Add plain citation parsing with LLM

InAnYan · Sep 25, 2024 · 7d0c04f · 7d0c04f
1 parent 91d0709
commit 7d0c04f
Show file tree

Hide file tree

Showing 11 changed files with 217 additions and 23 deletions.
diff --git a/src/main/java/org/jabref/gui/bibtexextractor/BibtexExtractorViewModel.java b/src/main/java/org/jabref/gui/bibtexextractor/BibtexExtractorViewModel.java
@@ -4,15 +4,23 @@
 
 import javax.swing.undo.UndoManager;
 
+import javafx.beans.property.ListProperty;
+import javafx.beans.property.ObjectProperty;
+import javafx.beans.property.SimpleListProperty;
+import javafx.beans.property.SimpleObjectProperty;
 import javafx.beans.property.SimpleStringProperty;
 import javafx.beans.property.StringProperty;
+import javafx.collections.FXCollections;
 
 import org.jabref.gui.DialogService;
 import org.jabref.gui.StateManager;
 import org.jabref.gui.externalfiles.ImportHandler;
 import org.jabref.gui.preferences.GuiPreferences;
+import org.jabref.logic.ai.AiService;
 import org.jabref.logic.importer.FetcherException;
 import org.jabref.logic.importer.fetcher.GrobidCitationFetcher;
+import org.jabref.logic.importer.fetcher.LlmCitationFetcher;
+import org.jabref.logic.importer.fetcher.OnlinePlainCitationParser;
 import org.jabref.logic.l10n.Localization;
 import org.jabref.logic.preferences.CliPreferences;
 import org.jabref.logic.util.BackgroundTask;
@@ -35,23 +43,29 @@ public class BibtexExtractorViewModel {
     private static final Logger LOGGER = LoggerFactory.getLogger(BibtexExtractorViewModel.class);
 
     private final boolean onlineMode;
+
     private final DialogService dialogService;
+    private final AiService aiService;
     private final CliPreferences preferences;
     private final TaskExecutor taskExecutor;
 
     private final ImportHandler importHandler;
     private final StringProperty inputTextProperty = new SimpleStringProperty("");
+    private final ListProperty<OnlinePlainCitationParser> onlinePlainCitationParsers = new SimpleListProperty<>(FXCollections.observableArrayList(OnlinePlainCitationParser.values()));
+    private final ObjectProperty<OnlinePlainCitationParser> parserChoice;
 
     public BibtexExtractorViewModel(boolean onlineMode,
                                     BibDatabaseContext bibdatabaseContext,
                                     DialogService dialogService,
+                                    AiService aiService,
                                     GuiPreferences preferences,
                                     FileUpdateMonitor fileUpdateMonitor,
                                     TaskExecutor taskExecutor,
                                     UndoManager undoManager,
                                     StateManager stateManager) {
         this.onlineMode = onlineMode;
         this.dialogService = dialogService;
+        this.aiService = aiService;
         this.preferences = preferences;
         this.taskExecutor = taskExecutor;
         this.importHandler = new ImportHandler(
@@ -62,6 +76,8 @@ public BibtexExtractorViewModel(boolean onlineMode,
                 stateManager,
                 dialogService,
                 taskExecutor);
+
+        this.parserChoice = new SimpleObjectProperty<>(preferences.getImporterPreferences().defaultOnlinePlainCitationParserProperty().get());
     }
 
     public void startParsing() {
@@ -78,25 +94,38 @@ private void startParsingOffline() {
     }
 
