extern/pdf-to-markdown
1
0
Fork 0
mirror of https://github.com/jzillmann/pdf-to-markdown.git synced 2024-11-22 07:43:46 +01:00
Code Issues Packages Projects Releases Wiki Activity

WIP merge successive code blocks

Browse Source
This commit is contained in:
Johannes Zillmann 2017-02-18 12:33:21 +01:00
parent e7ff939351
commit bed3fd357b
2 changed files with 18 additions and 10 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

18
src/javascript/models/transformations/DetectCodeBlocks.jsx
View File

@ -23,7 +23,7 @@ export default class DetectCodeBlocks extends ToPdfBlockViewTransformation {
// TODO ==> combine quotes follow each other // TODO ==> combine quotes follow each other
transform(parseResult:ParseResult) { transform(parseResult:ParseResult) {
const {mostUsedHeight} = parseResult.globals; const {mostUsedHeight, mostUsedDistance} = parseResult.globals;
var foundBlocks = 0; var foundBlocks = 0;
const textCombiner = new TextItemCombiner({}); const textCombiner = new TextItemCombiner({});
@ -49,20 +49,32 @@ export default class DetectCodeBlocks extends ToPdfBlockViewTransformation {
return true; return true;
}; };
const newBlocks = []; const newBlocks = [];
var preceedingCodeBlock;
page.blocks.forEach(block => { page.blocks.forEach(block => {
if (block.type) { if (block.type) {
newBlocks.push(block); newBlocks.push(block);
preceedingCodeBlock = null;
} else { } else {
if (itemAreSuitable(block.textItems)) { if (itemAreSuitable(block.textItems)) {
const mergeWithPreceedingCodeBlock = preceedingCodeBlock && preceedingCodeBlock.textItems[preceedingCodeBlock.textItems.length - 1].y - block.textItems[0].y < mostUsedDistance * 2;
if (mergeWithPreceedingCodeBlock) {
newBlocks.pop();
}
block.annotation = REMOVED_ANNOTATION; block.annotation = REMOVED_ANNOTATION;
newBlocks.push(block); newBlocks.push(block);
newBlocks.push(new PdfBlock({ if (mergeWithPreceedingCodeBlock) {
preceedingCodeBlock.textItems = preceedingCodeBlock.textItems.concat(textCombiner.combine(block.textItems));
} else {
preceedingCodeBlock = new PdfBlock({
type: CODE_BLOCK, type: CODE_BLOCK,
annotation: ADDED_ANNOTATION, annotation: ADDED_ANNOTATION,
textItems: textCombiner.combine(block.textItems) textItems: textCombiner.combine(block.textItems)
})); });
}
newBlocks.push(preceedingCodeBlock);
} else { } else {
newBlocks.push(block); newBlocks.push(block);
preceedingCodeBlock = null;
} }
} }
}); });

4
src/javascript/models/transformations/DetectPdfBlocks.jsx
View File

@ -10,10 +10,6 @@ export default class DetectPdfBlocks extends ToPdfBlockViewTransformation {
super("Detect Blocks"); super("Detect Blocks");
} }
showModificationCheckbox() {
return false;
}
createSummaryView(parseResult:ParseResult) { createSummaryView(parseResult:ParseResult) {
return <div> return <div>
Splitted into Splitted into