Fulltext fetcher for IACR eprints (#9651)

* Add fulltext fetcher for IACR eprints * Add IACR full text fetcher to CHANGELOG * Add tests and fixes for IACR full text fetcher * Make checkstyle happy * simplified test * Update CHANGELOG.md * Update CHANGELOG.md --------- Co-authored-by: Siedlerchr <siedlerkiller@gmail.com> Co-authored-by: Carl Christian Snethlage <50491877+calixtus@users.noreply.github.com>
JabRef · Mar 6, 2023 · 85ab410 · 85ab410
1 parent bcd808b
commit 85ab410
Show file tree

Hide file tree

Showing 4 changed files with 74 additions and 2 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -12,7 +12,10 @@ Note that this project **does not** adhere to [Semantic Versioning](http://semve
 ### Added
 
 - We added a field showing the BibTeX/biblatex source for added and deleted entries in the "External Changes Resolver" dialog. [#9509](https://github.com/JabRef/jabref/issues/9509)
-- Add "Attach file from URL" to right-click context menu which downloads file from URL and stores it with reference library.
+- We added a full text fetcher for IACR eprints. [#9651](https://github.com/JabRef/jabref/pull/9651)
+- We added "Attach file from URL" to right-click context menu to download and store a file with the reference library. [#9646](https://github.com/JabRef/jabref/issues/9646)
+- We enabled updating an existing entry with data from InspireHEP. [#9351](https://github.com/JabRef/jabref/issues/9351)
+
 
 
 

diff --git a/src/main/java/org/jabref/logic/importer/WebFetchers.java b/src/main/java/org/jabref/logic/importer/WebFetchers.java
@@ -196,6 +196,7 @@ public static Set<FulltextFetcher> getFullTextFetchers(ImportFormatPreferences i
         fetchers.add(new ArXivFetcher(importFormatPreferences));
         fetchers.add(new IEEE(importFormatPreferences, importerPreferences));
         fetchers.add(new ApsFetcher());
+        fetchers.add(new IacrEprintFetcher(importFormatPreferences));
 
         // Meta search
         // fetchers.add(new JstorFetcher(importFormatPreferences));

diff --git a/src/main/java/org/jabref/logic/importer/fetcher/IacrEprintFetcher.java b/src/main/java/org/jabref/logic/importer/fetcher/IacrEprintFetcher.java
@@ -1,11 +1,14 @@
 package org.jabref.logic.importer.fetcher;
 
 import java.io.IOException;
+import java.net.URL;
+import java.util.Objects;
 import java.util.Optional;
 import java.util.function.Predicate;
 import java.util.regex.Pattern;
 
 import org.jabref.logic.importer.FetcherException;
+import org.jabref.logic.importer.FulltextFetcher;
 import org.jabref.logic.importer.IdBasedFetcher;
 import org.jabref.logic.importer.ImportFormatPreferences;
 import org.jabref.logic.importer.ParseException;
@@ -17,7 +20,7 @@
 import org.jabref.model.strings.StringUtil;
 import org.jabref.model.util.DummyFileUpdateMonitor;
 
-public class IacrEprintFetcher implements IdBasedFetcher {
+public class IacrEprintFetcher implements FulltextFetcher, IdBasedFetcher {
 
     public static final String NAME = "IACR eprints";
 
@@ -26,6 +29,7 @@ public class IacrEprintFetcher implements IdBasedFetcher {
     private static final Predicate<String> IDENTIFIER_PREDICATE = Pattern.compile("\\d{4}/\\d{3,5}").asPredicate();
     private static final String CITATION_URL_PREFIX = "https://eprint.iacr.org/";
     private static final String DESCRIPTION_URL_PREFIX = "https://eprint.iacr.org/";
+    private static final String FULLTEXT_URL_PREFIX = "https://eprint.iacr.org/";
     private static final String VERSION_URL_PREFIX = "https://eprint.iacr.org/archive/versions/";
 
