@tricoteuses/senat 2.15.7 → 2.16.0

package/lib/scripts/retrieve_videos.js

@@ -7,8 +7,9 @@ import path from "path";
 import { AGENDA_FOLDER, DATA_TRANSFORMED_FOLDER, iterLoadSenatAgendas } from "../loaders";
 import { getSessionsFromStart } from "../types/sessions";
 import { commonOptions } from "./shared/cli_helpers";
+import { decodeHtmlEntities } from "../model/util";
 // ===================== Constants =====================
-const MATCH_THRESHOLD = 0.6;
+const MATCH_THRESHOLD = 0.56;
 const MAX_CANDIDATES = 15;
 const MAX_PAGES = 3;
 const STATS = { total: 0, accepted: 0 };
@@ -145,7 +146,9 @@ function extractCandidatesFromSearchHtml(html) {
         const pageUrl = `https://videos.senat.fr/video.${id}_${hash}.html`;
         const ctx = html.slice(Math.max(0, m.index - 240), Math.min(html.length, m.index + 240));
         const t = ctx.match(/title="([^"]+)"/i) || ctx.match(/>([^<]{10,200})</);
-        out.push({ id, hash, pageUrl, title: t?.[1] });
+        const title = t?.[1]?.trim();
+        const isSeancePublique = title?.toLowerCase().includes("séance publique") ?? false;
+        out.push({ id, hash, pageUrl, title, isSeancePublique });
     }
     const seen = new Set();
     return out.filter((c) => {
@@ -157,20 +160,28 @@ function extractCandidatesFromSearchHtml(html) {
     });
 }
 function parseDataNvs(nvs) {
-    const epoch = nvs.match(/<metadata\s+name="date"\s+value="(\d+)"/i)?.[1];
-    const title = nvs.match(/<metadata\s+name="title"\s+value="([^"]+)"/i)?.[1];
-    return { epoch: epoch ? Number(epoch) : undefined, title };
+    const epochStr = nvs.match(/<metadata\s+name="date"\s+value="(\d+)"/i)?.[1];
+    const epoch = epochStr ? Number(epochStr) : undefined;
+    const organesTag = nvs.match(/<metadata\b[^>]*\bname="organes"[^>]*>/i)?.[0];
+    let organeLabel;
+    let organeValue;
+    if (organesTag) {
+        organeLabel = organesTag.match(/\blabel="([^"]+)"/i)?.[1];
+        organeValue = organesTag.match(/\bvalue="([^"]+)"/i)?.[1];
+    }
+    const organeRaw = organeLabel ?? organeValue;
+    const organe = decodeHtmlEntities(organeRaw)?.trim();
+    const firstChapterLabel = decodeHtmlEntities(nvs.match(/<chapter\b[^>]*\blabel="([^"]+)"/i)[1]).trim();
+    return { epoch, organe, firstChapterLabel };
 }
-// nvsText = contenu texte de data.nvs (utf-8)
-// finalText = contenu texte de finalplayer.nvs (utf-8)
-export function buildSenatVodMasterM3u8FromNvs(nvsText, finalText) {
-    // 1) Base Akamai depuis data.nvs (mp4 "serverfiles://senat/YYYY/MM/encoderX_YYYYMMDDHHMMSS_1.mp4")
+function buildSenatVodMasterM3u8FromNvs(nvsText, finalText) {
+    // 1) Base Akamai from data.nvs (mp4 "serverfiles://senat/YYYY/MM/encoderX_YYYYMMDDHHMMSS_1.mp4")
     const baseMatch = nvsText.match(/serverfiles:\/\/senat\/(\d{4})\/(\d{2})\/(encoder\d)_(\d{14})/i);
     if (!baseMatch)
         return null;
     const [, yyyy, mm, encoder, stamp] = baseMatch;
     const base = `https://vodsenat.akamaized.net/senat/${yyyy}/${mm}/${encoder}_${stamp}`;
-    // 2) start/end depuis finalplayer.nvs
+    // 2) start/end from finalplayer.nvs
     let start = null, end = null;
     const playerAttr = finalText.match(/player[^>]*\bstarttime="(\d+)"[^>]*\bendtime="(\d+)"/i);
     if (playerAttr) {
@@ -178,12 +189,11 @@ export function buildSenatVodMasterM3u8FromNvs(nvsText, finalText) {
         end = parseInt(playerAttr[2], 10);
     }
     else {
-        // fallback: prendre le plus petit timecode des <synchro timecode="...">
+        // fallback: take smallest timecode of <synchro timecode="...">
         const tc = Array.from(finalText.matchAll(/timecode="(\d+)"/g)).map((m) => parseInt(m[1], 10));
         if (tc.length)
             start = Math.min(...tc);
     }
-    // 3) si pas d'end, on peut déduire via "duree" (en secondes) de data.nvs
     if (end == null) {
         const durMeta = nvsText.match(/<metadata[^>]*\bname="duree"[^>]*\bvalue="(\d+)"[^>]*>/i);
         if (durMeta && start != null) {
@@ -191,16 +201,14 @@ export function buildSenatVodMasterM3u8FromNvs(nvsText, finalText) {
             end = start + durMs;
         }
     }
-    // 4) Construction de l’URL
-    //    - si on a start & end → utiliser ps/pd (robuste et conforme à ce que sert le Sénat)
-    //    - sinon fallback sans suffixe (souvent valide aussi)
+    // 4) Construct URL
     if (start != null && end != null && end > start) {
