Examples with Annotation annis.model.Annotation used on opensource projects

Example 1 with Annotation

use of annis.model.Annotation in project ANNIS by korpling.

the class HTMLVis method createHTML.

public String createHTML(SDocumentGraph graph, VisualizationDefinition[] definitions) {
    HashMap<VisualizationDefinition, Integer> instruction_priorities = new HashMap<>();
    SortedMap<Long, List<OutputItem>> outputStartTags = new TreeMap<>();
    SortedMap<Long, List<OutputItem>> outputEndTags = new TreeMap<>();
    StringBuilder sb = new StringBuilder();
    List<SToken> token = graph.getSortedTokenByText();
    // Get metadata for visualizer if stylesheet requires it
    // First check the stylesheet
    Boolean bolMetaTypeFound = false;
    HashMap<String, String> meta = new HashMap<>();
    int def_priority = 0;
    for (VisualizationDefinition vis : definitions) {
        if (vis.getOutputter().getType() == SpanHTMLOutputter.Type.META_NAME) {
            bolMetaTypeFound = true;
        } else // not a meta-annotation, remember order in config file to set priority
        {
            if (vis.getMatcher() instanceof AnnotationNameMatcher) {
                instruction_priorities.put(vis, def_priority);
            } else if (vis.getMatcher() instanceof AnnotationNameAndValueMatcher) {
                instruction_priorities.put(vis, def_priority);
            } else if (vis.getMatcher() instanceof TokenMatcher) {
                instruction_priorities.put(vis, def_priority);
            }
            def_priority--;
        }
        vis.getOutputter().setMeta(meta);
    }
    if (// Metadata is required, get corpus and document name
    bolMetaTypeFound == true) {
        // Get corpus and document name
        String strDocName = "";
        String strCorpName = "";
        strDocName = graph.getDocument().getName();
        List<String> corpusPath = CommonHelper.getCorpusPath(graph.getDocument().getGraph(), graph.getDocument());
        strCorpName = corpusPath.get(corpusPath.size() - 1);
        // Get metadata and put in hashmap
        List<Annotation> metaData = Helper.getMetaDataDoc(strCorpName, strDocName);
        for (Annotation metaDatum : metaData) {
            meta.put(metaDatum.getName(), metaDatum.getValue());
        }
    }
    for (SToken t : token) {
        tokenColor = "";
        if (mc.containsKey(t) && hitMark) {
            tokenColor = MatchedNodeColors.getHTMLColorByMatch(mc.get(t));
        }
        for (VisualizationDefinition vis : definitions) {
            String matched = vis.getMatcher().matchedAnnotation(t);
            if (matched != null) {
                vis.getOutputter().outputHTML(t, matched, outputStartTags, outputEndTags, tokenColor, Objects.firstNonNull(instruction_priorities.get(vis), 0));
            }
        }
    }
    List<SSpan> spans = graph.getSpans();
    for (VisualizationDefinition vis : definitions) {
        for (SSpan span : spans) {
            tokenColor = "";
            if (mc.containsKey(span) && hitMark) {
                tokenColor = MatchedNodeColors.getHTMLColorByMatch(mc.get(span));
            }
            String matched = vis.getMatcher().matchedAnnotation(span);
            if (matched != null) {
                vis.getOutputter().outputHTML(span, matched, outputStartTags, outputEndTags, tokenColor, Objects.firstNonNull(instruction_priorities.get(vis), 0));
            }
        }
    }
    int minStartTagPos = outputStartTags.firstKey().intValue();
    int maxEndTagPos = outputEndTags.lastKey().intValue();
    // Find BEGIN and END instructions if available
    for (VisualizationDefinition vis : definitions) {
        if (vis.getMatcher() instanceof PseudoRegionMatcher) {
            PseudoRegionMatcher.PseudoRegion psdRegionType = ((PseudoRegionMatcher) vis.getMatcher()).getPsdRegion();
            int positionStart = 0;
            int positionEnd = 0;
            if (!outputEndTags.isEmpty() && !outputStartTags.isEmpty() && psdRegionType != null) {
                switch(psdRegionType) {
                    case BEGIN:
                        positionStart = positionEnd = Integer.MIN_VALUE;
                        // def_priority is now lower than all normal annotation
                        instruction_priorities.put(vis, def_priority);
                        break;
                    case END:
                        positionStart = positionEnd = Integer.MAX_VALUE;
                        // def_priority is now lower than all normal annotation
