${displayTs}` + contentHtml + `
` + ); + } + const audioMeta = audioSource ? ` +` : ""; + const html = [ + '', + ` +${audioMeta}`, + ...paragraphs, + "", + "" + ].join("\n"); + return { html, mismatch }; +} function lineGroup(line) { if (line.type === "meta" || line.type === "dependent") return "structural"; @@ -36317,6 +37140,11 @@ var PseudObsPlugin = class extends import_obsidian14.Plugin { name: t("command.exportAsCha"), callback: () => void this.exportCurrentFileAsFormat("cha") }); + this.addCommand({ + id: "export-as-html", + name: t("command.exportAsHtml"), + callback: () => void this.exportCurrentFileAsHtml() + }); this.addCommand({ id: "pseudonymize-selection", name: t("command.pseudonymizeSelection"), @@ -36482,32 +37310,32 @@ var PseudObsPlugin = class extends import_obsidian14.Plugin { * Calcule le chemin de l'export final selon les paramètres de destination. * Retourne { vaultPath } si dans le vault, { externalPath } si hors vault. */ + /** Retourne le sous-dossier de classe (vide si pas de mirroring ou pas de classe). */ + getClassSub(file) { + const s = this.settings; + if (!s.exportMirrorClasses) + return ""; + const transcRoot = s.transcriptionsFolder; + const fileFolder = file.parent?.path ?? ""; + const rawBase = file.basename.replace(/\.pseudonymized$/, ""); + if (fileFolder.startsWith(transcRoot)) { + return fileFolder.slice(transcRoot.length).replace(/^\//, ""); + } + const mappingFile = this.findInMappings(`${rawBase}.mapping.json`); + if (mappingFile) { + const mFolder = mappingFile.parent?.path ?? ""; + if (mFolder.startsWith(s.mappingFolder)) { + return mFolder.slice(s.mappingFolder.length).replace(/^\//, ""); + } + } + return ""; + } resolveExportPath(file, ext) { const s = this.settings; const rawBase = file.basename.replace(/\.pseudonymized$/, ""); const outputName = `${rawBase}.pseudonymized.${ext}`; - const fileFolder = file.parent?.path ?? ""; - let classSub = ""; - if (s.exportMirrorClasses) { - const transcRoot = s.transcriptionsFolder; - const fileFolder2 = file.parent?.path ?? ""; - if (fileFolder2.startsWith(transcRoot)) { - classSub = fileFolder2.slice(transcRoot.length).replace(/^\//, ""); - } else { - const mappingFile = this.findInMappings(`${rawBase}.mapping.json`); - if (mappingFile) { - const mFolder = mappingFile.parent?.path ?? ""; - const mRoot = s.mappingFolder; - if (mFolder.startsWith(mRoot)) { - classSub = mFolder.slice(mRoot.length).replace(/^\//, ""); - } - } - } - } + const classSub = this.getClassSub(file); const join2 = (base) => classSub ? `${base}/${classSub}/${outputName}` : `${base}/${outputName}`; - if (s.exportDestinationType === "next-to-source") { - return { vaultPath: `${fileFolder}/${outputName}`, externalPath: null }; - } if (s.exportDestinationType === "external" && s.exportExternalPath) { return { vaultPath: null, externalPath: join2(s.exportExternalPath) }; } @@ -36890,6 +37718,11 @@ var PseudObsPlugin = class extends import_obsidian14.Plugin { } // --- Conversion automatique --- async autoConvert(file) { + const s = this.settings; + const fileFolder = file.parent?.path ?? ""; + const outputFolders = [s.exportsFolder, s.mappingFolder, s.exportFinalFolder].filter(Boolean); + if (outputFolders.some((f) => fileFolder === f || fileFolder.startsWith(f + "/"))) + return; try { const raw = await this.app.vault.read(file); const ext = file.extension.toLowerCase(); @@ -36932,8 +37765,8 @@ var PseudObsPlugin = class extends import_obsidian14.Plugin { } await this.app.vault.create(mdPath, mdContent); const transcRoot = this.settings.transcriptionsFolder; - const fileFolder = file.parent?.path ?? ""; - const relSubFolder = fileFolder.startsWith(transcRoot) ? fileFolder.slice(transcRoot.length).replace(/^\//, "") : ""; + const