Compare commits
2 Commits
1c6b0e561d
...
31d8fd53c8
| Author | SHA1 | Date | |
|---|---|---|---|
| 31d8fd53c8 | |||
| 757528b952 |
8
rangement/.idea/.gitignore
vendored
8
rangement/.idea/.gitignore
vendored
@ -1,8 +0,0 @@
|
||||
# Default ignored files
|
||||
/shelf/
|
||||
/workspace.xml
|
||||
# Datasource local storage ignored files
|
||||
/dataSources/
|
||||
/dataSources.local.xml
|
||||
# Editor-based HTTP Client requests
|
||||
/httpRequests/
|
||||
@ -1,6 +0,0 @@
|
||||
<?xml version="1.0" encoding="UTF-8"?>
|
||||
<project version="4">
|
||||
<component name="JavaScriptSettings">
|
||||
<option name="languageLevel" value="ES6" />
|
||||
</component>
|
||||
</project>
|
||||
@ -1,8 +0,0 @@
|
||||
<?xml version="1.0" encoding="UTF-8"?>
|
||||
<project version="4">
|
||||
<component name="ProjectModuleManager">
|
||||
<modules>
|
||||
<module fileurl="file://$PROJECT_DIR$/.idea/rangement.iml" filepath="$PROJECT_DIR$/.idea/rangement.iml" />
|
||||
</modules>
|
||||
</component>
|
||||
</project>
|
||||
@ -1,8 +0,0 @@
|
||||
<?xml version="1.0" encoding="UTF-8"?>
|
||||
<module type="WEB_MODULE" version="4">
|
||||
<component name="NewModuleRootManager">
|
||||
<content url="file://$MODULE_DIR$" />
|
||||
<orderEntry type="inheritedJdk" />
|
||||
<orderEntry type="sourceFolder" forTests="false" />
|
||||
</component>
|
||||
</module>
|
||||
@ -1,6 +0,0 @@
|
||||
<?xml version="1.0" encoding="UTF-8"?>
|
||||
<project version="4">
|
||||
<component name="VcsDirectoryMappings">
|
||||
<mapping directory="$PROJECT_DIR$/.." vcs="Git" />
|
||||
</component>
|
||||
</project>
|
||||
@ -1,4 +1,12 @@
|
||||
# Devine le rangement
|
||||
/!\ Ce dépot a déménagé ici:
|
||||
https://forge.chapril.org/tykayn/rangement /!\
|
||||
|
||||
installez le avec:
|
||||
```bash
|
||||
npm i -g rangement
|
||||
```
|
||||
----
|
||||
|
||||
script qui devine comment renommer des fichiers selon un pattern de date
|
||||
trouve des infos exif et prend la plus ancienne pour renseigner le nom de fichier.
|
||||
|
||||
@ -1,6 +0,0 @@
|
||||
|
||||
export const tagSeparator = ' '
|
||||
export const tagSectionSeparator = '--'
|
||||
export const enableTestsLocally = false
|
||||
export const reportStatistics = false
|
||||
export const version = '1.0.0'
|
||||
@ -1,244 +0,0 @@
|
||||
/**
|
||||
* la classe qui repère des patterns
|
||||
*/
|
||||
import { tagSectionSeparator, tagSeparator } from './configs.mjs'
|
||||
import exifr from 'exifr'
|
||||
import moment from 'moment'
|
||||
import path from 'path'
|
||||
|
||||
/**
|
||||
* finds patterns for file name
|
||||
*/
|
||||
export default class finder {
|
||||
|
||||
static statistics = {
|
||||
filesModified: 0,
|
||||
}
|
||||
static patternsFiles = {
|
||||
'downloaded_pic': /^\-\w{15}\.jpg/, // FyB8cZnWIAc21rw.jpg
|
||||
'telegram_pic': /^\-\d{19}_\d{4}/, // -4900281569878475578_1109.jpg
|
||||
'open_camera': /^IMG_OC_\d{8}/i, // IMG_OC_20230617_092120_3.jpg
|
||||
'screenshot': /^Screenshot/i, // Screenshot 2023-06-15 at 15-26-04 Instance Panoramax OSM-FR.png
|
||||
}
|
||||
|
||||
static reportStatistics () {
|
||||
console.log('statistics',
|
||||
this.statistics)
|
||||
}
|
||||
|
||||
static findScreenshot (inputString) {
|
||||
return inputString.match(/screenshot/i) || inputString.match(/capture d'écran/i)
|
||||
}
|
||||
|
||||
static findFormattedDate (filepath) {
|
||||
let match = filepath.match(/\d{4}-\d{2}-\d{2}T\d{2}:\d{2}:\d{2}/ig)
|
||||
// console.log('match findFormattedDate', match)
|
||||
let result = ''
|
||||
if (match && match[0]) {
|
||||
result = match[0]
|
||||
}
|
||||
return result
|
||||
}
|
||||
|
||||
static findFileExtension (inputString) {
|
||||
let result = inputString.match(/\.\w{3,4}$/i)
|
||||
return result
|
||||
}
|
||||
|
||||
/**
|
||||
* find the section of file name which contains the free text to describe the picture
|
||||
* @param fileName
|
||||
* @returns {*|string}
|
||||
*/
|
||||
static findFileNameFreeTextPart (fileName) {
|
||||
fileName = fileName.replace(this.findFileExtension(fileName), '')
