Skip to content

Commit

Permalink
Merge branch 'bugfix/fix-dropped-misclassified-text' into bugfix/fix-…
Browse files Browse the repository at this point in the history 
…header-misclassification-dl
  • Loading branch information
@lfoppiano
lfoppiano committed Dec 22, 2024
2 parents 1fa792c + 09c824d commit 77cf7a2
Showing 1 changed file with 91 additions and 51 deletions.
142 changes: 91 additions & 51 deletions grobid-core/src/main/java/org/grobid/core/engines/FullTextParser.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -249,25 +249,25 @@ else if (config.getConsolidateCitations() == 2)

// full text processing
featSeg = getBodyTextFeatured(doc, documentBodyParts);
String resultBody = null;
LayoutTokenization layoutTokenization = null;
String bodyResults = null;
LayoutTokenization bodyLayoutTokens = null;
List<Figure> figures = null;
List<Table> tables = null;
List<Equation> equations = null;
if (featSeg != null && isNotBlank(featSeg.getLeft())) {
// if featSeg is null, it usually means that the fulltext body is not found in the
// document segmentation
String bodytext = featSeg.getLeft();
layoutTokenization = featSeg.getRight();
String bodyText = featSeg.getLeft();
bodyLayoutTokens = featSeg.getRight();
//tokenizationsBody = featSeg.getB().getTokenization();
//layoutTokensBody = featSeg.getB().getLayoutTokens();

resultBody = label(bodytext);
bodyResults = label(bodyText);
//Correct subsequent I-<figure> or I-<table>
resultBody = LabelUtils.postProcessFulltextFixInvalidTableOrFigure(resultBody);
bodyResults = LabelUtils.postProcessFulltextFixInvalidTableOrFigure(bodyResults);

// we apply now the figure and table models based on the fulltext labeled output
figures = processFigures(resultBody, layoutTokenization.getTokenization(), doc);
figures = processFigures(bodyResults, bodyLayoutTokens.getTokenization(), doc);
// further parse the caption
for(Figure figure : figures) {
if (CollectionUtils.isNotEmpty(figure.getCaptionLayoutTokens()) ) {
Expand All @@ -277,18 +277,27 @@ else if (config.getConsolidateCitations() == 2)
}
}

long numberFiguresFulltextModel = Arrays.stream(bodyResults.split("\n"))
.filter(r -> r.endsWith("I-" + TaggingLabels.FIGURE_LABEL))
.count();

List<Figure> badFigures = figures.stream()
.filter(f -> !f.isCompleteForTEI())
.collect(Collectors.toList());

LOGGER.warn("Identified bad figures: " + badFigures.size());
resultBody = revertResultsForBadItems(badFigures, resultBody, TaggingLabels.FIGURE_LABEL);
LOGGER.info("Number of figures badly formatted or incomplete we identified: " + badFigures.size());
bodyResults = revertResultsForBadItems(badFigures, bodyResults, TaggingLabels.FIGURE_LABEL,
!(figures.size() > numberFiguresFulltextModel));

figures = figures.stream()
.filter(f -> !badFigures.contains(f))
.collect(Collectors.toList());

tables = processTables(resultBody, layoutTokenization.getTokenization(), doc);
tables = processTables(bodyResults, bodyLayoutTokens.getTokenization(), doc);

long numberTablesFulltextModel = Arrays.stream(bodyResults.split("\n"))
.filter(r -> r.endsWith("I-" + TaggingLabels.TABLE_LABEL))
.count();

//We deal with tables considered bad by reverting them as <paragraph>, to reduce the risk them to be
// dropped later on.
Expand All @@ -299,8 +308,9 @@ else if (config.getConsolidateCitations() == 2)
.filter(t -> !(t.isCompleteForTEI() && t.validateTable()))
.collect(Collectors.toList());

