doaj.getBibjson().getOa_start());
doajModel.setEissn(doaj.getBibjson().getEissn());
doajModel.setIssn(doaj.getBibjson().getPissn());
doajModel.setJournalTitle(doaj.getBibjson().getTitle());
doajModel.setReviewProcess(doaj.getBibjson().getEditorial().getReview_process());
return doajModel;
}
}

View File

@ -2,6 +2,7 @@
package eu.dnetlib.dhp.oa.graph.hostedbymap.model;
import java.io.Serializable;
import java.util.List;
import com.opencsv.bean.CsvBindByName;
@ -17,7 +18,17 @@ public class DOAJModel implements Serializable {
private String eissn;
@CsvBindByName(column = "Review process")
private String reviewProcess;
private List<String> reviewProcess;
private Integer oaStart;
public Integer getOaStart() {
return oaStart;
}
public void setOaStart(Integer oaStart) {
this.oaStart = oaStart;
}
public String getJournalTitle() {
return journalTitle;
@ -43,11 +54,11 @@ public class DOAJModel implements Serializable {
this.eissn = eissn;
}
public String getReviewProcess() {
public List<String> getReviewProcess() {
return reviewProcess;
}
public void setReviewProcess(String reviewProcess) {
public void setReviewProcess(List<String> reviewProcess) {
this.reviewProcess = reviewProcess;
}
}

View File

@ -0,0 +1,35 @@
package eu.dnetlib.dhp.oa.graph.hostedbymap.model.doaj;
import java.io.Serializable;
import java.util.List;
public class APC implements Serializable {
private Boolean has_apc;
private String url;
private List<Max> max;
public List<Max> getMax() {
return max;
}
public void setMax(List<Max> max) {
this.max = max;
}
public Boolean getHas_apc() {
return has_apc;
}
public void setHas_apc(Boolean has_apc) {
this.has_apc = has_apc;
}
public String getUrl() {
return url;
}
public void setUrl(String url) {
this.url = url;
}
}

View File

@ -0,0 +1,25 @@
package eu.dnetlib.dhp.oa.graph.hostedbymap.model.doaj;
import java.io.Serializable;
public class Admin implements Serializable {
private Boolean ticked;
private Boolean seal;
public Boolean getTicked() {
return ticked;
}
public void setTicked(Boolean ticked) {
this.ticked = ticked;
}
public Boolean getSeal() {
return seal;
}
public void setSeal(Boolean seal) {
this.seal = seal;
}
}

View File

@ -0,0 +1,44 @@
package eu.dnetlib.dhp.oa.graph.hostedbymap.model.doaj;
import java.io.Serializable;
import java.util.List;
public class Article implements Serializable {
private String license_display_example_url;
private List<String> license_display;
private Boolean orcid;
private Boolean i4oc_open_citations;
public String getLicense_display_example_url() {
return license_display_example_url;
}
public void setLicense_display_example_url(String license_display_example_url) {
this.license_display_example_url = license_display_example_url;
}
public List<String> getLicense_display() {
return license_display;
}
public void setLicense_display(List<String> license_display) {
this.license_display = license_display;
}
public Boolean getOrcid() {
return orcid;
}
public void setOrcid(Boolean orcid) {
this.orcid = orcid;
}
public Boolean getI4oc_open_citations() {
return i4oc_open_citations;
}
public void setI4oc_open_citations(Boolean i4oc_open_citations) {
this.i4oc_open_citations = i4oc_open_citations;
}
}

