extern/pdf-to-markdown
1
0
Fork 0
mirror of https://github.com/jzillmann/pdf-to-markdown.git synced 2025-06-25 03:51:33 +02:00
Code Issues Packages Projects Releases Wiki Activity

Fix duplicate headline bug

Browse Source
This commit is contained in:
Johannes Zillmann 2024-05-17 12:22:34 -06:00
parent 491e8c549a
commit c952409c0e
1 changed files with 5 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
src/transformer/DetectHeaders.ts
View File

@ -37,6 +37,7 @@ export default class DetectHeaders extends ItemTransformer {
const itemsByLine = groupByLine(inputItems); const itemsByLine = groupByLine(inputItems);
const itemToLevel: Map<string, HeadlineType> = new Map(); const itemToLevel: Map<string, HeadlineType> = new Map();
// TODO move the seperate parts to different transformations (easier debuggable/testable)
// Handle title pages: Title pages often have multiple lines of extraordinary height. // Handle title pages: Title pages often have multiple lines of extraordinary height.
// Starting the leveling here would already consume most of the available headline levels. // Starting the leveling here would already consume most of the available headline levels.
@ -145,7 +146,10 @@ export default class DetectHeaders extends ItemTransformer {
items: inputItems.map((item) => { items: inputItems.map((item) => {
const headerType = itemToLevel.get(item.uuid); const headerType = itemToLevel.get(item.uuid);
if (headerType) { if (headerType) {
return itemWithType(item, headerType); const hasAlreadyHeadline = item.data['types'] || [].find((t) => isHeadline(t));
if (!hasAlreadyHeadline) {
return itemWithType(item, headerType);
}
} }
return item; return item;
}), }),