|
||||
let boom = fileName.split(tagSectionSeparator)
|
||||
if (boom.length) {
|
||||
let freeTextPart = boom[0].trim()
|
||||
console.log('freeTextPart', freeTextPart)
|
||||
return freeTextPart
|
||||
}
|
||||
return fileName.trim()
|
||||
}
|
||||
|
||||
/**
|
||||
* find an array of tags
|
||||
* @param inputString
|
||||
* @returns {[]}
|
||||
*/
|
||||
static findTagSectionInString (inputString) {
|
||||
let listOfTags = []
|
||||
// remove extension
|
||||
let extensionFile = finder.findFileExtension(inputString)
|
||||
|
||||
if (extensionFile) {
|
||||
extensionFile = extensionFile[0]
|
||||
} else {
|
||||
console.log('no extensionFile', extensionFile, inputString)
|
||||
extensionFile = ''
|
||||
}
|
||||
inputString = inputString.replace(extensionFile, '')
|
||||
|
||||
// console.log('extensionFile', extensionFile)
|
||||
if (inputString.includes(tagSectionSeparator)) {
|
||||
// console.log('inputString', inputString)
|
||||
if (inputString.length) {
|
||||
|
||||
let boom = inputString.split(tagSectionSeparator)
|
||||
// console.log('boom', boom)
|
||||
if (boom.length) {
|
||||
let fileSectionsName = boom.splice(tagSeparator)
|
||||
listOfTags = [...fileSectionsName[1].trim().split(tagSeparator)]
|
||||
// console.log('listOfTags', listOfTags)
|
||||
} else {
|
||||
console.log('no boom', boom)
|
||||
}
|
||||
}
|
||||
}
|
||||
return listOfTags
|
||||
}
|
||||
|
||||
static cleanSpaces (inputString) {
|
||||
return inputString.trim().replace(/ *g/, ' ')
|
||||
}
|
||||
|
||||
static searchAndReplaceInFileName (searchString, replaceString, fileName) {
|
||||
return this.cleanSpaces(fileName.replace(searchString, replaceString))
|
||||
}
|
||||
|
||||
/**
|
||||
* search screenshot clues and rename
|
||||
*/
|
||||
static searchAndRenameScreenshots (fileName) {
|
||||
if (finder.findScreenshot(fileName)) {
|
||||
let tags = this.findTagSectionInString(fileName)
|
||||
console.log('tags', tags)
|
||||
if (!tags.includes('screenshot')) {
|
||||
|
||||
fileName = this.addTagInFileName('screenshot', fileName)
|
||||
fileName = this.searchAndReplaceInFileName('Screenshot', '', fileName)
|
||||
console.log('screenShotMockFileName:', fileName)
|
||||
return this.cleanSpaces(fileName)
|
||||
}
|
||||
console.log('is a screenshot, remove screenshot in name, and add tag screenshot')
|
||||
} else {
|
||||
return null
|
||||
}
|
||||
}
|
||||
|
||||
static addTagInFileName (tagName, fileName) {
|
||||
|
||||
let tags = this.findTagSectionInString(fileName)
|
||||
let firstPart = this.findFileNameFreeTextPart(fileName)
|
||||
|
||||
tags.push(tagName)
|
||||
let uniqueArray = [...new Set(tags)]
|
||||
|
||||
let newFileName = firstPart + ' ' + tagSectionSeparator + ' ' + tags.join(tagSeparator)
|
||||
newFileName = newFileName.replace(/ {*}/, '') + this.findFileExtension(fileName)
|
||||
return this.cleanSpaces(newFileName)
|
||||
}
|
||||
|
||||
/**
|
||||
* convertit un nom de fichier en une structure décrivant plusieurs parties correspondant au pattern d'archivage
|
||||
* @param fullPath
|
||||
* @returns {{extension: *, dateStamp: string, freeText: (*|string), tags: *[]}}
|
||||
*/
|
||||
static destructurateFileName (fullPath) {
|
||||
let [folderPath, fileNameOriginal] = this.findFolderPath(fullPath)
|
||||
let dateStampInFileNameOriginal = this.findFormattedDate(fileNameOriginal)
|
||||
return {
|
||||
fullPath,
|
||||
folderPath,
|
||||
fileNameOriginal,
|
||||
dateStampInFileNameOriginal,
|
||||
dateStampExif: '',
|
||||
freeText: this.findFileNameFreeTextPart(fileNameOriginal),
|
||||
tags: this.findTagSectionInString(fileNameOriginal),
|
||||
extension: this.findFileExtension(fileNameOriginal),
|
||||
}
|
||||
|
||||
}
|
||||
|
||||
/**
|
||||
* finds the earliest part in several exif date info
|
||||
* @param exifData
|
||||
* @returns {string}
|
||||
*/
|
||||
static findEarliestDateInExifData (exifData) {
|
||||
if (exifData) {
|
||||
|
||||
let moments = []
|
||||
|
||||
// console.log('exif data : ', exifData) // Do something with your data!