View File

@ -0,0 +1,253 @@
package eu.dnetlib.dhp.oa.graph.hostedbymap.model.doaj;
import java.io.Serializable;
import java.util.List;
import com.fasterxml.jackson.annotation.JsonIgnore;
public class BibJson implements Serializable {
private Editorial editorial;
private PidScheme pid_scheme;
private Copyright copyright;
private List<String> keywords;
private Plagiarism plagiarism;
private List<Subject> subject;
private String eissn;
private String pissn;
private List<String> language;
private String title;
private Article article;
private Institution institution;
private Preservation preservation;
private List<License> license;
private Ref ref;
private Integer oa_start;
private APC apc;
private OtherCharges other_charges;
private Integer publication_time_weeks;
private DepositPolicy deposit_policy;
private Publisher publisher;
private Boolean boai;
private Waiver waiver;
private String alternative_title;
private List<String> is_replaced_by;
private List<String> replaces;
private String discontinued_date;
public String getDiscontinued_date() {
return discontinued_date;
}
public void setDiscontinued_date(String discontinued_date) {
this.discontinued_date = discontinued_date;
}
public List<String> getReplaces() {
return replaces;
}
public void setReplaces(List<String> replaces) {
this.replaces = replaces;
}
public List<String> getIs_replaced_by() {
return is_replaced_by;
}
public void setIs_replaced_by(List<String> is_replaced_by) {
this.is_replaced_by = is_replaced_by;
}
public String getAlternative_title() {
return alternative_title;
}
public void setAlternative_title(String alternative_title) {
this.alternative_title = alternative_title;
}
public String getPissn() {
return pissn;
}
public void setPissn(String pissn) {
this.pissn = pissn;
}
public Editorial getEditorial() {
return editorial;
}
public void setEditorial(Editorial editorial) {
this.editorial = editorial;
}
public PidScheme getPid_scheme() {
return pid_scheme;
}
public void setPid_scheme(PidScheme pid_scheme) {
this.pid_scheme = pid_scheme;
}
public Copyright getCopyright() {
return copyright;
}
public void setCopyright(Copyright copyright) {
this.copyright = copyright;
}
public List<String> getKeywords() {
return keywords;
}
public void setKeywords(List<String> keywords) {
this.keywords = keywords;
}
public Plagiarism getPlagiarism() {
return plagiarism;
}
public void setPlagiarism(Plagiarism plagiarism) {
this.plagiarism = plagiarism;
}
public List<Subject> getSubject() {
return subject;
}
public void setSubject(List<Subject> subject) {
this.subject = subject;
}
public String getEissn() {
return eissn;
}
public void setEissn(String eissn) {
this.eissn = eissn;
}
public List<String> getLanguage() {
return language;
}
public void setLanguage(List<String> language) {
this.language = language;
}
public String getTitle() {
return title;
}
public void setTitle(String title) {
this.title = title;
}
public Article getArticle() {
return article;
}
public void setArticle(Article article) {
this.article = article;
}
public Institution getInstitution() {
return institution;
}
public void setInstitution(Institution institution) {
this.institution = institution;
}
public Preservation getPreservation() {
return preservation;
}
public void setPreservation(Preservation preservation) {
this.preservation = preservation;
}
public List<License> getLicense() {
return license;
}
public void setLicense(List<License> license) {
this.license = license;
}
public Ref getRef() {
return ref;
}
public void setRef(Ref ref) {
this.ref = ref;
}
public Integer getOa_start() {
return oa_start;
}
public void setOa_start(Integer oa_start) {
this.oa_start = oa_start;
}
public APC getApc() {
return apc;
}
public void setApc(APC apc) {
this.apc = apc;
}
public OtherCharges getOther_charges() {
return other_charges;
}
public void setOther_charges(OtherCharges other_charges) {
this.other_charges = other_charges;
}
public Integer getPublication_time_weeks() {
return publication_time_weeks;
}
public void setPublication_time_weeks(Integer publication_time_weeks) {
this.publication_time_weeks = publication_time_weeks;
}
public DepositPolicy getDeposit_policy() {
return deposit_policy;
}
public void setDeposit_policy(DepositPolicy deposit_policy) {
this.deposit_policy = deposit_policy;
}
public Publisher getPublisher() {
return publisher;
}
public void setPublisher(Publisher publisher) {
this.publisher = publisher;
}
public Boolean getBoai() {
return boai;
}
public void setBoai(Boolean boai) {
this.boai = boai;
}
public Waiver getWaiver() {
return waiver;
}
public void setWaiver(Waiver waiver) {
this.waiver = waiver;
}
}

View File

@ -0,0 +1,25 @@
package eu.dnetlib.dhp.oa.graph.hostedbymap.model.doaj;
import java.io.Serializable;
public class Copyright implements Serializable {
private Boolean author_retains;
private String url;
public Boolean getAuthor_retains() {
return author_retains;
}
public void setAuthor_retains(Boolean author_retains) {
this.author_retains = author_retains;
}
public String getUrl() {
return url;
}
public void setUrl(String url) {
this.url = url;
}
}