fileFolder2 = file.parent?.path ?? ""; + const relSubFolder = fileFolder2.startsWith(transcRoot) ? fileFolder2.slice(transcRoot.length).replace(/^\//, "") : ""; const mappingDir = relSubFolder ? `${this.settings.mappingFolder}/${relSubFolder}` : this.settings.mappingFolder; await this.ensureFolder(mappingDir); const mappingPath = `${mappingDir}/${basename}.mapping.json`; @@ -37106,8 +37939,10 @@ var PseudObsPlugin = class extends import_obsidian14.Plugin { } else { pseudonymized = engine.pseudonymize(content, rules, marker); } - await this.ensureFolder(this.settings.exportsFolder); - const outputPath = `${this.settings.exportsFolder}/${file.basename}.pseudonymized.${ext}`; + const classSub = this.getClassSub(file); + const exportBase = classSub ? `${this.settings.exportsFolder}/${classSub}` : this.settings.exportsFolder; + await this.ensureFolder(exportBase); + const outputPath = `${exportBase}/${file.basename}.pseudonymized.${ext}`; const existing = this.app.vault.getAbstractFileByPath(outputPath); if (existing instanceof import_obsidian14.TFile) { await this.app.vault.modify(existing, pseudonymized); @@ -37276,6 +38111,37 @@ var PseudObsPlugin = class extends import_obsidian14.Plugin { } await this.writeExport(file, "vtt", vtt); } + /** + * Re-exporte le Markdown noScribe actif au format HTML noScribe pseudonymisé. + * Lit le .words.json correspondant pour les timestamps précis. + */ + async exportCurrentFileAsHtml() { + const file = this.getActiveOrLastFile(); + if (!file || file.extension !== "md") { + new import_obsidian14.Notice(t("notice.noActiveFile")); + return; + } + const content = await this.app.vault.read(file); + const formatMatch = /^pseudobs-format:\s*(\w+)/m.exec(content); + if (formatMatch?.[1] !== "html") { + new import_obsidian14.Notice(t("notice.notNoScribeFormat")); + return; + } + const rawBasename = file.basename.replace(/\.pseudonymized$/, ""); + const wordsJson = await this.findWordsJson(rawBasename); + if (!wordsJson) { + new import_obsidian14.Notice(t("notice.wordsJsonMissing", rawBasename)); + return; + } + const audioMatch = /^pseudobs-audio:\s*"([^"]+)"/m.exec(content); + const audioSource = audioMatch?.[1]; + const wordData = JSON.parse(wordsJson); + const { html, mismatch } = markdownToNoScribeHtml(content, wordData, audioSource); + if (mismatch) { + new import_obsidian14.Notice(t("notice.vttMismatch")); + } + await this.writeExport(file, "html", html, "notice.htmlExported"); + } /** * Re-exporte le Markdown pseudonymisé courant au format d'origine (srt ou cha). */ @@ -37301,7 +38167,7 @@ var PseudObsPlugin = class extends import_obsidian14.Plugin { * Écrit le contenu d'un export final selon les paramètres de destination. * Gère vault, next-to-source et external. */ - async writeExport(file, ext, content) { + async writeExport(file, ext, content, noticeKey = "notice.vttExported") { const dest = this.resolveExportPath(file, ext); if (dest.externalPath) { try { @@ -37309,7 +38175,7 @@ var PseudObsPlugin = class extends import_obsidian14.Plugin { const nodePath = require("path"); await nodeFs.promises.mkdir(nodePath.dirname(dest.externalPath), { recursive: true }); await nodeFs.promises.writeFile(dest.externalPath, content, "utf-8"); - new import_obsidian14.Notice(t("notice.vttExported", dest.externalPath)); + new import_obsidian14.Notice(t(noticeKey, dest.externalPath)); } catch (e) { new import_obsidian14.Notice(`Export error: ${e.message}`); } @@ -37325,7 +38191,7 @@ var PseudObsPlugin = class extends import_obsidian14.Plugin { } else { await this.app.vault.create(outputPath, content); } - new import_obsidian14.Notice(t("notice.vttExported", outputPath)); + new import_obsidian14.Notice(t(noticeKey, outputPath)); } /** Chercheavec deux ancres ts_ : le timestamp d'affichage + * et le texte. Les timestamps word-level ne sont pas reconstitués (le texte + * pseudonymisé ne s'aligne plus mot à mot sur l'original). + */ +export function markdownToNoScribeHtml( + mdContent: string, + wordData: VttCueData[], + audioSource?: string, +): { html: string; mismatch: boolean } { + const bodyMatch = /^---\n[\s\S]*?