|
||||
if (exifData.DateTimeOriginal) {
|
||||
// console.log('image créée le : DateTimeOriginal : ', exifData.DateTimeOriginal) // Do something with your data!
|
||||
moments.push(exifData.DateTimeOriginal)
|
||||
}
|
||||
if (exifData.ModificationDateTime) {
|
||||
// console.log('image créée le : ModificationDateTime : ', exifData.ModificationDateTime) // Do something with your data!
|
||||
moments.push(exifData.ModificationDateTime)
|
||||
}
|
||||
if (exifData.ModifyDate) {
|
||||
// console.log('image créée le : ModifyDate : ', exifData.ModifyDate) // Do something with your data!
|
||||
moments.push(exifData.ModifyDate)
|
||||
}
|
||||
if (exifData.FileAccessDateTime) {
|
||||
moments.push(exifData.FileAccessDateTime)
|
||||
}
|
||||
if (exifData.FileInodeChangeDateTime) {
|
||||
moments.push(exifData.FileInodeChangeDateTime)
|
||||
}
|
||||
if (exifData.FileModificationDateTime) {
|
||||
// console.log('image créée le : FileModificationDateTime : ', exifData.FileModificationDateTime) // Do something with your data!
|
||||
moments.push(exifData.FileModificationDateTime)
|
||||
}
|
||||
if (exifData.CreateDate) {
|
||||
// console.log('image créée le : CreateDate : ', exifData.CreateDate) // Do something with your data!
|
||||
moments.push(exifData.CreateDate)
|
||||
}
|
||||
|
||||
moments = moments.map(d => {
|
||||
let newdate = moment(d)
|
||||
return newdate
|
||||
})
|
||||
let minDate = moment.min(moments)
|
||||
|
||||
// console.log('minDate :::::::::', minDate)
|
||||
console.log('minDate :::::::::', minDate.format('yyyy-MM-DDTHH:mm:ss'))
|
||||
|
||||
return minDate.format('yyyy-MM-DDTHH:mm:ss')
|
||||
} else {
|
||||
console.log('pas de exif data')
|
||||
return ''
|
||||
}
|
||||
}
|
||||
|
||||
/**
|
||||
* examine plusieurs propriétés exif de date et retourne la plus ancienne
|
||||
* @param filepath
|
||||
*/
|
||||
static async findExifCreationDate (filepath) {
|
||||
|
||||
console.log('filepath', filepath)
|
||||
let dateAlreadyInFileName = finder.findFormattedDate(filepath)
|
||||
if (dateAlreadyInFileName) {
|
||||
|
||||
console.log('------ dateAlreadyInFileName', dateAlreadyInFileName)
|
||||
}
|
||||
|
||||
return await exifr.parse(filepath)
|
||||
|
||||
}
|
||||
|
||||
static findFolderPath (filePath) {
|
||||
let folders = filePath.split('/')
|
||||
let fileName = folders.pop()
|
||||
folders = filePath.replace(fileName, '')
|
||||
|
||||
console.log('\n - folders', folders)
|
||||
console.log(' - fileName', fileName, '\n')
|
||||
return [folders, fileName]
|
||||
|
||||
}
|
||||
}
|
||||
@ -1,134 +0,0 @@
|
||||
/**---------------------
|
||||
* @name tykayn Rangement
|
||||
* @description Rangement sorts and rename files depending on their exif data
|
||||
* @contact contact@cipherbliss.com
|
||||
--------------------- */
|
||||
/** ---------------------
|
||||
libs
|
||||
--------------------- */
|
||||
import fs from 'node-fs'
|
||||
import minimist from 'minimist'
|
||||
/** ---------------------
|
||||
custom utilities and configuration
|
||||
--------------------- */
|
||||
import { enableTestsLocally, reportStatistics,tagSectionSeparator, tagSeparator } from './configs.mjs'
|
||||
import {
|
||||
TestFindFormattedDate,
|
||||
TestScreenShotIsFoundAndRenamed,
|
||||
TestTagsAreDetectedInFileName
|
||||
} from './testFunctions.mjs'
|
||||
import finder from './finder.mjs'
|
||||
|
||||
let mini_arguments