     private void startParsingOnline() {
-        GrobidCitationFetcher grobidCitationFetcher = new GrobidCitationFetcher(preferences.getGrobidPreferences(), preferences.getImportFormatPreferences());
-        BackgroundTask.wrap(() -> grobidCitationFetcher.performSearch(inputTextProperty.getValue()))
-                      .onRunning(() -> dialogService.notify(Localization.lang("Your text is being parsed...")))
-                      .onFailure(e -> {
-                          if (e instanceof FetcherException) {
-                              String msg = Localization.lang("There are connection issues with a JabRef server. Detailed information: %0",
-                                      e.getMessage());
-                              dialogService.notify(msg);
-                          } else {
-                              LOGGER.warn("Missing exception handling.", e);
-                          }
-                      })
-                      .onSuccess(parsedEntries -> {
-                          dialogService.notify(Localization.lang("%0 entries were parsed from your query.", String.valueOf(parsedEntries.size())));
-                          importHandler.importEntries(parsedEntries);
-                      }).executeWith(taskExecutor);
+        BackgroundTask
+                .wrap(() -> switch (parserChoice.get()) {
+                    case GROBID ->
+                            new GrobidCitationFetcher(preferences.getGrobidPreferences(), preferences.getImportFormatPreferences()).performSearch(inputTextProperty.getValue());
+                    case LLM ->
+                            new LlmCitationFetcher(preferences.getImportFormatPreferences(), aiService.getChatLanguageModel()).performSearch(inputTextProperty.getValue());
+                })
+                .onRunning(() -> dialogService.notify(Localization.lang("Your text is being parsed...")))
+                .onFailure(e -> {
+                    if (e instanceof FetcherException) {
+                        String msg = Localization.lang("There are connection issues with a JabRef server. Detailed information: %0",
+                                e.getMessage());
+                        dialogService.notify(msg);
+                    } else {
+                        LOGGER.warn("Missing exception handling.", e);
+                    }
+                })
+                .onSuccess(parsedEntries -> {
+                    dialogService.notify(Localization.lang("%0 entries were parsed from your query.", String.valueOf(parsedEntries.size())));
+                    importHandler.importEntries(parsedEntries);
+                }).executeWith(taskExecutor);
     }
 
     public StringProperty inputTextProperty() {
         return this.inputTextProperty;
     }
+
+    public ListProperty<OnlinePlainCitationParser> onlinePlainCitationParsers() {
+        return this.onlinePlainCitationParsers;
+    }
+
+    public ObjectProperty<OnlinePlainCitationParser> parserChoice() {
+        return this.parserChoice;
+    }
 }
diff --git a/src/main/java/org/jabref/gui/bibtexextractor/ExtractBibtexDialog.fxml b/src/main/java/org/jabref/gui/bibtexextractor/ExtractBibtexDialog.fxml
@@ -4,12 +4,21 @@
 <?import javafx.scene.control.DialogPane?>
 <?import javafx.scene.control.TextArea?>
 <?import javafx.scene.layout.VBox?>
+<?import javafx.scene.layout.HBox?>
+<?import javafx.scene.control.Label?>
+<?import javafx.scene.control.ComboBox?>
 <DialogPane prefHeight="430.0" prefWidth="586.0" xmlns="http://javafx.com/javafx/8.0.171"
             xmlns:fx="http://javafx.com/fxml/1" fx:controller="org.jabref.gui.bibtexextractor.ExtractBibtexDialog">
     <content>
-        <VBox fx:id="contentVbox" minHeight="-Infinity" prefHeight="200.0" prefWidth="100.0">
+        <VBox fx:id="contentVbox" spacing="10" minHeight="-Infinity" prefHeight="200.0" prefWidth="100.0">
             <children>
-                <TextArea fx:id="input" minHeight="-Infinity" prefHeight="350.0" prefWidth="586.0" wrapText="true"/>
+                <TextArea fx:id="input" minHeight="-Infinity" prefHeight="450.0" prefWidth="586.0" wrapText="true"/>
+                <HBox alignment="BASELINE_LEFT" spacing="10">
+                    <children>
+                        <Label text="%Parser choice (only for online extraction):"/>
+                        <ComboBox fx:id="parserChoice" HBox.hgrow="ALWAYS"/>
+                    </children>
+                </HBox>
             </children>
         </VBox>
     </content>

diff --git a/src/main/java/org/jabref/gui/bibtexextractor/ExtractBibtexDialog.java b/src/main/java/org/jabref/gui/bibtexextractor/ExtractBibtexDialog.java
@@ -6,6 +6,7 @@
 import javafx.fxml.FXML;
 import javafx.scene.control.Button;
 import javafx.scene.control.ButtonType;
+import javafx.scene.control.ComboBox;
 import javafx.scene.control.TextArea;
 import javafx.scene.control.Tooltip;
 
@@ -14,6 +15,9 @@
 import org.jabref.gui.StateManager;
 import org.jabref.gui.preferences.GuiPreferences;
 import org.jabref.gui.util.BaseDialog;
+import org.jabref.gui.util.ViewModelListCellFactory;
+import org.jabref.logic.ai.AiService;
+import org.jabref.logic.importer.fetcher.OnlinePlainCitationParser;
 import org.jabref.logic.l10n.Localization;
 import org.jabref.logic.util.TaskExecutor;
 import org.jabref.model.database.BibDatabaseContext;
@@ -34,21 +38,25 @@ public class ExtractBibtexDialog extends BaseDialog<Void> {
 