                        instruction_priorities.put(vis, def_priority);
                        break;
                    case ALL:
                        // use same position as last and first key
                        positionStart = minStartTagPos;
                        positionEnd = maxEndTagPos;
                        // The ALL pseudo-range must enclose everything, thus it get the
                        // priority which is one lower than the smallest non BEGIN/END
                        // priority.
                        instruction_priorities.put(vis, def_priority);
                        break;
                    default:
                        break;
                }
            }
            switch(vis.getOutputter().getType()) {
                case META_NAME:
                    String strMetaVal = meta.get(vis.getOutputter().getMetaname().trim());
                    if (strMetaVal == null) {
                        throw new NullPointerException("no such metadata name in document: '" + vis.getOutputter().getMetaname().trim() + "'");
                    } else {
                        vis.getOutputter().outputAny(positionStart, positionEnd, ((PseudoRegionMatcher) vis.getMatcher()).getAnnotationName(), strMetaVal, outputStartTags, outputEndTags, Objects.firstNonNull(instruction_priorities.get(vis), 0));
                    }
                    break;
                case CONSTANT:
                    vis.getOutputter().outputAny(positionStart, positionEnd, ((PseudoRegionMatcher) vis.getMatcher()).getAnnotationName(), vis.getOutputter().getConstant(), outputStartTags, outputEndTags, Objects.firstNonNull(instruction_priorities.get(vis), 0));
                    break;
                case EMPTY:
                    vis.getOutputter().outputAny(positionStart, positionEnd, ((PseudoRegionMatcher) vis.getMatcher()).getAnnotationName(), "", outputStartTags, outputEndTags, Objects.firstNonNull(instruction_priorities.get(vis), 0));
                    break;
                case ANNO_NAME:
                    // this shouldn't happen, since the BEGIN/END instruction has no triggering annotation name or value
                    break;
                case VALUE:
                    // this shouldn't happen, since the BEGIN/END instruction has no triggering annotation name or value
                    break;
                case ESCAPED_VALUE:
                    // this shouldn't happen, since the BEGIN/END instruction has no triggering annotation name or value
                    break;
                default:
            }
        }
    }
    // get all used indexes
    Set<Long> indexes = new TreeSet<>();
    indexes.addAll(outputStartTags.keySet());
    indexes.addAll(outputEndTags.keySet());
    for (Long i : indexes) {
        // output all strings belonging to this token position
        // first the start tags for this position
        // add priorities from instruction_priorities for sorting length ties
        List<OutputItem> unsortedStart = outputStartTags.get(i);
        SortedSet<OutputItem> itemsStart = new TreeSet();
        if (unsortedStart != null) {
            Iterator<OutputItem> it = unsortedStart.iterator();
            while (it.hasNext()) {
                OutputItem s = it.next();
                itemsStart.add(s);
            }
        }
        {
            Iterator<OutputItem> it = itemsStart.iterator();
            boolean first = true;
            while (it.hasNext()) {
                OutputItem s = it.next();
                if (!first) {
                    sb.append("-->");
                }
                first = false;
                sb.append(s.getOutputString());
                if (it.hasNext()) {
                    sb.append("<!--\n");
                }
            }
        }
        // then the end tags for this position, but inverse their order
        List<OutputItem> unsortedEnd = outputEndTags.get(i);
        SortedSet<OutputItem> itemsEnd = new TreeSet();
        if (unsortedEnd != null) {
            Iterator<OutputItem> it = unsortedEnd.iterator();
            while (it.hasNext()) {
                OutputItem s = it.next();
                itemsEnd.add(s);
            }
        }
        {
            List<OutputItem> itemsEndReverse = new LinkedList<>(itemsEnd);
            Collections.reverse(itemsEndReverse);
            for (OutputItem s : itemsEndReverse) {
                sb.append(s.getOutputString());
            }
        }
    }
    return sb.toString();
}