|
||||
console.log(' ')
|
||||
|
||||
function parseArguments () {
|
||||
mini_arguments = minimist(process.argv.slice(2))
|
||||
console.log('arguments', mini_arguments)
|
||||
}
|
||||
|
||||
parseArguments()
|
||||
|
||||
function renameFile (originalFileName, fileMixedNewName) {
|
||||
fs.rename(originalFileName, fileMixedNewName, function (err) {
|
||||
if (err) console.log('rename ERROR: ' + err)
|
||||
})
|
||||
}
|
||||
|
||||
function appendFileName (fileProperties, newText) {
|
||||
fileProperties.freeText = finder.cleanSpaces(fileProperties.freeText + ' ' + newText)
|
||||
return fileProperties
|
||||
}
|
||||
|
||||
function prependFileName (fileProperties, newText) {
|
||||
fileProperties.freeText = finder.cleanSpaces(newText + ' ' + fileProperties.freeText)
|
||||
return fileProperties
|
||||
}
|
||||
|
||||
function makeFileNameFromProperties(fileProperties) {
|
||||
|
||||
let tagPlace = ''
|
||||
if (fileProperties.tags.length) {
|
||||
tagPlace = ' ' + tagSectionSeparator + ' '
|
||||
}
|
||||
// return finder.cleanSpaces(fileProperties.dateStamp + ' ' + fileProperties.freeText + tagPlace + fileProperties.tags.join(tagSeparator) + fileProperties.extension).replace(+' ' + tagSectionSeparator + ' ' + '.', '.')
|
||||
return ''+fileProperties.dateStampExif + ' ' + fileProperties.freeText + tagPlace + fileProperties.tags.join(tagSeparator) + fileProperties.extension
|
||||
}
|
||||
|
||||
function shouldWeChangeName (structureForFile) {
|
||||
console.log(' ______ allez hop fini la recherche on fait un nouveau nom')
|
||||
console.log('structureForFile', structureForFile)
|
||||
let newName = makeFileNameFromProperties(structureForFile)
|
||||
if (structureForFile.fileNameOriginal !== newName) {
|
||||
|
||||
console.log('\n ancien nom :', structureForFile.fileNameOriginal)
|
||||
// console.log(' nouveau nom:', foundDate +structureForFile.freeText + structureForFile.tags.join(tagSeparator) + structureForFile.extension )
|
||||
console.log(' nouveau nom:', newName)
|
||||
} else {
|
||||
console.log(' rien à changer')
|
||||
}
|
||||
|
||||
}
|
||||
|
||||
async function guessFileNameOnAllFilesFromArguments () {
|
||||
|
||||
// parcourir les dossiers
|
||||
// parcourir les fichiers
|
||||
|
||||
console.log('liste des fichiers', mini_arguments._)
|
||||
let fileList = mini_arguments._
|
||||
|
||||
fileList.forEach(fullPath => {
|
||||
|
||||
let structureForFile = finder.destructurateFileName(fullPath)
|
||||
|
||||
// examiner les infos exif de chaque fichier pour proposer un nouveau nom
|
||||
if (!structureForFile.dateStampInFileNameOriginal) {
|
||||
console.log(' le nom de fichier ne contient pas de date formatée au début')
|
||||
|
||||
finder.findExifCreationDate(structureForFile.fullPath)
|
||||
.then(data => {
|
||||
console.log(' ... chercher la date de création')
|
||||
let foundDate = finder.findEarliestDateInExifData(data)
|
||||
|
||||
console.log(' =>>>>>>> foundDate : ', foundDate)
|
||||
if (foundDate) {
|
||||
|
||||
|
||||
// finder.findEarliestDateInExifData(fullPath).then(response => {
|
||||
// console.log(' ... trouvée')
|
||||
// if (response) {
|
||||
structureForFile.dateStampExif = foundDate
|
||||
|
||||
shouldWeChangeName(structureForFile)
|
||||
// }
|
||||
// })
|
||||
|
||||
} else {
|
||||
console.log('pas de date trouvée dans le nom')
|
||||
}
|
||||
|
||||
}
|
||||
,
|
||||
(error) => {
|