LOGGER.warn("Identified bad tables: " + badTables.size());
resultBody = revertResultsForBadItems(badTables, resultBody, TaggingLabels.TABLE_LABEL);
LOGGER.info("Number of tables badly formatted or incomplete we identified: " + badTables.size());
bodyResults = revertResultsForBadItems(badTables, bodyResults, TaggingLabels.TABLE_LABEL,
!(tables.size() > numberTablesFulltextModel));

tables = tables.stream()
.filter(t-> !badTables.contains(t))
Expand All @@ -320,38 +330,43 @@ else if (config.getConsolidateCitations() == 2)
}
}

equations = processEquations(resultBody, layoutTokenization.getTokenization(), doc);
equations = processEquations(bodyResults, bodyLayoutTokens.getTokenization(), doc);
} else {
LOGGER.debug("Fulltext model: The featured body is empty");
}

// possible annexes (view as a piece of full text similar to the body)
documentBodyParts = doc.getDocumentPart(SegmentationLabels.ANNEX);
featSeg = getBodyTextFeatured(doc, documentBodyParts);
String resultAnnex = null;
List<LayoutToken> tokenizationsBody2 = null;
String annexResults = null;
List<LayoutToken> annexTokens = null;
if (featSeg != null && isNotEmpty(trim(featSeg.getLeft()))) {
// if featSeg is null, it usually means that no body segment is found in the
// if featSeg is null, it usually means that no annex segment is found in the
// document segmentation
String bodytext = featSeg.getLeft();
tokenizationsBody2 = featSeg.getRight().getTokenization();
resultAnnex = label(bodytext);
//System.out.println(rese);
String annexFeatures = featSeg.getLeft();
annexTokens = featSeg.getRight().getTokenization();
annexResults = label(annexFeatures);
// System.out.println(annexResults);
}

// post-process reference and footnote callout to keep them consistent (e.g. for example avoid that a footnote
// callout in superscript is by error labeled as a numerical reference callout)
List<MarkerType> markerTypes = null;

if (resultBody != null)
markerTypes = postProcessCallout(resultBody, layoutTokenization);
if (bodyResults != null)
markerTypes = postProcessCallout(bodyResults, bodyLayoutTokens);

// final combination
toTEI(doc, // document
resultBody, resultAnnex, // labeled data for body and annex
layoutTokenization, tokenizationsBody2, // tokenization for body and annex
bodyResults,
annexResults, // labeled data for body and annex
bodyLayoutTokens,
annexTokens, // tokenization for body and annex
resHeader, // header
figures, tables, equations, markerTypes,
figures,
tables,
equations,
markerTypes,
config);
return doc;
} catch (GrobidException e) {
Expand All @@ -362,6 +377,10 @@ else if (config.getConsolidateCitations() == 2)
}

static String revertResultsForBadItems(List<? extends Figure> badFiguresOrTables, String resultBody, String itemLabel) {
return revertResultsForBadItems(badFiguresOrTables, resultBody, itemLabel, true);
}

static String revertResultsForBadItems(List<? extends Figure> badFiguresOrTables, String resultBody, String itemLabel, boolean strict) {
//LF: we update the resultBody sequence by reverting these tables as <paragraph> elements
if (CollectionUtils.isNotEmpty(badFiguresOrTables)) {
List<List<String>> labelledResultsAsList = Arrays.stream(resultBody.split("\n"))
Expand All @@ -371,7 +390,8 @@ static String revertResultsForBadItems(List<? extends Figure> badFiguresOrTables
for (Figure badItem : badFiguresOrTables) {
// Find the index of the first layoutToken of the table in the tokenization
List<LayoutToken> layoutTokenItem = badItem.getLayoutTokens();
List<Integer> candidateIndexes = findCandidateIndex(layoutTokenItem, labelledResultsAsList, itemLabel);
List<Integer> candidateIndexes = findCandidateIndex(layoutTokenItem, labelledResultsAsList,
itemLabel, strict);
if (candidateIndexes.isEmpty()) {
LOGGER.info("Cannot find the candidate index for fixing the tables.");
continue;
Expand Down Expand Up @@ -441,20 +461,38 @@ static int consolidateResultCandidateThroughSequence(List<Integer> candidateInde
return resultIndexCandidate;
}