View File

@ -0,0 +1,52 @@
package eu.dnetlib.dhp.oa.graph.hostedbymap.model.doaj;
import java.io.Serializable;
public class DOAJEntry implements Serializable {
private String last_updated;
private BibJson bibjson;
private Admin admin;
private String created_date;
private String id;
public String getLast_updated() {
return last_updated;
}
public void setLast_updated(String last_updated) {
this.last_updated = last_updated;
}
public BibJson getBibjson() {
return bibjson;
}
public void setBibjson(BibJson bibjson) {
this.bibjson = bibjson;
}
public Admin getAdmin() {
return admin;
}
public void setAdmin(Admin admin) {
this.admin = admin;
}
public String getCreated_date() {
return created_date;
}
public void setCreated_date(String created_date) {
this.created_date = created_date;
}
public String getId() {
return id;
}
public void setId(String id) {
this.id = id;
}
}

View File

@ -0,0 +1,35 @@
package eu.dnetlib.dhp.oa.graph.hostedbymap.model.doaj;
import java.io.Serializable;
import java.util.List;
public class DepositPolicy implements Serializable {
private List<String> service;
private String url;
private Boolean has_policy;
public List<String> getService() {
return service;
}
public void setService(List<String> service) {
this.service = service;
}
public String getUrl() {
return url;
}
public void setUrl(String url) {
this.url = url;
}
public Boolean getHas_policy() {
return has_policy;
}
public void setHas_policy(Boolean has_policy) {
this.has_policy = has_policy;
}
}

View File

@ -0,0 +1,35 @@
package eu.dnetlib.dhp.oa.graph.hostedbymap.model.doaj;
import java.io.Serializable;
import java.util.List;
public class Editorial implements Serializable {
private List<String> review_process;
private String review_url;
private String board_url;
public List<String> getReview_process() {
return review_process;
}
public void setReview_process(List<String> review_process) {
this.review_process = review_process;
}
public String getReview_url() {
return review_url;
}
public void setReview_url(String review_url) {
this.review_url = review_url;
}
public String getBoard_url() {
return board_url;
}
public void setBoard_url(String board_url) {
this.board_url = board_url;
}
}

View File

@ -0,0 +1,25 @@
package eu.dnetlib.dhp.oa.graph.hostedbymap.model.doaj;
import java.io.Serializable;
public class Institution implements Serializable {
private String country;
private String name;
public String getCountry() {
return country;
}
public void setCountry(String country) {
this.country = country;
}
public String getName() {
return name;
}
public void setName(String name) {
this.name = name;
}
}

View File

@ -0,0 +1,67 @@
package eu.dnetlib.dhp.oa.graph.hostedbymap.model.doaj;
import java.io.Serializable;
import com.fasterxml.jackson.annotation.JsonProperty;
public class License implements Serializable {
private Boolean nc;
private Boolean nd;
private Boolean by;
private String type;
private Boolean sa;
private String url;
public Boolean getnC() {
return nc;
}
@JsonProperty("NC")
public void setnC(Boolean NC) {
this.nc = NC;
}
public String getType() {
return type;
}
public void setType(String type) {
this.type = type;
}
public Boolean getNd() {
return nd;
}
@JsonProperty("ND")
public void setNd(Boolean nd) {
this.nd = nd;
}
public Boolean getBy() {
return by;
}
@JsonProperty("BY")
public void setBy(Boolean by) {
this.by = by;
}
public Boolean getSa() {
return sa;
}
@JsonProperty("SA")
public void setSa(Boolean sa) {
this.sa = sa;
}
public String getUrl() {
return url;
}
public void setUrl(String url) {
this.url = url;
}
}

View File

@ -0,0 +1,25 @@
package eu.dnetlib.dhp.oa.graph.hostedbymap.model.doaj;
import java.io.Serializable;
public class Max implements Serializable {
private Integer price;
private String currency;
public Integer getPrice() {
return price;
}
public void setPrice(Integer price) {
this.price = price;
}
public String getCurrency() {
return currency;
}
public void setCurrency(String currency) {
this.currency = currency;
}
}

View File

@ -0,0 +1,25 @@
package eu.dnetlib.dhp.oa.graph.hostedbymap.model.doaj;
import java.io.Serializable;
public class OtherCharges implements Serializable {
private Boolean has_other_charges;
private String url;
public String getUrl() {
return url;
}
public void setUrl(String url) {
this.url = url;
}
public Boolean getHas_other_charges() {
return has_other_charges;
}
public void setHas_other_charges(Boolean has_other_charges) {
this.has_other_charges = has_other_charges;
}
}