||||
console.log('/////////// Error in reading exif of file: ' + error.message)
|
||||
return ''
|
||||
})
|
||||
}
|
||||
}
|
||||
)
|
||||
|
||||
}
|
||||
|
||||
guessFileNameOnAllFilesFromArguments()
|
||||
|
||||
// run tests
|
||||
if (enableTestsLocally) {
|
||||
|
||||
TestTagsAreDetectedInFileName()
|
||||
TestFindFormattedDate()
|
||||
TestScreenShotIsFoundAndRenamed()
|
||||
}
|
||||
if (reportStatistics || mini_arguments.stats) {
|
||||
finder.reportStatistics()
|
||||
}
|
||||
@ -1,22 +0,0 @@
|
||||
import finder from "./finders.mjs";
|
||||
// const finders = require('./finders.mjs')
|
||||
|
||||
describe('rangement file name', () => {
|
||||
|
||||
test('detects date in file name', () => {
|
||||
expect(finder.findFormattedDate('2023-06-23T18.36.47 -- machin bidule.jpg')).toBe('2023-06-23T18.36.47');
|
||||
});
|
||||
test('detects file extension in file name', () => {
|
||||
expect(finder.findFileExtension()('2023-06-23T18.36.47 -- machin bidule.jpg')).toBe('jpg');
|
||||
});
|
||||
})
|
||||
console.log('finders', finder)
|
||||
|
||||
test('adding positive numbers is not zero', () => {
|
||||
for (let a = 1; a < 10; a++) {
|
||||
for (let b = 1; b < 10; b++) {
|
||||
expect(a + b).not.toBe(0);
|
||||
}
|
||||
}
|
||||
});
|
||||
|
||||
Binary file not shown.
|
Before Width: | Height: | Size: 1.7 MiB |
@ -1,48 +0,0 @@
|
||||
import finder from './finder.mjs'
|
||||
|
||||
|
||||
const pathFolder = '/home/poule/encrypted/stockage-syncable/photos/a_dispatcher/tout'
|
||||
const sortingFolder = '/home/poule/encrypted/stockage-syncable/photos/a_dispatcher'
|
||||
|
||||
const fileDefinition = {
|
||||
dateStamp: '',
|
||||
freeText: '',
|
||||
tags: [],
|
||||
extension: '',
|
||||
}
|
||||
|
||||
export function TestScreenShotIsFoundAndRenamed() {
|
||||
|
||||
let screenShotMockFileName = 'Screenshot 2023-06-15 at 15-28-21 Instance Panoramax OSM-FR.png'
|
||||
let screenShotMockFileNameExpected = '2023-06-15 at 15-28-21 Instance Panoramax OSM-FR -- screenshot.png'
|
||||
let found = finder.searchAndRenameScreenshots(screenShotMockFileName)
|
||||
console.log('found', found)
|
||||
if (found === screenShotMockFileNameExpected) {
|
||||
console.log('TestScreenShotIsFoundAndRenamed : test succès')
|
||||
} else {
|
||||
console.log('TestScreenShotIsFoundAndRenamed : FAIL:')
|
||||
console.log(found)
|
||||
console.log(screenShotMockFileNameExpected)
|
||||
}
|
||||
}
|
||||
|
||||
export function TestTagsAreDetectedInFileName() {
|
||||
let mockFileName = '2023-06-15T10:11:12 -- screeenshot festival.png'
|
||||
let expectedResult = ['screeenshot', 'festival']
|
||||
let found = finder.findTagSectionInString(mockFileName)
|
||||
if (found === expectedResult) {
|
||||
console.info('Succès')
|
||||
}
|
||||
}
|
||||
|
||||
|
||||
export function TestFindFormattedDate() {
|
||||
let mockFileName = 'Capture d\'écran 2023-06-15T10:11:12.png'
|
||||
let expectedResult = '2023-06-15T10:11:12'
|
||||
let found = finder.findFormattedDate(mockFileName)
|
||||
console.log('foundDate', found, expectedResult)
|
||||
console.log('foundDate', found)
|
||||
if (found === expectedResult) {
|
||||
console.info('Succès')
|
||||
}
|
||||
}
|
||||
@ -1,103 +0,0 @@
|
||||
import finder from './finders'
|
||||
import fs from 'node-fs'
|
||||
|
||||
/**
|
||||
----------------------- parties non réalisées -----------------------
|
||||
* work in progress