/**
* Find a set of candidates representing the indexes from the labelledResults which could correspond
* to the first token of the figure/table
*
* strict = True then it will check the items related to I-<table> or I-<figure> first
* and then the <table> or <figure> only if there are not candidates
* strict = False is usually necessary if there are more tables than I- token, this because a figure/table could be
* identified within the sequence initially provided by the fulltext model
*
*/
@NotNull
static List<Integer> findCandidateIndex(List<LayoutToken> layoutTokenItem, List<List<String>> labelledResultsAsList, String itemLabel) {
return findCandidateIndex(layoutTokenItem, labelledResultsAsList, itemLabel, true);
}

@NotNull
static List<Integer> findCandidateIndex(List<LayoutToken> layoutTokenItem, List<List<String>> labelledResultsAsList, String itemLabel, boolean strict) {
LayoutToken firstLayoutTokenItem = layoutTokenItem.get(0);

List<Integer> candidateIndexes = IntStream.range(0, labelledResultsAsList.size())
.filter(i -> labelledResultsAsList.get(i).get(0).equals(firstLayoutTokenItem.getText())
&& Iterables.getLast(labelledResultsAsList.get(i)).equals("I-"+ itemLabel))
&& Iterables.getLast(labelledResultsAsList.get(i)).equals("I-" + itemLabel))
.boxed()
.collect(Collectors.toList());

if (candidateIndexes.isEmpty()) {
if (candidateIndexes.isEmpty() || !strict) {
candidateIndexes = IntStream.range(0, labelledResultsAsList.size())
.filter(i -> labelledResultsAsList.get(i).get(0).equals(firstLayoutTokenItem.getText())
&& Iterables.getLast(labelledResultsAsList.get(i)).equals(itemLabel))
&& (
Iterables.getLast(labelledResultsAsList.get(i)).equals(itemLabel)
|| Iterables.getLast(labelledResultsAsList.get(i)).equals("I-" + itemLabel))
)
.boxed()
.collect(Collectors.toList());
}
Expand Down Expand Up @@ -2241,10 +2279,12 @@ protected Pair<String,String> processTrainingDataFigures(String rese,

// If there still an open figure
if (openFigure) {
while((tokenizationsFigure.size() > 0) &&
while(CollectionUtils.isNotEmpty(tokenizationsFigure) &&
(tokenizationsFigure.get(0).getText().equals("\n") ||
tokenizationsFigure.get(0).getText().equals(" ")) )
tokenizationsFigure.get(0).getText().equals(" "))
) {
tokenizationsFigure.remove(0);
}

// process the "accumulated" figure
Pair<String,String> trainingData = parsers.getFigureParser()
Expand Down Expand Up @@ -2639,8 +2679,8 @@ private static MarkerType getBestType(Map<MarkerType,Integer> markerTypeCount) {
* and body sections.
*/
private void toTEI(Document doc,
String reseBody,
String reseAnnex,
String bodyLabellingResult,
String annexLabellingResult,
LayoutTokenization layoutTokenization,
List<LayoutToken> tokenizationsAnnex,
BiblioItem resHeader,
Expand Down Expand Up @@ -2670,9 +2710,9 @@ private void toTEI(Document doc,
parsers.getFundingAcknowledgementParser().processingXmlFragment(acknowledgmentStmt.toString(), config);

if (localResult != null && localResult.getLeft() != null) {
String local_tei = localResult.getLeft().toXML();
local_tei = local_tei.replace(" xmlns=\"http://www.tei-c.org/ns/1.0\"", "");
annexStatements.add(local_tei);
String localTei = localResult.getLeft().toXML();
localTei = localTei.replace(" xmlns=\"http://www.tei-c.org/ns/1.0\"", "");
annexStatements.add(localTei);
}
else {
annexStatements.add(acknowledgmentStmt.toString());
Expand All @@ -2681,14 +2721,14 @@ private void toTEI(Document doc,
if (localResult != null && localResult.getRight() != null) {
if (localResult.getRight().getLeft() != null) {
List<Funding> localFundings = localResult.getRight().getLeft();
if (localFundings.size()>0) {
if (CollectionUtils.isNotEmpty(localFundings)) {
fundings.addAll(localFundings);
}
}

