danvics/openreader
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Improve space detection logic in PDF text extraction

Browse source
This commit is contained in:
Richard Roberson 2025-03-22 12:10:17 -06:00
parent 4ed6b17e98
commit 1bbd77f93b
1 changed files with 15 additions and 3 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

18
src/utils/pdf.ts
View file

@ -142,10 +142,22 @@ export async function extractTextFromPDF(
const currentStartX = item.transform[4]; const currentStartX = item.transform[4];
const space = currentStartX - prevEndX; const space = currentStartX - prevEndX;
if (space > ((item.width ?? 0) * 0.3)) { // Get average character width as fallback
lineText += ' ' + item.str; const avgCharWidth = (item.width ?? 0) / Math.max(1, item.str.length);
// Multiple conditions for space detection
const needsSpace =
// Primary check: significant gap between items
space > Math.max(avgCharWidth * 0.3, 2) ||
// Secondary check: natural word boundary
(!/^\W/.test(item.str) && !/\W$/.test(prevItem.str)) ||
// Tertiary check: items are far enough apart relative to their size
(space > ((prevItem.width ?? 0) * 0.25));
if (needsSpace) {
lineText += ' ' + item.str;
} else { } else {
lineText += item.str; lineText += item.str;
} }
} }
prevItem = item; prevItem = item;