Example 2 with Annotation

use of annis.model.Annotation in project ANNIS by korpling.

the class QueryDaoImpl method exportCorpus.

@Override
@Transactional(readOnly = true)
public void exportCorpus(String toplevelCorpus, File outputDirectory) {
    // check if the corpus really exists
    mapCorpusNameToId(toplevelCorpus);
    SaltProject corpusProject = SaltFactory.createSaltProject();
    SCorpusGraph corpusGraph = SaltFactory.createSCorpusGraph();
    corpusGraph.setSaltProject(corpusProject);
    SCorpus rootCorpus = corpusGraph.createCorpus(null, toplevelCorpus);
    // add all root metadata
    for (Annotation metaAnno : listCorpusAnnotations(toplevelCorpus)) {
        rootCorpus.createMetaAnnotation(metaAnno.getNamespace(), metaAnno.getName(), metaAnno.getValue());
    }
    File documentRootDir = new File(outputDirectory, toplevelCorpus);
    if (!outputDirectory.exists()) {
        if (!outputDirectory.mkdirs()) {
            log.warn("Could not create output directory \"{}\" for exporting the corpus", outputDirectory.getAbsolutePath());
        }
    }
    List<Annotation> docs = listDocuments(toplevelCorpus);
    int i = 1;
    for (Annotation docAnno : docs) {
        log.info("Loading document {} from database ({}/{})", docAnno.getName(), i, docs.size());
        SaltProject docProject = retrieveAnnotationGraph(toplevelCorpus, docAnno.getName(), null);
        if (docProject != null && docProject.getCorpusGraphs() != null && !docProject.getCorpusGraphs().isEmpty()) {
            List<Annotation> docMetaData = listCorpusAnnotations(toplevelCorpus, docAnno.getName(), true);
            SCorpusGraph docCorpusGraph = docProject.getCorpusGraphs().get(0);
            // TODO: we could re-use the actual corpus structure instead of just adding a flat list of documents
            if (docCorpusGraph.getDocuments() != null) {
                for (SDocument doc : docCorpusGraph.getDocuments()) {
                    log.info("Removing SFeatures from {} ({}/{})", docAnno.getName(), i, docs.size());
                    // remove all ANNIS specific features that require a special Java class
                    SDocumentGraph graph = doc.getDocumentGraph();
                    if (graph != null) {
                        if (graph.getNodes() != null) {
                            for (SNode n : graph.getNodes()) {
                                n.removeLabel(AnnisConstants.ANNIS_NS, AnnisConstants.FEAT_RELANNIS_NODE);
                            }
                        }
                        if (graph.getRelations() != null) {
                            for (SRelation e : graph.getRelations()) {
                                e.removeLabel(AnnisConstants.ANNIS_NS, AnnisConstants.FEAT_RELANNIS_EDGE);
                            }
                        }
                    }
                    log.info("Saving document {} ({}/{})", doc.getName(), i, docs.size());
                    SaltUtil.saveDocumentGraph(graph, URI.createFileURI(new File(documentRootDir, doc.getName() + "." + SaltUtil.FILE_ENDING_SALT_XML).getAbsolutePath()));
                    SDocument docCopy = corpusGraph.createDocument(rootCorpus, doc.getName());
                    log.info("Adding metadata to document {} ({}/{})", doc.getName(), i, docs.size());
                    for (Annotation metaAnno : docMetaData) {
                        docCopy.createMetaAnnotation(metaAnno.getNamespace(), metaAnno.getName(), metaAnno.getValue());
                    }
                }
            }
        }
        i++;
    }
    // end for each document
    // save the actual SaltProject
    log.info("Saving corpus structure");
    File projectFile = new File(outputDirectory, SaltUtil.FILE_SALT_PROJECT);
    SaltXML10Writer writer = new SaltXML10Writer(projectFile);
    writer.writeSaltProject(corpusProject);
}

Example 3 with Annotation

use of annis.model.Annotation in project ANNIS by korpling.

the class CSVHelper method exportCSVData.

public static void exportCSVData(Iterator<AnnotatedMatch> matches, SortedMap<Integer, SortedSet<String>> columnsByNodePos, PrintWriter w) {
    int count = columnsByNodePos.keySet().size();
    // print values
    while (matches.hasNext()) {
        AnnotatedMatch match = matches.next();
        List<String> line = new ArrayList<>();
        int k = 0;
        for (; k < match.size(); ++k) {
            AnnotatedSpan span = match.get(k);
            Map<String, String> valueByName = new HashMap<>();
            if (span != null) {
                if (span.getAnnotations() != null) {
                    for (Annotation annotation : span.getAnnotations()) {
                        valueByName.put("anno_" + annotation.getQualifiedName(), annotation.getValue());
                    }
                }
                if (span.getMetadata() != null) {
                    for (Annotation meta : span.getMetadata()) {
                        valueByName.put("meta_" + meta.getQualifiedName(), meta.getValue());
                    }
                }
                line.add("" + span.getId());
                line.add(span.getCoveredText().replace("\t", "\\t"));
            }
            for (String name : columnsByNodePos.get(k)) {
                if (valueByName.containsKey(name)) {
                    line.add(valueByName.get(name).replace("\t", "\\t"));
                } else {
                    line.add("'NULL'");
                }
            }
        }
        for (int l = k; l < count; ++l) {
            line.add("'NULL'");
            for (int m = 0; m <= columnsByNodePos.get(l).size(); ++m) {
                line.add("'NULL'");
            }
        }
        w.append(StringUtils.join(line, "\t"));
        w.append("\n");
    }
}

