Marmelator/layout-parser
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases 289 Wiki Activity

RED-7806 - Specific customer document cannot be processed

Browse Source 
- check for font name null before using to avoid the NPE
This commit is contained in:
Corina Olariu 2023-10-25 09:16:47 +03:00
parent 9abdc6d44d
commit efa3d75479
3 changed files with 13 additions and 10 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

14
layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/LayoutParsingPipeline.java
View File

@ -143,13 +143,13 @@ public class LayoutParsingPipeline {
return String.format("%d pages with %d sections, %d headlines, %d paragraphs, %d tables with %d cells, %d headers, and %d footers parsed", return String.format("%d pages with %d sections, %d headlines, %d paragraphs, %d tables with %d cells, %d headers, and %d footers parsed",
numberOfPages, numberOfPages,
semanticNodeCounts.get(NodeType.SECTION), semanticNodeCounts.get(NodeType.SECTION) == null ? 0 : semanticNodeCounts.get(NodeType.SECTION),
semanticNodeCounts.get(NodeType.HEADLINE), semanticNodeCounts.get(NodeType.HEADLINE) == null ? 0 : semanticNodeCounts.get(NodeType.HEADLINE),
semanticNodeCounts.get(NodeType.PARAGRAPH), semanticNodeCounts.get(NodeType.PARAGRAPH) == null ? 0 : semanticNodeCounts.get(NodeType.PARAGRAPH),
semanticNodeCounts.get(NodeType.TABLE), semanticNodeCounts.get(NodeType.TABLE) == null ? 0 : semanticNodeCounts.get(NodeType.TABLE),
semanticNodeCounts.get(NodeType.TABLE_CELL), semanticNodeCounts.get(NodeType.TABLE_CELL) == null ? 0 : semanticNodeCounts.get(NodeType.TABLE_CELL),
semanticNodeCounts.get(NodeType.HEADER), semanticNodeCounts.get(NodeType.HEADER) == null ? 0 : semanticNodeCounts.get(NodeType.HEADER),
semanticNodeCounts.get(NodeType.FOOTER)); semanticNodeCounts.get(NodeType.FOOTER) == null ? 0 : semanticNodeCounts.get(NodeType.FOOTER));
} }

7
layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/model/text/TextPositionSequence.java
View File

@ -224,6 +224,8 @@ public class TextPositionSequence implements CharSequence {
@JsonAttribute(ignore = true) @JsonAttribute(ignore = true)
public String getFont() { public String getFont() {
if (textPositions.get(0).getFontName() == null)
return "none";
return textPositions.get(0).getFontName().toLowerCase(Locale.ROOT).replaceAll(",bold", "").replaceAll(",italic", ""); return textPositions.get(0).getFontName().toLowerCase(Locale.ROOT).replaceAll(",bold", "").replaceAll(",italic", "");
} }
@ -231,7 +233,9 @@ public class TextPositionSequence implements CharSequence {
@JsonIgnore @JsonIgnore
@JsonAttribute(ignore = true) @JsonAttribute(ignore = true)
public String getFontStyle() { public String getFontStyle() {
if (textPositions.get(0).getFontName() == null) {
return "standard";
}
String lowercaseFontName = textPositions.get(0).getFontName().toLowerCase(Locale.ROOT); String lowercaseFontName = textPositions.get(0).getFontName().toLowerCase(Locale.ROOT);
if (lowercaseFontName.contains("bold") && lowercaseFontName.contains("italic")) { if (lowercaseFontName.contains("bold") && lowercaseFontName.contains("italic")) {
@ -243,7 +247,6 @@ public class TextPositionSequence implements CharSequence {
} else { } else {
return "standard"; return "standard";
} }
} }

2
layoutparser-service/layoutparser-service-processor/src/main/java/com/knecon/fforesight/service/layoutparser/processor/services/factory/SearchTextWithTextPositionFactory.java
View File

@ -133,7 +133,7 @@ public class SearchTextWithTextPositionFactory {
private static void addTextPositionWithFontType(RedTextPosition currentTextPosition, String fontType, List<Integer> fontTypePositions, int stringIdx) { private static void addTextPositionWithFontType(RedTextPosition currentTextPosition, String fontType, List<Integer> fontTypePositions, int stringIdx) {
if (currentTextPosition.getFontName().toLowerCase(Locale.ROOT).contains(fontType)) { if (currentTextPosition.getFontName() != null && currentTextPosition.getFontName().toLowerCase(Locale.ROOT).contains(fontType)) {
fontTypePositions.add(stringIdx); fontTypePositions.add(stringIdx);
} }
} }