         const pd = end - start;
         return `${base}_ps${start}_pd${pd}.smil/master.m3u8`;
     }
     return `${base}.smil/master.m3u8`;
 }
-export function score(agenda, agendaTs, videoTitle, videoEpoch) {
+function score(agenda, agendaTs, videoTitle, videoEpoch, videoOrgane) {
     const titleScore = dice(agenda.titre || "", videoTitle || "");
     let timeScore = 0;
     if (agendaTs && videoEpoch) {
@@ -209,15 +217,11 @@ export function score(agenda, agendaTs, videoTitle, videoEpoch) {
         // delta : 180min
         timeScore = Math.max(0, 1 - deltaMin / 180);
     }
-    let orgBonus = 0;
-    if (agenda.organe && videoTitle) {
-        const o = normalize(agenda.organe);
-        const t = normalize(videoTitle);
-        const first = o.split(" ").filter(Boolean)[0];
-        if (first && t.includes(first))
-            orgBonus = 0.15;
+    const orgScore = videoOrgane && agenda.organe ? dice(agenda.organe, videoOrgane) : 0;
+    if (orgScore === 0 && agenda.organe === "Séance publique") {
+        return 0.5 * titleScore + 0.5 * timeScore;
     }
-    return 0.3 * titleScore + 0.7 * timeScore + orgBonus; // Can be adjusted
+    return 0.4 * titleScore + 0.3 * timeScore + orgScore * 0.3;
 }
 /**
  * Build search strategies for senat's videos
@@ -263,19 +267,16 @@ async function fetchAllSearchPages(args, baseDir, strategyIndex, maxPages = MAX_
 async function processGroupedReunion(agenda, session, dataDir) {
     if (!agenda)
         return;
-    // 1) Garde-fous
+    // 1) GuardRails
     if (!agenda.captationVideo) {
-        if (!options["silent"])
-            console.log(`[skip] ${agenda.uid} captationVideo=false`);
+        // if (!options["silent"]) console.log(`[skip] ${agenda.uid} captationVideo=false`)
         return;
     }
     if (!agenda.date || !agenda.startTime) {
-        if (!options["silent"])
-            console.log(`[skip] ${agenda.uid} date/hour missing`);
+        // if (!options["silent"]) console.log(`[skip] ${agenda.uid} date/hour missing`)
         return;
     }
     STATS.total++;
-    // 2) Dossier de sortie (utilise directement l'UID)
     const reunionUid = agenda.uid;
     const baseDir = path.join(dataDir, VIDEOS_ROOT_FOLDER, String(session), reunionUid);
     await fs.ensureDir(baseDir);
@@ -312,9 +313,29 @@ async function processGroupedReunion(agenda, session, dataDir) {
         if (!buf)
             continue;
         const meta = parseDataNvs(buf.toString("utf-8"));
-        const s = score(agenda, agendaTs, c.title ?? meta.title, meta.epoch);
+        // If organes are different, go to next candidates
+        if (meta.organe && agenda.organe) {
+            const videoOrgNorm = normalize(meta.organe);
+            const agendaOrgNorm = normalize(agenda.organe);
+            if (dice(agendaOrgNorm, videoOrgNorm) < 0.5) {
+                continue;
+            }
+        }
+        let videoTitle = c.title;
+        if (c.isSeancePublique && meta.firstChapterLabel) {
+            videoTitle = meta.firstChapterLabel;
+        }
+        const s = score(agenda, agendaTs, videoTitle, meta.epoch, meta.organe);
         if (!best || s > best.score) {
-            best = { id: c.id, hash: c.hash, pageUrl: c.pageUrl, epoch: meta.epoch, vtitle: c.title ?? meta.title, score: s };
+            best = {
+                id: c.id,
+                hash: c.hash,
+                pageUrl: c.pageUrl,
+                epoch: meta.epoch,
+                vtitle: videoTitle,
+                score: s,
+                vorgane: meta.organe,
+            };
         }
     }
     if (!best) {
@@ -326,7 +347,10 @@ async function processGroupedReunion(agenda, session, dataDir) {
     if (accepted)
         STATS.accepted++;
     if (!options["silent"]) {
-        console.log(`[pick] ${agenda.uid} best id=${best.id} hash=${best.hash} score=${best.score.toFixed(2)} accepted=${accepted} (strategy=${usedStrategy})`);
+        console.log(`[pick] ${agenda.uid} score=${best.score.toFixed(2)} 
+      agenda title="${agenda.titre ?? ""}" agenda organe="${agenda.organe ?? ""}"
+      best title="${best.vtitle ?? ""}" best organe="${best.vorgane ?? ""}"
+      accepted=${accepted} (strategy=${usedStrategy})`);
     }
     // ==== 3) Write metadata + NVS of the best candidate (always) ====
     const bestDt = best?.epoch ? epochToParisDateTime(best.epoch) : null;

