extern/pdf-to-markdown
1
0
Fork 0
mirror of https://github.com/jzillmann/pdf-to-markdown.git synced 2024-11-25 01:03:59 +01:00
Code Issues Packages Projects Releases Wiki Activity

Make everything ready for deploy

Browse Source
This commit is contained in:
Johannes Zillmann 2017-01-09 21:24:12 +01:00
parent 1a97cef440
commit 4ce59ac062
11 changed files with 58 additions and 18 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

33
.eslintrc Normal file
View File

@ -0,0 +1,33 @@
{
// Extend existing configuration
// from ESlint and eslint-plugin-react defaults.
"extends": [
"eslint:recommended", "plugin:react/recommended"
],
"parser": "babel-eslint",
// Enable ES6 support. If you want to use custom Babel
// features, you will need to enable a custom parser
// as described in a section below.
"parserOptions": {
"ecmaVersion": 6,
"sourceType": "module",
"ecmaFeatures": {
"jsx": true
}
},
"env": {
"browser": true,
"node": true,
"es6": true
},
// Enable custom plugin known as eslint-plugin-react
"plugins": [
"react"
],
"rules": {
// Disable `no-console` rule
"no-console": 1,
// Give a warning if identifiers contain underscores
"no-underscore-dangle": 1
}
}

4
package.json
View File

@ -6,7 +6,9 @@
"scripts": {
"watch": "webpack -d --watch",
"build": "webpack",
"lint": "eslint . --ext .js --ext .jsx --cache"
"lint": "eslint src --ext .js --ext .jsx --cache",
"release": "npm run lint && rm -rf build/* && NODE_ENV=production webpack -p",
"deploy": "npm run release && cp -r build/* docs/"
},
"keywords": [
"PDF",

12
src/javascript/components/LoadingView.jsx
View File

@ -1,6 +1,6 @@
import React from 'react';
import pdfjs from 'pdfjs-dist';
import pdfjs from 'pdfjs-dist'; // eslint-disable-line no-unused-vars
import { Line } from 'rc-progress';
import PdfPage from '../models/PdfPage.jsx';
@ -29,7 +29,7 @@ export default class LoadingView extends React.Component {
anouncePageParsed(index, textItems) {
//TODO might make problems.. concat unordered and order at the end ?
this.state.pdfPages[index].textItems = textItems;
this.state.pdfPages[index].textItems = textItems; // eslint-disable-line react/no-direct-mutation-state
this.setState({
parsedPages: this.state.parsedPages + 1
});
@ -39,8 +39,8 @@ export default class LoadingView extends React.Component {
componentWillMount() {
const anounceInitialParseFunction = this.anounceInitialParse.bind(this);
const anouncePageParsedFunction = this.anouncePageParsed.bind(this);
PDFJS.getDocument(this.props.fileBuffer).then(function(pdfDocument) {
console.log('Number of pages: ' + pdfDocument.numPages);
PDFJS.getDocument(this.props.fileBuffer).then(function(pdfDocument) { // eslint-disable-line no-undef
// console.log('Number of pages: ' + pdfDocument.numPages);
// console.debug(pdfDocument);
const numPages = pdfDocument.numPages;
// const numPages = 4; // hack
@ -51,8 +51,8 @@ export default class LoadingView extends React.Component {
}));
}
anounceInitialParseFunction(pdfPages);
for (var i = 1; i <= numPages; i++) {
pdfDocument.getPage(i).then(function(page) {
for (var j = 1; j <= numPages; j++) {
pdfDocument.getPage(j).then(function(page) {
page.getTextContent().then(function(textContent) {
// console.debug(textContent);
const textItems = textContent.items.map(function(item) {

1
src/javascript/components/PdfUploadView.jsx
View File

@ -17,7 +17,6 @@ export default class PdfUploadView extends React.Component {
}
onDrop(files) {
console.debug(files.length);
if (files.length > 1) {
alert(`Maximum one file allowed to upload, but not ${files.length}!`)
return

