javier/Stirling-PDF
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

formatting

Browse Source
This commit is contained in:
Anthony Stirling
2023-12-30 19:11:27 +00:00
parent 7b43fca6fc
commit 5f771b7851
155 changed files with 5539 additions and 4767 deletions
Download Patch File Download Diff File Expand all files Collapse all files

131
src/main/java/stirling/software/SPDF/pdf/ImageFinder.java
View File

@@ -48,83 +48,84 @@ public class ImageFinder extends org.apache.pdfbox.contentstream.PDFGraphicsStre
super.processOperator(operator, operands);
}
@Override
public void appendRectangle(Point2D p0, Point2D p1, Point2D p2, Point2D p3) throws IOException {
// TODO Auto-generated method stub
}
@Override
public void appendRectangle(Point2D p0, Point2D p1, Point2D p2, Point2D p3) throws IOException {
// TODO Auto-generated method stub
@Override
public void drawImage(PDImage pdImage) throws IOException {
// TODO Auto-generated method stub
}
}
@Override
public void clip(int windingRule) throws IOException {
// TODO Auto-generated method stub
}
@Override
public void drawImage(PDImage pdImage) throws IOException {
// TODO Auto-generated method stub
@Override
public void moveTo(float x, float y) throws IOException {
// TODO Auto-generated method stub
}
}
@Override
public void lineTo(float x, float y) throws IOException {
// TODO Auto-generated method stub
}
@Override
public void clip(int windingRule) throws IOException {
// TODO Auto-generated method stub
@Override
public void curveTo(float x1, float y1, float x2, float y2, float x3, float y3) throws IOException {
// TODO Auto-generated method stub
}
}
@Override
public Point2D getCurrentPoint() throws IOException {
// TODO Auto-generated method stub
return null;
}
@Override
public void moveTo(float x, float y) throws IOException {
// TODO Auto-generated method stub
@Override
public void closePath() throws IOException {
// TODO Auto-generated method stub
}
}
@Override
public void endPath() throws IOException {
// TODO Auto-generated method stub
}
@Override
public void lineTo(float x, float y) throws IOException {
// TODO Auto-generated method stub
@Override
public void strokePath() throws IOException {
// TODO Auto-generated method stub
}
}
@Override
public void fillPath(int windingRule) throws IOException {
// TODO Auto-generated method stub
}
@Override
public void curveTo(float x1, float y1, float x2, float y2, float x3, float y3)
throws IOException {
// TODO Auto-generated method stub
@Override
public void fillAndStrokePath(int windingRule) throws IOException {
// TODO Auto-generated method stub
}
}
@Override
public void shadingFill(COSName shadingName) throws IOException {
// TODO Auto-generated method stub
}
@Override
public Point2D getCurrentPoint() throws IOException {
// TODO Auto-generated method stub
return null;
}
@Override
public void closePath() throws IOException {
// TODO Auto-generated method stub
}
@Override
public void endPath() throws IOException {
// TODO Auto-generated method stub
}
@Override
public void strokePath() throws IOException {
// TODO Auto-generated method stub
}
@Override
public void fillPath(int windingRule) throws IOException {
// TODO Auto-generated method stub
}
@Override
public void fillAndStrokePath(int windingRule) throws IOException {
// TODO Auto-generated method stub
}
@Override
public void shadingFill(COSName shadingName) throws IOException {
// TODO Auto-generated method stub
}
// ... rest of the overridden methods
}