\n---\n+([\s\S]*)$/.exec(mdContent); + const body = bodyMatch ? bodyMatch[1] : mdContent; + + const cueLines = body.split('\n').filter((l) => MD_CUE_RE.test(l.trim())); + const mismatch = cueLines.length !== wordData.length; + const count = Math.min(cueLines.length, wordData.length); + + const P_STYLE = ' margin-top:0px; margin-bottom:0px; margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px;'; + const paragraphs: string[] = []; + + for (let i = 0; i < count; i++) { + const m = MD_CUE_RE.exec(cueLines[i].trim())!; + const speaker = m[1]?.trim() || wordData[i].speaker || ''; + const text = m[2]?.trim() ?? ''; + if (!text) continue; + + const cueData = wordData[i]; + const cueStartMs = tsToMs(cueData.startTime); + const cueEndMs = tsToMs(cueData.endTime); + const speakerId = speaker ? sanitizeSpeakerId(speaker) : ''; + const cueAnchor = `ts_${cueStartMs}_${cueEndMs}_${speakerId}`; + const displayTs = `[${cueData.startTime.slice(0, 8)}]`; + + // Ancres word-level : mapper les tokens pseudonymisés sur les timestamps d'origine + // par index proportionnel (floor). Un pseudonyme hérite du timestamp du nom original. + const origWords = cueData.words.filter(w => w.time !== ''); + const pseudoTokens = text.split(/\s+/).filter(Boolean); + + let contentHtml: string; + + if (origWords.length === 0 || pseudoTokens.length === 0) { + // Pas de timestamps word-level : ancre unique pour toute la réplique + const prefix = speaker ? `${escapeHtml(speaker)} : ` : ''; + contentHtml = `${prefix}${escapeHtml(text)}`; + } else { + const wordStartMs = origWords.map(w => tsToMs(w.time)); + const wordEndMs = origWords.map((_, j) => + j + 1 < origWords.length ? wordStartMs[j + 1] : cueEndMs + ); + + // Mapping floor : token pseudo j → orig word floor(j * M / N) + const mapToOrig = (j: number): number => + Math.min(Math.floor(j * origWords.length / pseudoTokens.length), origWords.length - 1); + + // Grouper les tokens consécutifs qui partagent le même index orig + interface Group { origIdx: number; tokens: string[] } + const groups: Group[] = []; + for (let j = 0; j < pseudoTokens.length; j++) { + const origIdx = mapToOrig(j); + if (groups.length === 0 || groups[groups.length - 1].origIdx !== origIdx) { + groups.push({ origIdx, tokens: [] }); + } + groups[groups.length - 1].tokens.push(pseudoTokens[j]); + } + + const parts: string[] = []; + for (let g = 0; g < groups.length; g++) { + const { origIdx, tokens } = groups[g]; + const anchorName = `ts_${wordStartMs[origIdx]}_${wordEndMs[origIdx]}_${speakerId}`; + let groupText = tokens.join(' '); + // Espace séparateur entre ancres (sauf la dernière) + if (g < groups.length - 1) groupText += ' '; + // Préfixe locuteur sur le premier groupe + const content = (g === 0 && speaker) + ? `${escapeHtml(speaker)} : ${escapeHtml(groupText)}` + : escapeHtml(groupText); + parts.push(`${content}`); + } + contentHtml = parts.join(''); + } + + paragraphs.push( + `
` + + `${displayTs}` + + contentHtml + + `