View File

@ -0,0 +1,26 @@
package eu.dnetlib.dhp.oa.graph.hostedbymap.model.doaj;
import java.io.Serializable;
import java.util.List;
public class PidScheme implements Serializable {
private List<String> scheme;
private Boolean has_pid_scheme;
public List<String> getScheme() {
return scheme;
}
public void setScheme(List<String> scheme) {
this.scheme = scheme;
}
public Boolean getHas_pid_scheme() {
return has_pid_scheme;
}
public void setHas_pid_scheme(Boolean has_pid_scheme) {
this.has_pid_scheme = has_pid_scheme;
}
}

View File

@ -0,0 +1,27 @@
package eu.dnetlib.dhp.oa.graph.hostedbymap.model.doaj;
import java.io.Serializable;
import javax.sql.rowset.serial.SerialArray;
public class Plagiarism implements Serializable {
private Boolean detection;
private String url;
public Boolean getDetection() {
return detection;
}
public void setDetection(Boolean detection) {
this.detection = detection;
}
public String getUrl() {
return url;
}
public void setUrl(String url) {
this.url = url;
}
}

View File

@ -0,0 +1,44 @@
package eu.dnetlib.dhp.oa.graph.hostedbymap.model.doaj;
import java.io.Serializable;
import java.util.List;
public class Preservation implements Serializable {
private Boolean has_preservation;
private List<String> service;
private List<String> national_library;
private String url;
public String getUrl() {
return url;
}
public void setUrl(String url) {
this.url = url;
}
public Boolean getHas_preservation() {
return has_preservation;
}
public void setHas_preservation(Boolean has_preservation) {
this.has_preservation = has_preservation;
}
public List<String> getService() {
return service;
}
public void setService(List<String> service) {
this.service = service;
}
public List<String> getNational_library() {
return national_library;
}
public void setNational_library(List<String> national_library) {
this.national_library = national_library;
}
}

View File

@ -0,0 +1,25 @@
package eu.dnetlib.dhp.oa.graph.hostedbymap.model.doaj;
import java.io.Serializable;
public class Publisher implements Serializable {
private String country;
private String name;
public String getCountry() {
return country;
}
public void setCountry(String country) {
this.country = country;
}
public String getName() {
return name;
}
public void setName(String name) {
this.name = name;
}
}

View File

@ -0,0 +1,52 @@
package eu.dnetlib.dhp.oa.graph.hostedbymap.model.doaj;
import java.io.Serializable;
public class Ref implements Serializable {
private String aims_scope;
private String journal;
private String oa_statement;
private String author_instructions;
private String license_terms;
public String getAims_scope() {
return aims_scope;
}
public void setAims_scope(String aims_scope) {
this.aims_scope = aims_scope;
}
public String getJournal() {
return journal;
}
public void setJournal(String journal) {
this.journal = journal;
}
public String getOa_statement() {
return oa_statement;
}
public void setOa_statement(String oa_statement) {
this.oa_statement = oa_statement;
}
public String getAuthor_instructions() {
return author_instructions;
}
public void setAuthor_instructions(String author_instructions) {
this.author_instructions = author_instructions;
}
public String getLicense_terms() {
return license_terms;
}
public void setLicense_terms(String license_terms) {
this.license_terms = license_terms;
}
}

View File

@ -0,0 +1,34 @@
package eu.dnetlib.dhp.oa.graph.hostedbymap.model.doaj;
import java.io.Serializable;
public class Subject implements Serializable {
private String code;
private String scheme;
private String term;
public String getCode() {
return code;
}
public void setCode(String code) {
this.code = code;
}
public String getScheme() {
return scheme;
}
public void setScheme(String scheme) {
this.scheme = scheme;
}
public String getTerm() {
return term;
}
public void setTerm(String term) {
this.term = term;
}
}

View File

@ -0,0 +1,25 @@
package eu.dnetlib.dhp.oa.graph.hostedbymap.model.doaj;
import java.io.Serializable;
public class Waiver implements Serializable {
private Boolean has_waiver;
private String url;
public String getUrl() {
return url;
}
public void setUrl(String url) {
this.url = url;
}
public Boolean getHas_waiver() {
return has_waiver;
}
public void setHas_waiver(Boolean has_waiver) {
this.has_waiver = has_waiver;
}
}