package/lib/types/agenda.d.ts

@@ -1,11 +1,11 @@
 export interface AgendaEvent {
     id: string;
     type: string | null;
-    date: string | null;
+    date: string;
     startTime: string | null;
     endTime: string | null;
     timeOriginal: string | null;
-    titre: string | null;
+    titre: string;
     organe: string | null;
     objet: string | null;
     lieu: string | null;

package/lib/utils/cr_spliting.js

@@ -349,12 +349,13 @@ export async function linkCRtoCommissionGroup(opts) {
             uid,
             chambre: "SN",
             date: dateISO,
-            type: organeDetected ?? "Commissions",
+            type: "Commission",
             startTime: hourShortToStartTime(hourShort),
             endTime: null,
             captationVideo: false,
             titre: titreGuess ?? "",
-            objet: "",
+            organe: organeDetected ?? "Commission",
+            objet: titreGuess ?? "",
             events: [],
             compteRenduRefUid: crUid,
         };
@@ -364,6 +365,7 @@ export async function linkCRtoCommissionGroup(opts) {
     else {
         group.compteRenduRefUid = crUid;
         updated = true;
+        console.log(`[AGENDA][COM] Updated group uid=${uid} for CR uid=${crUid}`);
     }
     // Lien CR
     // Enrichir depuis CR si vide