     @Inject protected StateManager stateManager;
     @Inject protected DialogService dialogService;
+    @Inject protected AiService aiService;
     @Inject protected FileUpdateMonitor fileUpdateMonitor;
     @Inject protected TaskExecutor taskExecutor;
     @Inject protected UndoManager undoManager;
     @Inject protected GuiPreferences preferences;
 
     @FXML protected TextArea input;
-    private final boolean onlineMode;
+    @FXML protected ButtonType parseButtonType;
+    @FXML protected ComboBox<OnlinePlainCitationParser> parserChoice;
 
-    @FXML private ButtonType parseButtonType;
+    private final boolean onlineMode;
 
     public ExtractBibtexDialog(boolean onlineMode) {
         this.onlineMode = onlineMode;
+
         ViewLoader.view(this)
                   .load()
                   .setAsDialogPane(this);
+
         if (onlineMode) {
             this.setTitle(Localization.lang("Plain References Parser (online)"));
         } else {
@@ -59,16 +67,25 @@ public ExtractBibtexDialog(boolean onlineMode) {
     @FXML
     private void initialize() {
         BibDatabaseContext database = stateManager.getActiveDatabase().orElseThrow(() -> new NullPointerException("Database null"));
+
         BibtexExtractorViewModel viewModel = new BibtexExtractorViewModel(
                 onlineMode,
                 database,
                 dialogService,
+                aiService,
                 preferences,
                 fileUpdateMonitor,
                 taskExecutor,
                 undoManager,
                 stateManager);
 
+        new ViewModelListCellFactory<OnlinePlainCitationParser>()
+                .withText(OnlinePlainCitationParser::getLocalizedName)
+                .install(parserChoice);
+        parserChoice.getItems().setAll(viewModel.onlinePlainCitationParsers());
+        parserChoice.valueProperty().bindBidirectional(viewModel.parserChoice());
+        parserChoice.setDisable(!onlineMode);
+
         input.textProperty().bindBidirectional(viewModel.inputTextProperty());
         String clipText = ClipBoardManager.getContents();
         if (StringUtil.isBlank(clipText)) {

diff --git a/src/main/java/org/jabref/gui/preferences/websearch/WebSearchTab.fxml b/src/main/java/org/jabref/gui/preferences/websearch/WebSearchTab.fxml
@@ -10,6 +10,7 @@
 <?import javafx.scene.layout.GridPane?>
 <?import javafx.scene.layout.HBox?>
 <?import javafx.scene.layout.VBox?>
+<?import javafx.scene.control.ComboBox?>
 <fx:root spacing="10.0" type="VBox"
          xmlns="http://javafx.com/javafx" xmlns:fx="http://javafx.com/fxml"
          fx:controller="org.jabref.gui.preferences.websearch.WebSearchTab">
@@ -21,6 +22,10 @@
     <CheckBox fx:id="warnAboutDuplicatesOnImport" text="%Warn about duplicates on import"/>
     <CheckBox fx:id="downloadLinkedOnlineFiles" text="%Download linked online files"/>
     <CheckBox fx:id="keepDownloadUrl" text="%Store url for downloaded file" />
+    <HBox alignment="BASELINE_LEFT" spacing="10">
+        <Label text="%Default online plain citation parser"/>
+        <ComboBox fx:id="defaultOnlinePlainCitationParser" HBox.hgrow="ALWAYS"/>
+    </HBox>
 
     <Label styleClass="sectionHeader" text="%Custom DOI URI"/>
     <HBox alignment="CENTER_LEFT" spacing="10.0">

diff --git a/src/main/java/org/jabref/gui/preferences/websearch/WebSearchTab.java b/src/main/java/org/jabref/gui/preferences/websearch/WebSearchTab.java
@@ -4,6 +4,7 @@
 import javafx.fxml.FXML;
 import javafx.scene.control.Button;
 import javafx.scene.control.CheckBox;
+import javafx.scene.control.ComboBox;
 import javafx.scene.control.SplitPane;
 import javafx.scene.control.TableColumn;
 import javafx.scene.control.TableView;
@@ -16,7 +17,9 @@
 import org.jabref.gui.preferences.AbstractPreferenceTabView;
 import org.jabref.gui.preferences.PreferencesTab;
 import org.jabref.gui.slr.StudyCatalogItem;
+import org.jabref.gui.util.ViewModelListCellFactory;
 import org.jabref.gui.util.ViewModelTableRowFactory;
+import org.jabref.logic.importer.fetcher.OnlinePlainCitationParser;
 import org.jabref.logic.l10n.Localization;
 import org.jabref.logic.preferences.FetcherApiKey;
 