View File

@ -74,7 +74,9 @@
<decision name="resume_from">
<switch>
<case to="produceHBM">${wf:conf('resumeFrom') eq 'ProduceHBM'}</case>
<case to="remove_hbmpath">${wf:conf('resumeFrom') eq 'download_csv'}</case>
<case to="fork_downloads_csv">${wf:conf('resumeFrom') eq 'DownloadBoth'}</case>
<case to="downloadGold">${wf:conf('resumeFrom') eq 'DownloadGold'}</case>
<case to="downloadDOAJ">${wf:conf('resumeFrom') eq 'DownloadDoaj'}</case>
<default to="prepareInfo"/>
</switch>
</decision>
@ -83,18 +85,9 @@
<message>Action failed, error message[${wf:errorMessage(wf:lastErrorNode())}]</message>
</kill>
<action name="remove_hbmpath">
<fs>
<delete path="${hostedByMapPath}"/>
<!-- <mkdir path="${hostedByMapPath}"/>-->
</fs>
<ok to="fork_downloads_csv"/>
<error to="Kill"/>
</action>
<fork name="fork_downloads_csv">
<fork name="fork_downloads_csv">
<path start="download_gold"/>
<path start="download_doaj"/>
<path start="download_doaj_json"/>
</fork>
<action name="download_gold">
@ -103,21 +96,43 @@
<arg>--hdfsNameNode</arg><arg>${nameNode}</arg>
<arg>--fileURL</arg><arg>${unibiFileURL}</arg>
<arg>--tmpFile</arg><arg>/tmp/unibi_gold_replaced.csv</arg>
<arg>--outputFile</arg><arg>${workingDir}/unibi_gold.json</arg>
<arg>--outputFile</arg><arg>/user/${wf:user()}/data/unibi_gold.json</arg>
<arg>--classForName</arg><arg>eu.dnetlib.dhp.oa.graph.hostedbymap.model.UnibiGoldModel</arg>
</java>
<ok to="join_download"/>
<error to="Kill"/>
</action>
<action name="download_doaj">
<action name="download_doaj_json">
<shell xmlns="uri:oozie:shell-action:0.2">
<job-tracker>${jobTracker}</job-tracker>
<name-node>${nameNode}</name-node>
<configuration>
<property>
<name>mapred.job.queue.name</name>
<value>${queueName}</value>
</property>
</configuration>
<exec>download.sh</exec>
<argument>${doajJsonFileURL}</argument>
<argument>${dumpPath}</argument>
<argument>${dumpFileName}</argument>
<env-var>HADOOP_USER_NAME=${wf:user()}</env-var>
<file>download.sh</file>
<capture-output/>
</shell>
<ok to="extractTarGzAndMap"/>
<error to="Kill"/>
</action>
<action name="extractTarGzAndMap">
<java>
<main-class>eu.dnetlib.dhp.oa.graph.hostedbymap.DownloadCSV2</main-class>
<main-class>eu.dnetlib.dhp.oa.graph.hostedbymap.ExtractAndMapDoajJson</main-class>
<arg>--hdfsNameNode</arg><arg>${nameNode}</arg>
<arg>--fileURL</arg><arg>${doajFileURL}</arg>
<arg>--tmpFile</arg><arg>/tmp/doaj_replaced.csv</arg>
<arg>--outputFile</arg><arg>${workingDir}/doaj.json</arg>
<arg>--classForName</arg><arg>eu.dnetlib.dhp.oa.graph.hostedbymap.model.DOAJModel</arg>
<arg>--compressedFile</arg><arg>${dumpPath}/${dumpFileName}</arg>
<arg>--workingPath</arg><arg>${workingDir}/DOAJ/</arg>
<arg>--outputPath</arg><arg>/user/${wf:user()}/data/doaj.json</arg>
</java>
<ok to="join_download"/>
<error to="Kill"/>
@ -125,6 +140,54 @@
<join name="join_download" to="produceHBM"/>
<action name="downloadGold">
<java>
<main-class>eu.dnetlib.dhp.oa.graph.hostedbymap.DownloadCSV</main-class>
<arg>--hdfsNameNode</arg><arg>${nameNode}</arg>
<arg>--fileURL</arg><arg>${unibiFileURL}</arg>
<arg>--tmpFile</arg><arg>/tmp/unibi_gold_replaced.csv</arg>
<arg>--outputFile</arg><arg>/user/${wf:user()}/data/unibi_gold.json</arg>
<arg>--classForName</arg><arg>eu.dnetlib.dhp.oa.graph.hostedbymap.model.UnibiGoldModel</arg>
</java>
<ok to="produceHBM"/>
<error to="Kill"/>
</action>
<action name="downloadDOAJ">
<shell xmlns="uri:oozie:shell-action:0.2">
<job-tracker>${jobTracker}</job-tracker>
<name-node>${nameNode}</name-node>
<configuration>
<property>
<name>mapred.job.queue.name</name>
<value>${queueName}</value>
</property>
</configuration>
<exec>download.sh</exec>
<argument>${doajJsonFileURL}</argument>
<argument>${dumpPath}</argument>
<argument>${dumpFileName}</argument>
<env-var>HADOOP_USER_NAME=${wf:user()}</env-var>
<file>download.sh</file>
<capture-output/>
</shell>
<ok to="extract"/>
<error to="Kill"/>
</action>
<action name="extract">
<java>
<main-class>eu.dnetlib.dhp.oa.graph.hostedbymap.ExtractAndMapDoajJson</main-class>
<arg>--hdfsNameNode</arg><arg>${nameNode}</arg>
<arg>--compressedFile</arg><arg>${dumpPath}/${dumpFileName}</arg>
<arg>--workingPath</arg><arg>${workingDir}/DOAJ/</arg>
<arg>--outputPath</arg><arg>/user/${wf:user()}/data/doaj.json</arg>
</java>
<ok to="produceHBM"/>
<error to="Kill"/>
</action>
<action name="produceHBM">
<spark xmlns="uri:oozie:spark-action:0.2">
<master>yarn-cluster</master>