package/lib/utils/reunion_grouping.d.ts

@@ -2,7 +2,7 @@ import { AgendaEvent, GroupedReunion, TimeSlot } from "../types/agenda";
 type KnownType = "SP" | "COM" | "MC" | "OD" | "ID";
 export declare function groupNonSPByTypeOrganeHour(events: AgendaEvent[]): Record<"IDC" | "IDM" | "IDO" | "IDI", GroupedReunion[]>;
 export declare function groupSeancePubliqueBySlot(events: AgendaEvent[]): Record<TimeSlot, GroupedReunion[]>;
-export declare function makeTypeGroupUid(dateISO: string, kind: KnownType, hourShort: string | null, organe?: string | null): string;
+export declare function makeTypeGroupUid(dateISO: string, kind: KnownType, agendaEventId: string, organe?: string | null): string;
 export declare function makeGroupUid(date: string, slot: TimeSlot): string;
 export declare function formatYYYYMMDD(dateYYYYMMDD: string): string;
 export declare function makeReunionUid(agenda: AgendaEvent): string;

package/lib/utils/reunion_grouping.js

@@ -28,56 +28,27 @@ export function groupNonSPByTypeOrganeHour(events) {
     const nonSP = events.filter((e) => !isSeancePublique(e?.type));
     if (nonSP.length === 0)
         return out;
-    const buckets = new Map();
     for (const e of nonSP) {
         const kind = classifyAgendaType(e?.type);
         if (!kind || kind === "SP")
             continue;
         const { startISO, endISO } = deriveTimesForEvent(e);
-        const hourShort = hourShortFromISO(startISO) ?? hourShortFromOriginal(e.timeOriginal);
-        const org = e.organe ?? "NA_ORG";
-        const key = [e.date, kind, org, hourShort || "NA"].join("|");
-        if (!buckets.has(key))
-            buckets.set(key, []);
-        buckets.get(key).push({ ...e, startTime: startISO ?? e.startTime, endTime: endISO ?? e.endTime });
-    }
-    for (const [key, list] of buckets) {
-        const [date, kindStr, organe, hourShort] = key.split("|");
-        const kind = kindStr;
-        const enriched = list
-            .map((ev) => {
-            const { startISO, endISO } = deriveTimesForEvent(ev);
-            return { ev, startISO: startISO ?? ev.startTime, endISO: endISO ?? ev.endTime };
-        })
-            .sort((a, b) => {
-            const ta = a.startISO ? (parseISO(a.startISO)?.toMillis() ?? Number.MAX_SAFE_INTEGER) : Number.MAX_SAFE_INTEGER;
-            const tb = b.startISO ? (parseISO(b.startISO)?.toMillis() ?? Number.MAX_SAFE_INTEGER) : Number.MAX_SAFE_INTEGER;
-            return ta - tb;
-        });
-        const startTime = enriched.find((x) => !!x.startISO)?.startISO ?? null;
-        const endTime = enriched.reduce((acc, x) => {
-            const de = x.endISO ? parseISO(x.endISO)?.toMillis() : null;
-            const accMs = acc ? parseISO(acc)?.toMillis() : null;
-            if (de != null && (accMs == null || de > accMs))
-                return x.endISO;
-            return acc;
-        }, null);
-        const any = enriched[0]?.ev;
-        const hour = hourShort !== "NA" ? hourShort : (hourShortFromISO(startTime) ?? hourShortFromOriginal(any?.timeOriginal));
-        const uid = makeTypeGroupUid(date, kind, hour ?? "", any?.organe ?? "");
+        const startTime = startISO ?? e.startTime ?? null;
+        const endTime = endISO ?? e.endTime ?? null;
+        const uid = makeTypeGroupUid(e.date, kind, e.id, e.organe ?? null);
         const suffix = (kind === "COM" ? "IDC" : kind === "MC" ? "IDM" : kind === "OD" ? "IDO" : "IDI");
         const group = {
             uid,
             chambre: "SN",
-            date,
-            type: any?.type || "",
-            organe: any?.organe || undefined,
+            date: e.date,
+            type: e.type || "",
+            organe: e.organe || undefined,
             startTime,
             endTime,
-            captationVideo: enriched.some((x) => x.ev.captationVideo === true),
-            titre: compactTitleList(enriched.map((x) => x.ev.titre || "").filter(Boolean), 8),
-            objet: joinObjets(enriched.map((x) => x.ev)),
-            events: enriched.map((x) => x.ev),
+            captationVideo: e.captationVideo === true,
+            titre: e.titre,
+            objet: e.objet || "",
+            events: [e],
         };
         out[suffix].push(group);
     }
@@ -270,12 +241,12 @@ function organeInitials(input, maxLen = 8) {
     const out = letters.join("");
     return out.slice(0, maxLen);
 }
-export function makeTypeGroupUid(dateISO, kind, hourShort, organe) {
+export function makeTypeGroupUid(dateISO, kind, agendaEventId, organe) {
     const ymd = dateISO ? formatYYYYMMDD(dateISO) : "00000000";
     const suffix = typeToSuffixStrict(kind);
-    const hh = hourShort ?? "NA";
     const org = organe ? organeInitials(organe) : "";
-    return `RUSN${ymd}${suffix}${org ? org : ""}-${hh}`;
+    let base = `RUSN${ymd}${suffix}${org ? org : ""}${agendaEventId}`;
+    return base;
 }
 function parseISO(isoLike) {
     if (!isoLike)