@@ -30,6 +33,7 @@ public class WebSearchTab extends AbstractPreferenceTabView<WebSearchTabViewMode
     @FXML private CheckBox warnAboutDuplicatesOnImport;
     @FXML private CheckBox downloadLinkedOnlineFiles;
     @FXML private CheckBox keepDownloadUrl;
+    @FXML private ComboBox<OnlinePlainCitationParser> defaultOnlinePlainCitationParser;
 
     @FXML private CheckBox useCustomDOI;
     @FXML private TextField useCustomDOIName;
@@ -69,6 +73,12 @@ public void initialize() {
         downloadLinkedOnlineFiles.selectedProperty().bindBidirectional(viewModel.shouldDownloadLinkedOnlineFiles());
         keepDownloadUrl.selectedProperty().bindBidirectional(viewModel.shouldKeepDownloadUrl());
 
+        new ViewModelListCellFactory<OnlinePlainCitationParser>()
+                .withText(OnlinePlainCitationParser::getLocalizedName)
+                .install(defaultOnlinePlainCitationParser);
+        defaultOnlinePlainCitationParser.itemsProperty().bind(viewModel.onlinePlainCitationParsers());
+        defaultOnlinePlainCitationParser.valueProperty().bindBidirectional(viewModel.defaultOnlinePlainCitationParserProperty());
+
         grobidEnabled.selectedProperty().bindBidirectional(viewModel.grobidEnabledProperty());
         grobidURL.textProperty().bindBidirectional(viewModel.grobidURLProperty());
         grobidURL.disableProperty().bind(grobidEnabled.selectedProperty().not());

diff --git a/src/main/java/org/jabref/gui/preferences/websearch/WebSearchTabViewModel.java b/src/main/java/org/jabref/gui/preferences/websearch/WebSearchTabViewModel.java
@@ -32,6 +32,7 @@
 import org.jabref.logic.importer.fetcher.CompositeSearchBasedFetcher;
 import org.jabref.logic.importer.fetcher.CustomizableKeyFetcher;
 import org.jabref.logic.importer.fetcher.GrobidPreferences;
+import org.jabref.logic.importer.fetcher.OnlinePlainCitationParser;
 import org.jabref.logic.l10n.Localization;
 import org.jabref.logic.net.URLDownload;
 import org.jabref.logic.os.OS;
@@ -47,6 +48,8 @@ public class WebSearchTabViewModel implements PreferenceTabViewModel {
     private final BooleanProperty warnAboutDuplicatesOnImportProperty = new SimpleBooleanProperty();
     private final BooleanProperty shouldDownloadLinkedOnlineFiles = new SimpleBooleanProperty();
     private final BooleanProperty shouldkeepDownloadUrl = new SimpleBooleanProperty();
+    private final ListProperty<OnlinePlainCitationParser> onlinePlainCitationParsers = new SimpleListProperty<>(FXCollections.observableArrayList(OnlinePlainCitationParser.values()));
+    private final ObjectProperty<OnlinePlainCitationParser> defaultOnlinePlainCitationParser = new SimpleObjectProperty<>();
 
     private final BooleanProperty useCustomDOIProperty = new SimpleBooleanProperty();
     private final StringProperty useCustomDOINameProperty = new SimpleStringProperty("");
@@ -85,6 +88,8 @@ public void setValues() {
         warnAboutDuplicatesOnImportProperty.setValue(importerPreferences.shouldWarnAboutDuplicatesOnImport());
         shouldDownloadLinkedOnlineFiles.setValue(filePreferences.shouldDownloadLinkedFiles());
         shouldkeepDownloadUrl.setValue(filePreferences.shouldKeepDownloadUrl());
+        defaultOnlinePlainCitationParser.setValue(importerPreferences.getDefaultOnlinePlainCitationParser());
+
         useCustomDOIProperty.setValue(doiPreferences.isUseCustom());
         useCustomDOINameProperty.setValue(doiPreferences.getDefaultBaseURI());
 