View File

@ -2,9 +2,10 @@ package eu.dnetlib.dhp.oa.graph.hostedbymap
import com.fasterxml.jackson.databind.ObjectMapper
import eu.dnetlib.dhp.application.ArgumentApplicationParser
import eu.dnetlib.dhp.common.HdfsSupport
import eu.dnetlib.dhp.oa.graph.hostedbymap.model.{DOAJModel, UnibiGoldModel}
import eu.dnetlib.dhp.schema.oaf.Datasource
import org.apache.commons.io.IOUtils
import org.apache.commons.io.{FileUtils, IOUtils}
import org.apache.hadoop.conf.Configuration
import org.apache.hadoop.fs.{FileSystem, Path}
import org.apache.hadoop.io.compress.GzipCodec
@ -13,7 +14,8 @@ import org.apache.spark.sql.{Dataset, Encoder, Encoders, SparkSession}
import org.json4s.DefaultFormats
import org.slf4j.{Logger, LoggerFactory}
import java.io.PrintWriter
import java.io.{File, PrintWriter}
import scala.collection.JavaConverters._
object SparkProduceHostedByMap {
@ -171,7 +173,16 @@ object SparkProduceHostedByMap {
}
def doajToHostedbyItemType(doaj: DOAJModel): HostedByItemType = {
if (doaj.getOaStart == null) {
return getHostedByItemType(
Constants.DOAJ,
doaj.getJournalTitle,
doaj.getIssn,
doaj.getEissn,
"",
true
)
}
return getHostedByItemType(
Constants.DOAJ,
doaj.getJournalTitle,
@ -256,6 +267,8 @@ object SparkProduceHostedByMap {
logger.info("Getting the Datasources")
HdfsSupport.remove(outputPath, spark.sparkContext.hadoopConfiguration)
Aggregators
.explodeHostedByItemType(
oaHostedByDataset(spark, datasourcePath)

