extern/pdf-to-markdown
1
0
Fork 0
mirror of https://github.com/jzillmann/pdf-to-markdown.git synced 2025-06-24 19:41:24 +02:00
Code Issues Packages Projects Releases Wiki Activity

No unused locals

Browse Source
This commit is contained in:
Johannes Zillmann 2021-03-21 08:39:42 +01:00
parent 17290cf746
commit f5a180113d
5 changed files with 6 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
core/src/transformer/RemoveRepetitiveItems.ts
View File

@ -20,8 +20,8 @@ const config = {
// Max number of lines at top/bottom (per page) which are getting evaluated for eviction // Max number of lines at top/bottom (per page) which are getting evaluated for eviction
maxNumberOffTopOrBottomLines: 3, maxNumberOffTopOrBottomLines: 3,
// From the absolute fringe elements (min/max y) how much y can item deviate before beeing disregarded. // From the absolute fringe elements (min/max y) how much y can item deviate before beeing disregarded.
maxDistanceFromFringeElements:30, maxDistanceFromFringeElements: 30,
// Max neighbour taken (in one direction) for detecting neighbour similarity. // Max neighbour taken (in one direction) for detecting neighbour similarity.
// Choosen number might be more effectful for PDFs with a strong odd/evan page differernce. // Choosen number might be more effectful for PDFs with a strong odd/evan page differernce.
@ -54,7 +54,7 @@ export default class RemoveRepetitiveItems extends ItemTransformer {
const yLines = pageExtracts const yLines = pageExtracts
.map((page) => page.lineByY(y)) .map((page) => page.lineByY(y))
.filter((line) => typeof line !== 'undefined') as Line[]; .filter((line) => typeof line !== 'undefined') as Line[];
const texts = yLines.map((line) => line.text());
const similarities = flatMap(yLines, (line, idx) => const similarities = flatMap(yLines, (line, idx) =>
adiacentLines(yLines, idx).map((adiacentLine) => calculateSimilarity(line, adiacentLine)), adiacentLines(yLines, idx).map((adiacentLine) => calculateSimilarity(line, adiacentLine)),
); );
@ -66,6 +66,7 @@ export default class RemoveRepetitiveItems extends ItemTransformer {
// - contain chapter highlights // - contain chapter highlights
// - contains rising number // - contains rising number
// const texts = yLines.map((line) => line.text());
// console.log('y' + y, texts, similarities, median(similarities)); // console.log('y' + y, texts, similarities, median(similarities));
return median(similarities) >= config.minSimilarity; return median(similarities) >= config.minSimilarity;
}); });

1
core/test/Files.test.ts
View File

@ -2,7 +2,6 @@ import { toMatchFile } from 'jest-file-snapshot';
import * as pdfjs from 'pdfjs-dist/es5/build/pdf'; import * as pdfjs from 'pdfjs-dist/es5/build/pdf';
import * as fs from 'fs'; import * as fs from 'fs';
import * as path from 'path';
import PdfParser from 'src/PdfParser'; import PdfParser from 'src/PdfParser';
import PdfPipeline from 'src/PdfPipeline'; import PdfPipeline from 'src/PdfPipeline';

1
core/test/Progress.test.ts
View File

@ -92,7 +92,6 @@ function expectTotalProgress(progress: Progress, expected: number) {
function expectStageInProgress(progress: Progress, stageIndex: number) { function expectStageInProgress(progress: Progress, stageIndex: number) {
for (let index = 0; index < progress.stageProgress.length; index++) { for (let index = 0; index < progress.stageProgress.length; index++) {
const stageProgress = progress.stageProgress[index];
if (index < stageIndex) { if (index < stageIndex) {
expect(progress.isProgressing(index)).toBe(false); expect(progress.isProgressing(index)).toBe(false);
expect(progress.isComplete(index)).toBe(true); expect(progress.isComplete(index)).toBe(true);

3
core/test/debug/detectChanges.test.ts
View File

@ -1,8 +1,7 @@
import ChangeTracker from 'src/debug/ChangeTracker'; import ChangeTracker from 'src/debug/ChangeTracker';
import { detectChanges } from 'src/debug/detectChanges'; import { detectChanges } from 'src/debug/detectChanges';
import { PositionChange, Direction, Addition, Removal } from 'src/debug/ChangeIndex'; import { PositionChange, Direction, Addition, Removal } from 'src/debug/ChangeIndex';
import { idItem, idItems, items } from 'test/testItems'; import { idItem, idItems } from 'test/testItems';
import Item from 'src/Item';
test('No changes', async () => { test('No changes', async () => {
const items = idItems(0, ['A', 'B', 'C']); const items = idItems(0, ['A', 'B', 'C']);

1
core/tsconfig.json
View File

@ -6,6 +6,7 @@
"outDir": "./lib", "outDir": "./lib",
"strict": true, "strict": true,
"noImplicitAny": false, "noImplicitAny": false,
"noUnusedLocals": true,
"baseUrl": "./", "baseUrl": "./",
"paths": { "paths": {
"src/*": ["src/*"], "src/*": ["src/*"],