Example 4 with Annotation

use of annis.model.Annotation in project ANNIS by korpling.

the class CSVHelper method exportCSVHeader.

public static SortedMap<Integer, SortedSet<String>> exportCSVHeader(Iterator<AnnotatedMatch> matches, PrintWriter w) {
    // figure out what annotations are used at each match position
    SortedMap<Integer, SortedSet<String>> columnsByNodePos = new TreeMap<>();
    while (matches.hasNext()) {
        AnnotatedMatch match = matches.next();
        for (int j = 0; j < match.size(); ++j) {
            AnnotatedSpan span = match.get(j);
            if (columnsByNodePos.get(j) == null) {
                columnsByNodePos.put(j, new TreeSet<String>());
            }
            if (span != null) {
                for (Annotation annotation : span.getAnnotations()) {
                    columnsByNodePos.get(j).add("anno_" + annotation.getQualifiedName());
                }
                for (Annotation meta : span.getMetadata()) {
                    columnsByNodePos.get(j).add("meta_" + meta.getQualifiedName());
                }
            }
        }
    }
    // important: don't close the wrapper CSVWriter!
    @SuppressWarnings("resource") CSVWriter csvWriter = new CSVWriter(w, '\t', CSVWriter.NO_QUOTE_CHARACTER, '\\');
    // print column names and data types
    int count = columnsByNodePos.keySet().size();
    ArrayList<String> headerLine = new ArrayList<>();
    for (int j = 0; j < count; ++j) {
        headerLine.add(fullColumnName(j + 1, "id"));
        headerLine.add(fullColumnName(j + 1, "span"));
        SortedSet<String> annotationNames = columnsByNodePos.get(j);
        for (String name : annotationNames) {
            headerLine.add(fullColumnName(j + 1, name));
        }
    }
    csvWriter.writeNext(headerLine.toArray(new String[headerLine.size()]));
    return columnsByNodePos;
}

Example 5 with Annotation

use of annis.model.Annotation in project ANNIS by korpling.

the class WekaHelper method exportArffData.

public static void exportArffData(Iterator<AnnotatedMatch> matches, SortedMap<Integer, SortedSet<String>> columnsByNodePos, PrintWriter w) {
    int count = columnsByNodePos.keySet().size();
    w.append("\n@data\n\n");
    // print values
    while (matches.hasNext()) {
        AnnotatedMatch match = matches.next();
        List<String> line = new ArrayList<>();
        int k = 0;
        for (; k < match.size(); ++k) {
            AnnotatedSpan span = match.get(k);
            Map<String, String> valueByName = new HashMap<>();
            if (span != null) {
                if (span.getAnnotations() != null) {
                    for (Annotation annotation : span.getAnnotations()) {
                        valueByName.put("anno_" + annotation.getQualifiedName(), annotation.getValue());
                    }
                }
                if (span.getMetadata() != null) {
                    for (Annotation meta : span.getMetadata()) {
                        valueByName.put("meta_" + meta.getQualifiedName(), meta.getValue());
                    }
                }
                line.add("'" + span.getId() + "'");
                line.add("'" + span.getCoveredText().replace("'", "\\'") + "'");
            }
            for (String name : columnsByNodePos.get(k)) {
                if (valueByName.containsKey(name)) {
                    line.add("'" + valueByName.get(name).replace("'", "\\'") + "'");
                } else {
                    line.add("'NULL'");
                }
            }
        }
        for (int l = k; l < count; ++l) {
            line.add("'NULL'");
            for (int m = 0; m <= columnsByNodePos.get(l).size(); ++m) {
                line.add("'NULL'");
            }
        }
        w.append(StringUtils.join(line, ","));
        w.append("\n");
    }
}