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@tricoteuses/senat",
-  "version": "2.15.7",
+  "version": "2.16.0",
   "description": "Handle French Sénat's open data",
   "keywords": [
     "France",

package/lib/model/compte_rendu.d.ts

@@ -1,9 +0,0 @@
-import { CompteRendu, Sommaire } from "../types/compte_rendu";
-import { TimeSlot } from "../types/agenda";
-export declare function parseCompteRenduSlotFromFile(xmlFilePath: string, wantedSlot: TimeSlot, firstSlotOfDay?: TimeSlot): Promise<CompteRendu | null>;
-export declare function sessionStartYearFromDate(d: Date): number;
-export declare function parseYYYYMMDD(yyyymmdd: string): Date | null;
-export declare function deriveTitreObjetFromSommaire(sommaire: Sommaire | undefined, slot: TimeSlot): {
-    titre: string;
-    objet: string;
-};

package/lib/model/compte_rendu.js

@@ -1,325 +0,0 @@
-import fs from "fs";
-import * as cheerio from "cheerio";
-import path from "path";
-import { computeIntervalsBySlot } from "../utils/cr_spliting";
-import { norm } from "./util";
-const asArray = (x) => x == null ? [] : Array.isArray(x) ? x : [x];
-const toInt = (s) => Number.isFinite(Number(s)) ? Number(s) : Number.POSITIVE_INFINITY;
-export async function parseCompteRenduSlotFromFile(xmlFilePath, wantedSlot, firstSlotOfDay) {
-    try {
-        const raw = fs.readFileSync(xmlFilePath, "utf8");
-        const $ = cheerio.load(raw, { xml: false });
-        const metadonnees = extractMetadonnees($, xmlFilePath);
-        const order = $("body *").toArray();
-        const idx = new Map(order.map((el, i) => [el, i]));
-        const intervalsAll = computeIntervalsBySlot($, idx, firstSlotOfDay);
-        const intervals = intervalsAll.filter(iv => iv.slot === wantedSlot);
-        if (intervals.length === 0) {
-            console.warn(`[CRI] no intervals for ${path.basename(xmlFilePath)} [${wantedSlot}]`);
-            return null;
-        }
-        metadonnees.sommaire = extractSommaireForIntervals($, idx, intervals);
-        const points = [];
-        let ordre = 0;
-        const addPoint = (p) => points.push({ ...p, ordre_absolu_seance: String(++ordre) });
-        // Titles
-        $("cri\\:titreS1 p.titre_S1").each((_, el) => {
-            if (!elementInAnyInterval(el, idx, intervals))
-                return;
-            const t = normalizeTitle(norm($(el).text() || ""));
-            if (t)
-                addPoint({ code_grammaire: "TITRE_TEXTE_DISCUSSION", texte: { _: t }, code_style: "Titre" });
-        });
-        // Interventions
-        $("div.intervenant").each((_, block) => {
-            if (!elementInAnyInterval(block, idx, intervals))
-                return;
-            const $block = $(block);
-            $block.find([
-                "p[class^='titre_S']",
-                "p.mention_titre",
-                "p.intitule_titre",
-                "p.mention_chapitre",
-                "p.intitule_chapitre",
-                "p.mention_article",
-                "p.intitule_article",
-                "p.mention_section",
-                "p.intitule_section",
-            ].join(",")).remove();
-            const firstP = $block.find("p").first();
-            const speakerLabelRaw = firstP.find(".orateur_nom").text() || firstP.find("a.lien_senfic").text() || "";
-            const speakerLabel = dedupeSpeaker(speakerLabelRaw);