|
||||
// TODO
|
||||
---------------------------------------------------------------------
|
||||
**/
|
||||
|
||||
function TestDownloadedTelegramPictureRename (fileName) {
|
||||
let fileProperties = destructurateFileName(fileName)
|
||||
}
|
||||
|
||||
function hasDifferentDateInNameThanExif (fileName) {
|
||||
let foundDate = finder.findFormattedDate(fileName)
|
||||
|
||||
if (foundDate && foundDate != getExifCreationDate(fileName)) {
|
||||
return true
|
||||
}
|
||||
return false
|
||||
}
|
||||
|
||||
function moveToArchive (targetDirectory, fileFullPath) {
|
||||
// find current directory,
|
||||
// rename file to move it
|
||||
}
|
||||
|
||||
function getStatisticsOnArchiveFolder (fileFullPath) {
|
||||
|
||||
return {
|
||||
foldersCount: 'TODO',
|
||||
filesWithoutSemanticName: 'TODO'
|
||||
}
|
||||
}
|
||||
|
||||
function getControlledVocabularyFromFiles (fileFullPath) {
|
||||
|
||||
let controlledVocabulary = ['TODO']
|
||||
// find all tags
|
||||
return controlledVocabulary
|
||||
}
|
||||
|
||||
function moveToSortingFolder (fileFullPath) {
|
||||
return 'TODO'
|
||||
}
|
||||
|
||||
/**
|
||||
* écrit un nouveau nom de fichier formatté
|
||||
* @param convertedToName
|
||||
* @param originalFileName
|
||||
* @returns {*}
|
||||
*/
|
||||
function mixDateNameWithFileName (convertedToName, originalFileName) {
|
||||
// enlever l'ancien timestamp si il existe
|
||||
// ajouter en début de nom le nouveau timestamp avec un espace et conserver le reste du nom
|
||||
return originalFileName
|
||||
}
|
||||
|
||||
function TestMixingName () {
|
||||
|
||||
let fileMixedNewName = mixDateNameWithFileName(convertedToName, originalFileName)
|
||||
console.log('new name', fileMixedNewName)
|
||||
|
||||
if (fileMixedNewName !== originalFileName) {
|
||||
console.log('renommage =>', fileMixedNewName)
|
||||
// renameFile(originalFileName, fileMixedNewName)
|
||||
}
|
||||
}
|
||||
|
||||
|
||||
|
||||
/**
|
||||
* obtenir une liste des dossiers uniquement dans le dossier courant
|
||||
* @param path
|
||||
* @returns {*}
|
||||
*/
|
||||
function getDirectories (path) {
|
||||
return fs.readdirSync(path).filter(function (file) {
|
||||
return fs.statSync(path + '/' + file).isDirectory()
|
||||
})
|
||||
}
|
||||
|
||||
|
||||
function convertDateToTimeInFileName (inputDate) {
|
||||
return inputDate.replace(' ', 'T')
|
||||
}
|
||||
|
||||
function testthings(){
|
||||
// let list = getDirectories(pathFolder)
|
||||
// console.log('list', list)
|
||||
|
||||
let originalFileName = '2015-04-30T09.09.02 -- scan papier.jpg'
|
||||
let formattedDatePIMBefore = finder.findFormattedDate(originalFileName)
|
||||
console.log('formattedDatePIMBefore', formattedDatePIMBefore)
|
||||
|
||||
let creationDateFound = finder.getExifCreationDate(pathFolder + '/' + originalFileName)
|
||||
let convertedToName = ''
|
||||
if (creationDateFound) {
|
||||
convertedToName = convertDateToTimeInFileName(creationDateFound)
|
||||
}
|
||||
console.log('convertedToName', convertedToName)
|
||||
}
|
||||
101
scraping/main.ts
Normal file
101
scraping/main.ts
Normal file
@ -0,0 +1,101 @@
|
||||
/**
|
||||
scrapping des livres de la médiathèque de briis
|
||||
**/
|
||||
// @ts-ignore
|
||||
import https from 'https';
|
||||
import WriteFile from "./utils";
|
||||
|
||||
const axios = require('axios');
|
||||
const cheerio = require('cheerio');
|
||||