if (localResult.getRight().getRight() != null) {
List<Affiliation> localAffiliations = localResult.getRight().getRight();
if (localAffiliations.size()>0) {
if (CollectionUtils.isNotEmpty(localAffiliations)) {
affiliations.addAll(localAffiliations);
}
}
Expand All @@ -2709,29 +2749,29 @@ private void toTEI(Document doc,
resCitations,
config);
}
if (fundingStmt.length() > 0) {
if (StringUtils.isNotBlank(fundingStmt)) {
MutablePair<Element, MutableTriple<List<Funding>,List<Person>,List<Affiliation>>> localResult =
parsers.getFundingAcknowledgementParser().processingXmlFragment(fundingStmt.toString(), config);

if (localResult != null && localResult.getLeft() != null) {
String local_tei = localResult.getLeft().toXML();
local_tei = local_tei.replace(" xmlns=\"http://www.tei-c.org/ns/1.0\"", "");
annexStatements.add(local_tei);
String localTEI = localResult.getLeft().toXML();
localTEI = localTEI.replace(" xmlns=\"http://www.tei-c.org/ns/1.0\"", "");
annexStatements.add(localTEI);
} else {
annexStatements.add(fundingStmt.toString());
}

if (localResult != null && localResult.getRight() != null) {
if (localResult.getRight().getLeft() != null) {
List<Funding> localFundings = localResult.getRight().getLeft();
if (localFundings.size()>0) {
if (CollectionUtils.isNotEmpty(localFundings)) {
fundings.addAll(localFundings);
}
}

if (localResult.getRight().getRight() != null) {
List<Affiliation> localAffiliations = localResult.getRight().getRight();
if (localAffiliations.size()>0) {
if (CollectionUtils.isNotEmpty(localAffiliations)) {
affiliations.addAll(localAffiliations);
}
}
Expand All @@ -2752,24 +2792,24 @@ private void toTEI(Document doc,
parsers.getFundingAcknowledgementParser().processingXmlFragment(fundingStmt.toString(), config);

if (localResult != null && localResult.getLeft() != null){
String local_tei = localResult.getLeft().toXML();
local_tei = local_tei.replace(" xmlns=\"http://www.tei-c.org/ns/1.0\"", "");
annexStatements.add(local_tei);
String localTEI = localResult.getLeft().toXML();
localTEI = localTEI.replace(" xmlns=\"http://www.tei-c.org/ns/1.0\"", "");
annexStatements.add(localTEI);
} else {
annexStatements.add(fundingStmt.toString());
}

if (localResult != null && localResult.getRight() != null) {
if (localResult.getRight().getLeft() != null) {
List<Funding> localFundings = localResult.getRight().getLeft();
if (localFundings.size()>0) {
if (CollectionUtils.isNotEmpty(localFundings)) {
fundings.addAll(localFundings);
}
}

if (localResult.getRight().getRight() != null) {
List<Affiliation> localAffiliations = localResult.getRight().getRight();
if (localAffiliations.size()>0) {
if (CollectionUtils.isNotEmpty(localAffiliations)) {
affiliations.addAll(localAffiliations);
}
}
Expand All @@ -2778,7 +2818,7 @@ private void toTEI(Document doc,

tei.append(teiFormatter.toTEIHeader(resHeader, null, resCitations, markerTypes, fundings, config));

tei = teiFormatter.toTEIBody(tei, reseBody, resHeader, resCitations,
tei = teiFormatter.toTEIBody(tei, bodyLabellingResult, resHeader, resCitations,
layoutTokenization, figures, tables, equations, markerTypes, doc, config);

tei.append("\t\t<back>\n");
Expand Down Expand Up @@ -2873,7 +2913,7 @@ else if (affiliation.getAffiliationString() != null && affiliation.getAffiliatio
tei.append(availabilityStmt.toString());
}

tei = teiFormatter.toTEIAnnex(tei, reseAnnex, resHeader, resCitations,
tei = teiFormatter.toTEIAnnex(tei, annexLabellingResult, resHeader, resCitations,
tokenizationsAnnex, markerTypes, doc, config);

tei = teiFormatter.toTEIReferences(tei, resCitations, config);
Expand Down

0 comments on commit 77cf7a2

Please sign in to comment.