@@ -112,6 +117,7 @@ public void storeSettings() {
         importerPreferences.setWarnAboutDuplicatesOnImport(warnAboutDuplicatesOnImportProperty.getValue());
         filePreferences.setDownloadLinkedFiles(shouldDownloadLinkedOnlineFiles.getValue());
         filePreferences.setKeepDownloadUrl(shouldkeepDownloadUrl.getValue());
+        importerPreferences.setDefaultOnlinePlainCitationParser(defaultOnlinePlainCitationParser.getValue());
         grobidPreferences.setGrobidEnabled(grobidEnabledProperty.getValue());
         grobidPreferences.setGrobidOptOut(grobidPreferences.isGrobidOptOut());
         grobidPreferences.setGrobidURL(grobidURLProperty.getValue());
@@ -137,6 +143,14 @@ public BooleanProperty generateKeyOnImportProperty() {
         return generateKeyOnImportProperty;
     }
 
+    public ListProperty<OnlinePlainCitationParser> onlinePlainCitationParsers() {
+        return onlinePlainCitationParsers;
+    }
+
+    public ObjectProperty<OnlinePlainCitationParser> defaultOnlinePlainCitationParserProperty() {
+        return defaultOnlinePlainCitationParser;
+    }
+
     public BooleanProperty useCustomDOIProperty() {
         return this.useCustomDOIProperty;
     }

diff --git a/src/main/java/org/jabref/logic/importer/ImporterPreferences.java b/src/main/java/org/jabref/logic/importer/ImporterPreferences.java
@@ -13,11 +13,11 @@
 import javafx.collections.ObservableList;
 import javafx.collections.ObservableSet;
 
+import org.jabref.logic.importer.fetcher.OnlinePlainCitationParser;
 import org.jabref.logic.importer.fileformat.CustomImporter;
 import org.jabref.logic.preferences.FetcherApiKey;
 
 public class ImporterPreferences {
-
     private final BooleanProperty importerEnabled;
     private final BooleanProperty generateNewKeyOnImport;
     private final BooleanProperty warnAboutDuplicatesOnImport;
@@ -26,14 +26,18 @@ public class ImporterPreferences {
     private final ObservableSet<CustomImporter> customImporters;
     private final BooleanProperty persistCustomKeys;
     private final ObservableList<String> catalogs;
+    private final ObjectProperty<OnlinePlainCitationParser> defaultOnlinePlainCitationParser;
+
     public ImporterPreferences(boolean importerEnabled,
                                boolean generateNewKeyOnImport,
                                Path importWorkingDirectory,
                                boolean warnAboutDuplicatesOnImport,
                                Set<CustomImporter> customImporters,
                                Set<FetcherApiKey> apiKeys,
                                boolean persistCustomKeys,
-                               List<String> catalogs) {
+                               List<String> catalogs,
+                               OnlinePlainCitationParser defaultOnlinePlainCitationParser
+    ) {
         this.importerEnabled = new SimpleBooleanProperty(importerEnabled);
         this.generateNewKeyOnImport = new SimpleBooleanProperty(generateNewKeyOnImport);
         this.importWorkingDirectory = new SimpleObjectProperty<>(importWorkingDirectory);
@@ -42,6 +46,7 @@ public ImporterPreferences(boolean importerEnabled,
         this.apiKeys = FXCollections.observableSet(apiKeys);
         this.persistCustomKeys = new SimpleBooleanProperty(persistCustomKeys);
         this.catalogs = FXCollections.observableArrayList(catalogs);
+        this.defaultOnlinePlainCitationParser = new SimpleObjectProperty<>(defaultOnlinePlainCitationParser);
     }
 
     public boolean areImporterEnabled() {
@@ -133,4 +138,16 @@ public void setCatalogs(List<String> catalogs) {
     public ObservableList<String> getCatalogs() {
           return catalogs;
     }
+
+    public OnlinePlainCitationParser getDefaultOnlinePlainCitationParser() {
+        return defaultOnlinePlainCitationParser.get();
+    }
+
+    public ObjectProperty<OnlinePlainCitationParser> defaultOnlinePlainCitationParserProperty() {
+        return defaultOnlinePlainCitationParser;
+    }
+
+    public void setDefaultOnlinePlainCitationParser(OnlinePlainCitationParser defaultOnlinePlainCitationParser) {
+        this.defaultOnlinePlainCitationParser.set(defaultOnlinePlainCitationParser);
+    }
 }