135
src/main/java/stirling/software/SPDF/pdf/TextFinder.java
View File

@@ -1,4 +1,5 @@
package stirling.software.SPDF.pdf;
import java.io.IOException;
import java.util.ArrayList;
import java.util.List;
@@ -13,78 +14,80 @@ import stirling.software.SPDF.model.PDFText;
public class TextFinder extends PDFTextStripper {
private final String searchText;
private final boolean useRegex;
private final boolean wholeWordSearch;
private final List<PDFText> textOccurrences = new ArrayList<>();
private final String searchText;
private final boolean useRegex;
private final boolean wholeWordSearch;
private final List<PDFText> textOccurrences = new ArrayList<>();
public TextFinder(String searchText, boolean useRegex, boolean wholeWordSearch) throws IOException {
this.searchText = searchText.toLowerCase();
this.useRegex = useRegex;
this.wholeWordSearch = wholeWordSearch;
setSortByPosition(true);
}
public TextFinder(String searchText, boolean useRegex, boolean wholeWordSearch)
throws IOException {
this.searchText = searchText.toLowerCase();
this.useRegex = useRegex;
this.wholeWordSearch = wholeWordSearch;
setSortByPosition(true);
}
private List<Integer> findOccurrencesInText(String searchText, String content) {
List<Integer> indexes = new ArrayList<>();
Pattern pattern;
private List<Integer> findOccurrencesInText(String searchText, String content) {
List<Integer> indexes = new ArrayList<>();
Pattern pattern;
if (useRegex) {
// Use regex-based search
pattern = wholeWordSearch
? Pattern.compile("(\\b|_|\\.)" + searchText + "(\\b|_|\\.)")
: Pattern.compile(searchText);
} else {
// Use normal text search
pattern = wholeWordSearch
? Pattern.compile("(\\b|_|\\.)" + Pattern.quote(searchText) + "(\\b|_|\\.)")
: Pattern.compile(Pattern.quote(searchText));
}
if (useRegex) {
// Use regex-based search
pattern =
wholeWordSearch
? Pattern.compile("(\\b|_|\\.)" + searchText + "(\\b|_|\\.)")
: Pattern.compile(searchText);
} else {
// Use normal text search
pattern =
wholeWordSearch
? Pattern.compile(
"(\\b|_|\\.)" + Pattern.quote(searchText) + "(\\b|_|\\.)")
: Pattern.compile(Pattern.quote(searchText));
}
Matcher matcher = pattern.matcher(content);
while (matcher.find()) {
indexes.add(matcher.start());
}
return indexes;
}
@Override
protected void writeString(String text, List<TextPosition> textPositions) {
for (Integer index : findOccurrencesInText(searchText, text.toLowerCase())) {
if (index + searchText.length() <= textPositions.size()) {
// Initial values based on the first character
TextPosition first = textPositions.get(index);
float minX = first.getX();
float minY = first.getY();
float maxX = first.getX() + first.getWidth();
float maxY = first.getY() + first.getHeight();
Matcher matcher = pattern.matcher(content);
while (matcher.find()) {
indexes.add(matcher.start());
}
return indexes;
}
// Loop over the rest of the characters and adjust bounding box values
for (int i = index; i < index + searchText.length(); i++) {
TextPosition position = textPositions.get(i);
minX = Math.min(minX, position.getX());
minY = Math.min(minY, position.getY());
maxX = Math.max(maxX, position.getX() + position.getWidth());
maxY = Math.max(maxY, position.getY() + position.getHeight());
}
@Override
protected void writeString(String text, List<TextPosition> textPositions) {
for (Integer index : findOccurrencesInText(searchText, text.toLowerCase())) {
if (index + searchText.length() <= textPositions.size()) {
// Initial values based on the first character
TextPosition first = textPositions.get(index);
float minX = first.getX();
float minY = first.getY();
float maxX = first.getX() + first.getWidth();
float maxY = first.getY() + first.getHeight();
textOccurrences.add(new PDFText(
getCurrentPageNo() - 1,
minX,
minY,
maxX,
maxY,
text
));
}
}
}
// Loop over the rest of the characters and adjust bounding box values
for (int i = index; i < index + searchText.length(); i++) {
TextPosition position = textPositions.get(i);
minX = Math.min(minX, position.getX());
minY = Math.min(minY, position.getY());
maxX = Math.max(maxX, position.getX() + position.getWidth());
maxY = Math.max(maxY, position.getY() + position.getHeight());
}
public List<PDFText> getTextLocations(PDDocument document) throws Exception {
this.getText(document);
System.out.println("Found " + textOccurrences.size() + " occurrences of '" + searchText + "' in the document.");
textOccurrences.add(
new PDFText(getCurrentPageNo() - 1, minX, minY, maxX, maxY, text));
}
}
}
return textOccurrences;
}
public List<PDFText> getTextLocations(PDDocument document) throws Exception {
this.getText(document);
System.out.println(
"Found "
+ textOccurrences.size()
+ " occurrences of '"
+ searchText
+ "' in the document.");
}
return textOccurrences;
}
}