View File

@ -55,7 +55,6 @@ public class DownloadCsvTest {
new DownloadCSV()
.doDownload(
fileURL,
workingDir + "/unibi_gold",
outputFile,
UnibiGoldModel.class.getName(),
',',
@ -91,56 +90,6 @@ public class DownloadCsvTest {
assertEquals(67028, count);
}
@Disabled
@Test
void getDoajFileTest() throws CollectorException, IOException, ClassNotFoundException {
String fileURL = "https://doaj.org/csv";
final String outputFile = workingDir + "/doaj.json";
new DownloadCSV()
.doDownload(
fileURL,
workingDir + "/doaj",
outputFile,
DOAJModel.class.getName(),
',',
fs);
BufferedReader in = new BufferedReader(new InputStreamReader(fs.open(new Path(outputFile))));
String line;
int count = 0;
while ((line = in.readLine()) != null) {
DOAJModel doaj = new ObjectMapper().readValue(line, DOAJModel.class);
if (count == 0) {
assertEquals("0001-3765", doaj.getIssn());
assertEquals("1678-2690", doaj.getEissn());
assertEquals("Anais da Academia Brasileira de Ciências", doaj.getJournalTitle());
}
if (count == 22) {
log.info(new ObjectMapper().writeValueAsString(doaj));
System.out.println(new ObjectMapper().writeValueAsString(doaj));
}
if (count == 7904) {
// log.info(new ObjectMapper().writeValueAsString(doaj));
assertEquals("", doaj.getIssn());
assertEquals("2055-7159", doaj.getEissn());
assertEquals("BJR|case reports", doaj.getJournalTitle());
}
if (count == 16707) {
assertEquals("2783-1043", doaj.getIssn());
assertEquals("2783-1051", doaj.getEissn());
assertEquals("فیزیک کاربردی ایران", doaj.getJournalTitle());
}
count += 1;
}
assertEquals(16715, count);
}
@AfterAll
public static void cleanup() {
FileUtils.deleteQuietly(new File(workingDir));