` + ); + } + + const audioMeta = audioSource + ? `\n` + : ''; + + const html = [ + '', + `\n${audioMeta}`, + ...paragraphs, + '', + '', + ].join('\n'); + + return { html, mismatch }; +} + function lineGroup(line: ChatLine): 'structural' | 'turn' | null { if (line.type === 'meta' || line.type === 'dependent') return 'structural'; if (line.type === 'turn' || line.type === 'continuation') return 'turn'; diff --git a/src/settings.ts b/src/settings.ts index ced3b12..c407c73 100644 --- a/src/settings.ts +++ b/src/settings.ts @@ -5,7 +5,7 @@ import { OnboardingModal } from './ui/OnboardingModal'; import { t, setLocale, AVAILABLE_LANGUAGES } from './i18n'; export type NerBackend = 'none' | 'spacy' | 'transformers-js'; -export type ExportDestinationType = 'vault' | 'next-to-source' | 'external'; +export type ExportDestinationType = 'vault' | 'external'; export interface PseudObsSettings { transcriptionsFolder: string; diff --git a/src/ui/CorpusModal.ts b/src/ui/CorpusModal.ts index 2bf3aa1..98cb74b 100644 --- a/src/ui/CorpusModal.ts +++ b/src/ui/CorpusModal.ts @@ -21,10 +21,12 @@ export function getCorpusClasses(app: App, transcriptionsFolder: string): string */ export class CorpusModal extends Modal { private plugin: PseudObsPlugin; + private onCloseCb?: () => void; - constructor(app: App, plugin: PseudObsPlugin) { + constructor(app: App, plugin: PseudObsPlugin, onCloseCb?: () => void) { super(app); this.plugin = plugin; + this.onCloseCb = onCloseCb; } onOpen(): void { @@ -145,6 +147,7 @@ export class CorpusModal extends Modal { onClose(): void { this.contentEl.empty(); + this.onCloseCb?.(); } } diff --git a/src/ui/PseudonymizationView.ts b/src/ui/PseudonymizationView.ts index 6d8bffe..e1a12ec 100644 --- a/src/ui/PseudonymizationView.ts +++ b/src/ui/PseudonymizationView.ts @@ -3,7 +3,7 @@ import { t } from '../i18n'; import type { DictionaryFile } from '../types'; import type PseudObsPlugin from '../main'; import { scanOccurrences } from '../scanner/OccurrenceScanner'; -import { getCorpusClasses } from './CorpusModal'; +import { getCorpusClasses, CorpusModal } from './CorpusModal'; import { FolderSuggest } from './FolderSuggest'; import { EditRuleModal } from './EditRuleModal'; import { RuleModal } from './RuleModal'; @@ -211,9 +211,8 @@ export class PseudonymizationView extends ItemView { // Sélecteur de type const typeRow = exportSection.createDiv('pseudobs-corpus-export-type-row'); for (const [val, labelKey] of [ - ['vault', 'panel.corpus.exportDest.vault'], - ['next-to-source', 'panel.corpus.exportDest.nextToSource'], - ['external', 'panel.corpus.exportDest.external'], + ['vault', 'panel.corpus.exportDest.vault'], + ['external', 'panel.corpus.exportDest.external'], ] as [string, string][]) { const lbl = typeRow.createEl('label', { cls: 'pseudobs-corpus-export-type-label' }); const radio = lbl.createEl('input'); @@ -270,12 +269,8 @@ export class PseudonymizationView extends ItemView { text: `+ ${t('panel.corpus.addClass')}`, cls: 'pseudobs-corpus-add-class-btn', }); - addClassBtn.addEventListener('click', async () => { - const name = this.promptText(t('corpus.modal.classNamePlaceholder')); - if (!name) return; - await this.plugin.ensureFolder(`${transcRoot}/${name}`); - await this.plugin.ensureFolder(`${s.mappingFolder}/${name}`); - void this.renderTab('corpus'); + addClassBtn.addEventListener('click', () => { + new CorpusModal(this.app, this.plugin, () => void this.renderTab('corpus')).open(); }); // ---- Liste des fichiers par classe ---------------------------- @@ -723,6 +718,16 @@ export class PseudonymizationView extends ItemView { void this.plugin.exportCurrentFileAsVtt(); }) ); + if (format === 'html') { + new Setting(el) + .setName(t('panel.exports.exportHtml')) + .setDesc(t('panel.exports.exportHtml.desc')) + .addButton((btn) => + btn.setButtonText(t('command.exportAsHtml')).onClick(() => { + void this.plugin.exportCurrentFileAsHtml(); + }) + ); + } } else if (format === 'srt') { new Setting(el) .setName(t('panel.exports.exportSrt'))