-            const { mat, nom: nomCRI, qua: quaCRI } = readIntervenantMeta($block);
-            const qualFromSpans = extractAndRemoveLeadingQualite($, $block);
-            const qualite = norm(decodeHtmlEntities(quaCRI || "")) || qualFromSpans;
-            const canonicalName = dedupeSpeaker(nomCRI || speakerLabel);
-            const role = roleForSpeaker(speakerLabel) || roleForSpeaker(qualite) || roleForSpeaker(quaCRI || "");
-            const speechHtml = sanitizeInterventionHtml($, $block);
-            if (!norm(cheerio.load(speechHtml).text() || ""))
-                return;
-            addPoint({
-                code_grammaire: "PAROLE_GENERIQUE",
-                roledebat: role,
-                orateurs: { orateur: { nom: canonicalName, id: mat || "", qualite } },
-                texte: { _: speechHtml },
-            });
-        });
-        const contenu = {
-            quantiemes: { journee: metadonnees.dateSeance, session: metadonnees.session },
-            point: points,
-        };
-        return {
-            uid: "CRSSN" + xmlFilePath.replace(/^.*?(\d{8}).*$/i, "$1") + `-${wantedSlot}`,
-            seanceRef: "RUSN" + xmlFilePath.replace(/^.*?(\d{8}).*$/i, "$1") + "IDS-" + wantedSlot,
-            sessionRef: metadonnees.session,
-            metadonnees,
-            contenu,
-        };
-    }
-    catch (e) {
-        console.error(`[CRI] parseSlot error file=${xmlFilePath} slot=${wantedSlot}:`, e);
-        return null;
-    }
-}
-export function sessionStartYearFromDate(d) {
-    // Session (1th oct N → 30 sept N+1)
-    const m = d.getMonth();
-    const y = d.getFullYear();
-    return m >= 9 ? y : y - 1;
-}
-export function parseYYYYMMDD(yyyymmdd) {
-    if (!/^\d{8}$/.test(yyyymmdd))
-        return null;
-    const y = Number(yyyymmdd.slice(0, 4));
-    const m = Number(yyyymmdd.slice(4, 6)) - 1;
-    const d = Number(yyyymmdd.slice(6, 8));
-    const dt = new Date(y, m, d);
-    return Number.isFinite(dt.getTime()) ? dt : null;
-}
-export function deriveTitreObjetFromSommaire(sommaire, slot) {
-    const items = extractLevel1Items(sommaire);
-    const meaningful = items.filter(it => !isBoilerplate(it.label));
-    if (meaningful.length === 0) {
-        return {
-            titre: `Séance publique ${slotLabel(slot)}`,
-            objet: "",
-        };
-    }
-    const titre = meaningful[0].label;
-    const objet = meaningful.slice(0, 3).map(it => it.label).join(" ; ");
-    return { titre, objet };
-}
-function slotLabel(slot) {
-    switch (slot) {
-        case "MATIN": return "du matin";
-        case "APRES-MIDI": return "de l’après-midi";
-        case "SOIR": return "du soir";
-        default: return "";
-    }
-}
-const BOILERPLATE_PATTERNS = [
-    /proc(?:è|e)s-?verbal/i,
-    /hommages?/i,
-    /désignation des vice-?président/i,
-    /candidatures? aux?/i,
-    /ordre du jour/i,
-    /rappels? au règlement/i,
-    /communications?/i,
-    /dépôts?/i,
-    /proclamation/i,
-    /présidence de/i,
-    /questions? diverses?/i,
-    /ouverture de la séance/i,
-    /clo(?:t|̂)ure de la séance/i,
-];
-const isBoilerplate = (label) => !label?.trim() || BOILERPLATE_PATTERNS.some(rx => rx.test(label));
-function extractLevel1Items(sommaire) {
-    const level1 = asArray(sommaire?.sommaire1);
-    return level1
-        .map(el => ({
-        numero: toInt(el?.valeur_pts_odj),