View File

@ -1,25 +1,25 @@
{"journalTitle":"Lëd i Sneg","issn":"2076-6734","eissn":"2412-3765","reviewProcess":"Double blind peer review"}
{"journalTitle":"Компьютерные исследования и моделирование","issn":"2076-7633","eissn":"2077-6853","reviewProcess":"Blind peer review"}
{"journalTitle":" Историко-биологические исследования","issn":"2076-8176","eissn":"2500-1221","reviewProcess":"Double blind peer review"}
{"journalTitle":"Інформаційні технології і засоби навчання","issn":"2076-8184","eissn":"","reviewProcess":"Double blind peer review"}
{"journalTitle":"Revue Internationale de Pédagogie de lEnseignement Supérieur","issn":"","eissn":"2076-8427","reviewProcess":"Double blind peer review"}
{"journalTitle":"Проблемы развития территории","issn":"2076-8915","eissn":"2409-9007","reviewProcess":"Double blind peer review"}
{"journalTitle":"Rambam Maimonides Medical Journal","issn":"","eissn":"2076-9172","reviewProcess":"Peer review"}
{"journalTitle":"Membranes","issn":"2077-0375","eissn":"","reviewProcess":"Blind peer review"}
{"journalTitle":"Journal of Clinical Medicine","issn":"","eissn":"2077-0383","reviewProcess":"Blind peer review"}
{"journalTitle":"Agriculture","issn":"","eissn":"2077-0472","reviewProcess":"Blind peer review"}
{"journalTitle":"Standartnye Obrazcy","issn":"2077-1177","eissn":"","reviewProcess":"Double blind peer review"}
{"journalTitle":"Металл и литье Украины","issn":"2077-1304","eissn":"2706-5529","reviewProcess":"Double blind peer review"}
{"journalTitle":"Journal of Marine Science and Engineering","issn":"","eissn":"2077-1312","reviewProcess":"Blind peer review"}
{"journalTitle":"Religions","issn":"","eissn":"2077-1444","reviewProcess":"Double blind peer review"}
{"journalTitle":"GW-Unterricht","issn":"2077-1517","eissn":"2414-4169","reviewProcess":"Double blind peer review"}
{"journalTitle":"UCV-Scientia","issn":"2077-172X","eissn":"","reviewProcess":"Peer review"}
{"journalTitle":"Sovremennye Issledovaniâ Socialʹnyh Problem","issn":"2077-1770","eissn":"2218-7405","reviewProcess":"Double blind peer review"}
{"journalTitle":"Granì","issn":"2077-1800","eissn":"2413-8738","reviewProcess":"Double blind peer review"}
{"journalTitle":"Journal of Economics Finance and Administrative Science","issn":"2077-1886","eissn":"2218-0648","reviewProcess":"Double blind peer review"}
{"journalTitle":"Science Education International","issn":"","eissn":"2077-2327","reviewProcess":"Double blind peer review"}
{"journalTitle":"Edumecentro","issn":"","eissn":"2077-2874","reviewProcess":"Double blind peer review"}
{"journalTitle":"Monteverdia","issn":"","eissn":"2077-2890","reviewProcess":"Double blind peer review"}
{"journalTitle":"Transformación","issn":"","eissn":"2077-2955","reviewProcess":"Double blind peer review"}
{"journalTitle":"Journal of Space Technology","issn":"2077-3099","eissn":"2411-5029","reviewProcess":"Double blind peer review"}
{"journalTitle":"Revue de Primatologie","issn":"","eissn":"2077-3757","reviewProcess":"Peer review"}
{"journalTitle":"Lëd i Sneg","issn":"2076-6734","eissn":"2412-3765","reviewProcess":["Double blind peer review"],"oaStart":2015}
{"journalTitle":"Компьютерные исследования и моделирование","issn":"2076-7633","eissn":"2077-6853","reviewProcess":["Blind peer review"],"oaStart":2009}
{"journalTitle":" Историко-биологические исследования","issn":"2076-8176","eissn":"2500-1221","reviewProcess":["Double blind peer review"],"oaStart":2010}
{"journalTitle":"Інформаційні технології і засоби навчання","issn":"2076-8184","eissn":null,"reviewProcess":["Double blind peer review"],"oaStart":2006}
{"journalTitle":"Revue Internationale de Pédagogie de lEnseignement Supérieur","issn":null,"eissn":"2076-8427","reviewProcess":["Double blind peer review"],"oaStart":2009}
{"journalTitle":"Проблемы развития территории","issn":"2076-8915","eissn":"2409-9007","reviewProcess":["Double blind peer review"],"oaStart":2008}
{"journalTitle":"Rambam Maimonides Medical Journal","issn":null,"eissn":"2076-9172","reviewProcess":["Peer review"],"oaStart":2010}
{"journalTitle":"Membranes","issn":"2077-0375","eissn":null,"reviewProcess":["Blind peer review"],"oaStart":2011}
{"journalTitle":"Journal of Clinical Medicine","issn":null,"eissn":"2077-0383","reviewProcess":["Blind peer review"],"oaStart":2012}
{"journalTitle":"Agriculture","issn":null,"eissn":"2077-0472","reviewProcess":["Blind peer review"],"oaStart":2011}
{"journalTitle":"Standartnye Obrazcy","issn":"2077-1177","eissn":null,"reviewProcess":["Double blind peer review"],"oaStart":2014}
{"journalTitle":"Металл и литье Украины","issn":"2077-1304","eissn":"2706-5529","reviewProcess":["Double blind peer review"],"oaStart":2019}
{"journalTitle":"Journal of Marine Science and Engineering","issn":null,"eissn":"2077-1312","reviewProcess":["Blind peer review"],"oaStart":2013}
{"journalTitle":"Religions","issn":null,"eissn":"2077-1444","reviewProcess":["Double blind peer review"],"oaStart":2010}
{"journalTitle":"GW-Unterricht","issn":"2077-1517","eissn":"2414-4169","reviewProcess":["Double blind peer review"],"oaStart":2010}
{"journalTitle":"UCV-Scientia","issn":"2077-172X","eissn":null,"reviewProcess":["Peer review"],"oaStart":2009}
{"journalTitle":"Sovremennye Issledovaniâ Socialʹnyh Problem","issn":"2077-1770","eissn":"2218-7405","reviewProcess":["Double blind peer review"],"oaStart":2010}
{"journalTitle":"Granì","issn":"2077-1800","eissn":"2413-8738","reviewProcess":["Double blind peer review"],"oaStart":2014}
{"journalTitle":"Journal of Economics Finance and Administrative Science","issn":"2077-1886","eissn":"2218-0648","reviewProcess":["Double blind peer review"],"oaStart":2017}
{"journalTitle":"Science Education International","issn":null,"eissn":"2077-2327","reviewProcess":["Double blind peer review"],"oaStart":2017}
{"journalTitle":"Edumecentro","issn":null,"eissn":"2077-2874","reviewProcess":["Double blind peer review"],"oaStart":2013}
{"journalTitle":"Monteverdia","issn":null,"eissn":"2077-2890","reviewProcess":["Double blind peer review"],"oaStart":2008}
{"journalTitle":"Transformación","issn":null,"eissn":"2077-2955","reviewProcess":["Double blind peer review"],"oaStart":2010}
{"journalTitle":"Journal of Space Technology","issn":"2077-3099","eissn":"2411-5029","reviewProcess":["Double blind peer review"],"oaStart":2011}
{"journalTitle":"Revue de Primatologie","issn":null,"eissn":"2077-3757","reviewProcess":["Peer review"],"oaStart":2009}