const url: string = "www.mediatheque-de-briis-sous-forges.net";
|
||||
const fetching_path: string = "/mediatheque-de-briis-sous-forges.net/opac/recherche/catalogue?node=0&value=0&page=2";
|
||||
|
||||
interface Book {
|
||||
author: string
|
||||
title: string
|
||||
description: string
|
||||
img: string
|
||||
}
|
||||
const books: Book[] = [];
|
||||
// autres pages:
|
||||
// http://www.mediatheque-de-briis-sous-forges.net/mediatheque-de-briis-sous-forges.net/opac/recherche/catalogue?node=0&value=0&page=2
|
||||
const page_max = 1927
|
||||
|
||||
const getTables = (html: string): any => {
|
||||
const $ = cheerio.load(html);
|
||||
const tableElements = $(
|
||||
"table.notice"
|
||||
);
|
||||
return tableElements;
|
||||
};
|
||||
|
||||
const getHtml = async (hostname: string, path: string): Promise<string> =>
|
||||
new Promise((resolve, reject) => {
|
||||
https
|
||||
.get(
|
||||
{
|
||||
hostname,
|
||||
path,
|
||||
// port:80,
|
||||
method: "GET",
|
||||
},
|
||||
(res) => {
|
||||
let html = "";
|
||||
res.on("data", function (chunk) {
|
||||
html += chunk;
|
||||
});
|
||||
res.on("end", function () {
|
||||
resolve(html);
|
||||
});
|
||||
}
|
||||
)
|
||||
.on("error", (error) => {
|
||||
console.error(error);
|
||||
reject(error);
|
||||
});
|
||||
});
|
||||
|
||||
// fetchData(url).then((res: any) => {
|
||||
// const html = res.data;
|
||||
// const $ = cheerio.load(html);
|
||||
// const statsTable :any = $('table.notice');
|
||||
// console.log('statsTable', statsTable)
|
||||
// statsTable.each(function(){
|
||||
// let elem:any = this;
|
||||
// let author = $(elem).find('td').eq(2).text();
|
||||
// // let img = $(this).find('img').attr('src');
|
||||
// // console.log(elem);
|
||||
// console.log(author);
|
||||
// });
|
||||
// })
|
||||
|
||||
|
||||
function writeBookScrapping() {
|
||||
WriteFile('books.json', JSON.stringify(books, null , 2))
|
||||
}
|
||||
|
||||
getHtml(url, fetching_path)
|
||||
.then(getTables)
|
||||
.then(
|
||||
(tables: any) => tables.each(
|
||||
(_: any, table: any) => {
|
||||
const $ = cheerio.load(table);
|
||||
// console.log('une table')
|
||||
// let author = $().find('a.notice').text();
|
||||
let text_description = $(table).find('td').eq(1).text();
|
||||
let boom = text_description.split('\n');
|
||||
let splitting = boom[1].split('/')
|
||||
let img_src = $(table).find('td img').attr('src');
|
||||
console.log(img_src);
|
||||
books.push({
|
||||
author: boom[0],
|
||||
title: splitting[0],
|
||||
description: splitting[1],
|
||||
img: img_src
|
||||
})
|
||||
// console.log(cheerio.load(table).html())
|
||||
}
|
||||
)
|
||||
)
|
||||
.then(writeBookScrapping)
|
||||
.catch((error) => console.log(error));
|
||||
62
scraping/output/books.json
Normal file
62
scraping/output/books.json
Normal file
@ -0,0 +1,62 @@
|
||||
[
|
||||
{
|
||||
"author": "ALLENDE, Isabel",
|
||||
"title": "Portrait sépia ",
|
||||
"description": " Isabel Allende ; Trad. de l'espagnol par Claude de Frayssinet. - Paris : Grasset et Fasquelle, 2001. - 1 vol. , 391 p. : couv. ill. ; 24 x 15 cm.",
|
||||
"img": "http://images-eu.amazon.com/images/P/2246617715.08.MZZZZZZZ.jpg"
|
||||
},
|
||||
{
|
||||
"author": "AMETTE, Jacques-Pierre",
|
||||
"title": "La Maîtresse de Brecht : roman ",
|
||||
"description": " Jacques-Pierre Amette. - Paris : Albin Michel, 2003. - 300 p. : jaquette ill. ; 20 cm.",
|
||||