-        label: String(el?.titreStruct?.intitule ?? "").trim(),
-    }))
-        .filter(it => !!it.label)
-        .sort((a, b) => a.numero - b.numero);
-}
-function stripTrailingPunct(s) { return s.replace(/\s*([:,.;])\s*$/u, "").trim(); }
-function dedupeSpeaker(raw) {
-    let s = norm(raw);
-    s = stripTrailingPunct(s);
-    const dupPatterns = [/^(.+?)\s*[.]\s*\1$/u, /^(.+?)\s*,\s*\1,?$/u, /^(.+?)\s+\1$/u];
-    for (const re of dupPatterns) {
-        const m = s.match(re);
-        if (m) {
-            s = m[1];
-            break;
-        }
-    }
-    return s.replace(/\.\s*$/, "");
-}
-function decodeHtmlEntities(s) {
-    return s.replace(/&#(\d+);/g, (_, d) => String.fromCharCode(parseInt(d, 10)))
-        .replace(/&#x([0-9a-fA-F]+);/g, (_, h) => String.fromCharCode(parseInt(h, 16)));
-}
-function fixApostrophes(s) {
-    let out = s;
-    out = out.replace(/\s*’\s*/g, "’");
-    out = out.replace(/\b([dljctmsn])\s*’/gi, (_, m) => m + "’");
-    out = out.replace(/’\s+([A-Za-zÀ-ÖØ-öø-ÿ])/g, "’$1");
-    out = out.replace(/\s+([,;:.!?])/g, "$1");
-    return out;
-}
-function normalizeTitle(text) { return text.replace(/^PR[ÉE]SIDENCE DE\b/i, "Présidence de "); }
-function roleForSpeaker(labelOrQualite) {
-    const s = (labelOrQualite || "").toLowerCase();
-    if (/^(m\.|mme)?\s*(le|la)\s+pr[ée]sident(e)?\b/.test(s) || /\bpr[ée]sident[e]?\s+de\s+séance\b/.test(s))
-        return "président";
-    return "";
-}
-function readIntervenantMeta($block) {
-    const int = $block.find('cri\\:intervenant').first();
-    if (int.length)
-        return { mat: int.attr("mat") || undefined, nom: int.attr("nom") || undefined, qua: int.attr("qua") || undefined };
-    const html = $block.html() || "";
-    const m = html.match(/<!--\s*cri:intervenant\b([^>]+)-->/i);
-    if (!m)
-        return {};
-    const out = {};
-    const re = /(\w+)="([^"]*)"/g;
-    let a;
-    while ((a = re.exec(m[1])))
-        out[a[1]] = decodeHtmlEntities(a[2]);
-    return { mat: out["mat"], nom: out["nom"], qua: out["qua"] };
-}
-function extractAndRemoveLeadingQualite($, $block) {
-    const firstP = $block.find("p").first();
-    if (firstP.length === 0)
-        return "";
-    const parts = [];
-    let stop = false;
-    firstP.contents().each((_, node) => {
-        if (stop)
-            return;
-        if (node.type === "tag") {
-            const $node = $(node);
-            if ($node.hasClass("orateur_nom")) {
-                $node.remove();
-                return;
-            }
-            if ($node.hasClass("orateur_qualite")) {
-                parts.push($node.text() || "");
-                $node.remove();
-                return;
-            }
-            const t = norm($node.text() || "");
-            if (t)
-                stop = true;
-            else
-                $node.remove();
-        }
-        else if (node.type === "text") {
-            const t = norm(node.data || "");
-            if (!t || /^[:.,;–—-]+$/.test(t)) {
-                node.data = "";
-                return;
-            }
-            stop = true;
-        }
-    });
-    return fixApostrophes(norm(parts.join(" ")));
-}
-function sanitizeInterventionHtml($, $block) {
-    const $clone = $block.clone();
-    $clone.find('a[name]').remove();
-    $clone.find('div[align="right"]').remove();
-    $clone.find('a.link').remove();
-    $clone.find('img').remove();