     private final ImportFormatPreferences prefs;
@@ -130,4 +134,27 @@ private boolean isFromOrAfterYear2000(BibEntry entry) throws FetcherException {
     public String getName() {
         return NAME;
     }
+
+    @Override
+    public Optional<URL> findFullText(BibEntry entry) throws IOException, FetcherException {
+        Objects.requireNonNull(entry);
+
+        Optional<String> urlField = entry.getField(StandardField.URL);
+        if (urlField.isPresent()) {
+            String descriptiveHtml = getHtml(urlField.get());
+            String startOfFulltextLink = "<a class=\"btn btn-sm btn-outline-dark\"";
+            String fulltextLinkAsInHtml = getRequiredValueBetween(startOfFulltextLink, ".pdf", descriptiveHtml);
+            // There is an additional "\n           href=\"/archive/" we have to remove - and for some reason,
+            // getRequiredValueBetween refuses to match across the line break.
+            fulltextLinkAsInHtml = fulltextLinkAsInHtml.replaceFirst(".*href=\"/", "").trim();
+            String fulltextLink = FULLTEXT_URL_PREFIX + fulltextLinkAsInHtml + ".pdf";
+            return Optional.of(new URL(fulltextLink));
+        }
+        return Optional.empty();
+    }
+
+    @Override
+    public TrustLevel getTrustLevel() {
+        return TrustLevel.PREPRINT;
+    }
 }
diff --git a/src/test/java/org/jabref/logic/importer/fetcher/IacrEprintFetcherTest.java b/src/test/java/org/jabref/logic/importer/fetcher/IacrEprintFetcherTest.java
@@ -1,5 +1,7 @@
 package org.jabref.logic.importer.fetcher;
 
+import java.io.IOException;
+import java.net.URL;
 import java.util.ArrayList;
 import java.util.Arrays;
 import java.util.Collection;
@@ -36,6 +38,7 @@ public class IacrEprintFetcherTest {
 
     private IacrEprintFetcher fetcher;
     private BibEntry abram2017;
+    private BibEntry abram2017noVersion;
     private BibEntry beierle2016;
     private BibEntry delgado2017;
 
@@ -55,6 +58,17 @@ public void setUp() {
                 .withField(StandardField.VERSION, "20171124:064527")
                 .withField(StandardField.YEAR, "2017");
 
+        abram2017noVersion = new BibEntry(StandardEntryType.Misc)
+                .withCitationKey("cryptoeprint:2017/1118")
+                .withField(StandardField.ABSTRACT, "dummy")
+                .withField(StandardField.AUTHOR, "Ittai Abraham and Dahlia Malkhi and Kartik Nayak and Ling Ren and Alexander Spiegelman")
+                .withField(StandardField.DATE, "2017-11-24")
+                .withField(StandardField.HOWPUBLISHED, "Cryptology ePrint Archive, Paper 2017/1118")
+                .withField(StandardField.NOTE, "\\url{https://eprint.iacr.org/2017/1118}")
+                .withField(StandardField.TITLE, "Solida: A Blockchain Protocol Based on Reconfigurable Byzantine Consensus")
+                .withField(StandardField.URL, "https://eprint.iacr.org/2017/1118")
+                .withField(StandardField.YEAR, "2017");
+
         beierle2016 = new BibEntry(StandardEntryType.Misc)
                 .withCitationKey("cryptoeprint:2016/119")
                 .withField(StandardField.ABSTRACT, "dummy")
@@ -185,4 +199,31 @@ private static Stream<String> allNonWithdrawnIdsWithOldHtmlFormat() {
         ids.removeAll(withdrawnIds);
         return ids.stream();
     }
+
+    @Test
+    public void getFulltextWithVersion() throws FetcherException, IOException {
+        Optional<URL> pdfUrl = fetcher.findFullText(abram2017);
+        assertEquals(Optional.of("https://eprint.iacr.org/archive/2017/1118/1511505927.pdf"), pdfUrl.map(URL::toString));
+    }
+
+    @Test
+    public void getFulltextWithoutVersion() throws FetcherException, IOException {
+        Optional<URL> pdfUrl = fetcher.findFullText(abram2017noVersion);
+        assertEquals(Optional.of("https://eprint.iacr.org/2017/1118.pdf"), pdfUrl.map(URL::toString));
+    }
+
+    @Test
+    public void getFulltextWithoutUrl() throws FetcherException, IOException {
+        BibEntry abram2017WithoutUrl = abram2017;
+        abram2017WithoutUrl.clearField(StandardField.URL);
+        Optional<URL> pdfUrl = fetcher.findFullText(abram2017WithoutUrl);
+        assertEquals(Optional.empty(), pdfUrl);
+    }
+
+    @Test
+    public void getFulltextWithNonIACRUrl() throws IOException {
+        BibEntry abram2017WithNonIACRUrl = abram2017;
+        abram2017WithNonIACRUrl.setField(StandardField.URL, "https://example.com");
+        assertThrows(FetcherException.class, () -> fetcher.findFullText(abram2017WithNonIACRUrl));
+    }
 }