"img": "http://images-eu.amazon.com/images/P/2226141634.08.MZZZZZZZ.jpg"
|
||||
},
|
||||
{
|
||||
"author": "ANDRIÂC, Ivo",
|
||||
"title": "Mara la courtisane : et autres nouvelles ",
|
||||
"description": " Ivo Andriâc ; Trad. du serbo-croate par Pascale Delpech. - Paris : Belfond, 1999. - 234 p. : couv. ill. en coul. ; 23 cm. - (Littérature étrangère).",
|
||||
"img": "http://images-eu.amazon.com/images/P/2714435572.08.MZZZZZZZ.jpg"
|
||||
},
|
||||
{
|
||||
"author": "ANGLADE, Jean",
|
||||
"title": "Un Lit d'aubépine : roman ",
|
||||
"description": " Jean Anglade. - Paris : Presses de la Cité, 1995. - 325 p. : couv. ill. en coul. ; 23 cm. - (Production Jeannine Balland).",
|
||||
"img": "http://images-eu.amazon.com/images/P/2258039568.08.MZZZZZZZ.jpg"
|
||||
},
|
||||
{
|
||||
"author": "ARNOTHY, Christine",
|
||||
"title": "J'ai quinze ans et je ne veux pas mourir ; (suivi de) Il n'est pas si facile de vivre ",
|
||||
"description": " Christine Arnothy. - Paris : France loisirs, 1981. - 330 p ; 23 cm.",
|
||||
"img": "http://images-eu.amazon.com/images/P/2724211065.08.MZZZZZZZ.jpg"
|
||||
},
|
||||
{
|
||||
"author": "CHAUVIN, Rémy",
|
||||
"title": "Le Monde animal et ses comportements complexes ",
|
||||
"description": " Rémy Chauvin, Bernadette Chauvin. - Paris : Plon, 1977. - 282 p : ill ; 21 cm.",
|
||||
"img": "http://images-eu.amazon.com/images/P/2259002331.08.MZZZZZZZ.jpg"
|
||||
},
|
||||
{
|
||||
"author": "D'ARZO, Silvio",
|
||||
"title": "Maison des autres , (Contient) Un moment comme ça ",
|
||||
"description": " texte de Silvio D'Arzo ; Trad. de l'italien par Bernard Simeone, Philippe Renard ; Préf. Attilio Bertolucci. - Lagrasse : Verdier, 1997. - 1 vol. , 86 p. : - ; 22 x 14 cm. - (Terra d'altri, ISSN 0989-4160).",
|
||||
"img": "http://images-eu.amazon.com/images/P/2864322838.08.MZZZZZZZ.jpg"
|
||||
},
|
||||
{
|
||||
"author": "ASSOULINE, Pierre",
|
||||
"title": "La Cliente : roman ",
|
||||
"description": " Pierre Assouline. - Paris : Gallimard, 1998. - 191 p. ; 21 cm.",
|
||||
"img": "http://images-eu.amazon.com/images/P/207075278X.08.MZZZZZZZ.jpg"
|
||||
},
|
||||
{
|
||||
"author": "ATKINSON, Kate",
|
||||
"title": "Dans les replis du temps ",
|
||||
"description": " texte de Kate Atkinson ; Trad. de l'anglais par Jean Bourdier. - Paris : Librairie générale française, 1999. - 1 vol. , 403 p. : ill., couv. ill. en coul. ; 18 x 11 cm. - (Le livre de poche ; 14687).",
|
||||
"img": "http://images-eu.amazon.com/images/P/2253146870.08.MZZZZZZZ.jpg"
|
||||
},
|
||||
{
|
||||
"author": "ATKINSON, Kate",
|
||||
"title": "Dans les coulisses du musée : roman ",
|
||||
"description": " Kate Atkinson ; Trad. de l'anglais par Jean Bourdier. - Paris : Bernard de Fallois, 1996. - 348 p. : couv. ill. en coul. ; 23 cm.",
|
||||
"img": "http://images-eu.amazon.com/images/P/2877062775.08.MZZZZZZZ.jpg"
|
||||
}
|
||||
]
|
||||
15
scraping/utils.ts
Normal file
15
scraping/utils.ts
Normal file
@ -0,0 +1,15 @@
|
||||
import * as fs from "node:fs";
|
||||
|
||||
export default function WriteFile(fileName: string, fileContent: any) {
|
||||
console.log('write file', fileName)
|
||||
return fs.writeFile(
|
||||
`./output/${fileName}`,
|
||||
fileContent,
|
||||
'utf8',
|
||||
(err) => {
|
||||
if (err) {
|
||||
console.log(`Error writing file: ${err}`)
|
||||
}
|
||||
}
|
||||
)
|
||||
}
|
||||
Loading…
Reference in New Issue
Block a user