3
src/javascript/components/PdfView.jsx
View File

@ -1,6 +1,5 @@
import React from 'react';
import Badge from 'react-bootstrap/lib/Badge'
import ButtonToolbar from 'react-bootstrap/lib/ButtonToolbar'
import ButtonGroup from 'react-bootstrap/lib/ButtonGroup'
import Button from 'react-bootstrap/lib/Button'
@ -54,8 +53,6 @@ export default class PdfView extends React.Component {
const {currentTransformation, pageNr} = this.state;
const {pdfPages, transformations} = this.props;
const header = "Parsed " + pdfPages.length + " pages!"
const currentTransformationName = transformations[currentTransformation].name;
const transformedPdfPages = pdfPages.filter((elem, i) => pageNr == -1 || i == pageNr).map(pdfPage => {

7
src/javascript/models/transformations/CombineSameYTransformation.jsx
View File

@ -1,5 +1,6 @@
import Transformation from './Transformation.jsx';
import TextItem from '../TextItem.jsx';
import PdfPage from '../PdfPage.jsx';
export default class CombineSameYTransformation extends Transformation {
@ -18,9 +19,9 @@ export default class CombineSameYTransformation extends Transformation {
if (textItem.y == lastTextItem.y) {
//combine
console.debug("last=" + lastTextItem.text + ", x=" + lastTextItem.x + ", width=" + lastTextItem.width);
console.debug("new=" + textItem.text + ", x=" + textItem.x + ", width=" + textItem.width);
console.debug("diff=" + (textItem.x - lastTextItem.x - lastTextItem.width));
// console.debug("last=" + lastTextItem.text + ", x=" + lastTextItem.x + ", width=" + lastTextItem.width);
// console.debug("new=" + textItem.text + ", x=" + textItem.x + ", width=" + textItem.width);
// console.debug("diff=" + (textItem.x - lastTextItem.x - lastTextItem.width));
var combinedText = lastTextItem.text;
//TODO make 5 dependent on text size or biggest gap?

1
src/javascript/models/transformations/NoOpTransformation.jsx
View File

@ -1,4 +1,5 @@
import Transformation from './Transformation.jsx';
import PdfPage from '../PdfPage.jsx';
export default class NoOpTransformation extends Transformation {

1
src/javascript/models/transformations/RoundCoordinatesTransformation.jsx
View File

@ -1,4 +1,5 @@
import Transformation from './Transformation.jsx';
import PdfPage from '../PdfPage.jsx';
export default class RoundCoordinatesTransformation extends Transformation {

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

@ -1,4 +1,4 @@
import AppState from '../PdfPage.jsx';
import PdfPage from '../PdfPage.jsx';
// A transformation from an PdfPage to an PdfPage
export default class Transformation {
@ -13,7 +13,7 @@ export default class Transformation {
this.name = name;
}
transform(pdfPage:PdfPage) {
transform(pdfPage:PdfPage) { // eslint-disable-line no-unused-vars
throw new TypeError("Do not call abstract method foo from child.");
}
}

4
src/store.js
View File

@ -30,7 +30,7 @@ export default {
var lineY;
// console.debug("Page " + rawPage.index + "-------");
rawPage.textItems.forEach(textItem => {
console.debug(textItem);
// console.debug(textItem);
const yRounded = Math.round(textItem.y);
if (!line) {
// console.debug("First line: " + textItem.text);
@ -57,7 +57,7 @@ export default {
},
upload: function(pages) {
console.debug("Store: upload");
// console.debug("Store: upload");
this.state.uploaded = true;
this.state.pages = pages;
},

6
webpack.config.js
View File

@ -76,6 +76,12 @@ module.exports = {
new HtmlWebpackPlugin({
template: SOURCE_DIR + '/index.html'
}),
new webpack.DefinePlugin({
'process.env': {
'version': JSON.stringify(process.env.npm_package_version),
NODE_ENV: JSON.stringify(process.env.NODE_ENV || 'development')
}
}),
new CopyWebpackPlugin([
{
from: NODEMODULES_DIR + '/pdfjs-dist/build/pdf.worker.js',