-    $clone.find('a#ameli_amendement_cri_phrase, a#ameli_amendement_cra_contenu, a#ameli_amendement_cra_objet').remove();
-    $clone.find(".orateur_nom, .orateur_qualite").remove();
-    let html = $clone.html() || "";
-    html = html.replace(/<!--[\s\S]*?-->/g, "");
-    return html.trim();
-}
-function extractSommaireForIntervals($, idx, intervals) {
-    const inIv = (el) => elementInAnyInterval(el, idx, intervals);
-    const root = $("body");
-    const sommaire = { presidentSeance: { _: "" }, sommaire1: [] };
-    // (1) Présidence (tm2) — première ligne dans l’intervalle 
-    const pres = root.find("p.tm2").filter((_, el) => inIv(el)).first();
-    if (pres.length)
-        sommaire.presidentSeance = { _: norm(pres.text()) };
-    // (2) Paras tm5 présents dans l’intervalle 
-    const paras = [];
-    root.find("p.tm5").each((_, el) => {
-        if (!inIv(el))
-            return;
-        const t = norm($(el).text());
-        if (t)
-            paras.push({ _: t });
-    });
-    if (paras.length)
-        sommaire.para = paras.length === 1 ? paras[0] : paras;
-    // (3) Items de 1er niveau (tm3) présents dans l’intervalle 
-    const items = [];
-    root.find("p.tm3").each((_, el) => {
-        if (!inIv(el))
-            return;
-        const $p = $(el);
-        const full = norm($p.text() || "");
-        if (!full)
-            return;
-        const numMatch = full.match(/^(\d+)\s*[.\-–—]\s*/);
-        const valeur = numMatch ? numMatch[1] : undefined;
-        // prefere intitule in ancre <a> if present
-        const a = $p.find("a").first();
-        const intituleRaw = a.length ? a.text() : full.replace(/^(\d+)\s*[.\-–—]\s*/, "");
-        const intitule = norm(intituleRaw);
-        // id_syceron from href="#Niv1_SOMx"
-        const href = (a.attr("href") || "").trim();
-        const idSyceron = href.startsWith("#") ? href.slice(1) : href;
-        const titreStruct = { id_syceron: idSyceron || "", intitule };
-        items.push({ valeur_pts_odj: valeur, titreStruct });
-    });
-    if (items.length)
-        sommaire.sommaire1 = items;
-    return sommaire;
-}
-function extractMetadonnees($, filePath) {
-    let dateText = norm($("h1, h2, .page-title").first().text() || "");
-    if (!dateText)
-        dateText = norm($("p").first().text() || "");
-    const dateMatch = dateText.match(/\b(\d{1,2}\s+\w+\s+\d{4})\b/i);
-    const allText = norm($("body").text() || "");
-    const sessionMatch = allText.match(/\bsession\s+(\d{4}-\d{4})\b/i);
-    let dateSeance = dateMatch?.[1] || "";
-    if (!dateSeance) {
-        const m = filePath.match(/d(\d{4})(\d{2})(\d{2})\.xml$/i);
-        if (m)
-            dateSeance = `${m[1]}-${m[2]}-${m[3]}`;
-    }
-    return {
-        dateSeance,
-        dateSeanceJour: dateSeance,
-        numSeanceJour: "",
-        numSeance: "",
-        typeAssemblee: "SN",
-        legislature: "",
-        session: sessionMatch?.[1] || "",
-        nomFichierJo: "",
-        validite: "",
-        etat: "",
-        diffusion: "",
-        version: "1.0",
-        environnement: "",
-        heureGeneration: new Date()
-    };
-}
-function elementInAnyInterval(el, idx, intervals) {
-    const p = idx.get(el);
-    if (p == null)
-        return false;
-    for (const iv of intervals)
-        if (p >= iv.start && p < iv.end)
-            return true;
-    return false;
-}