@tricoteuses/senat 2.22.13 → 2.22.14

package/lib/utils/cr_spliting.js

@@ -1,265 +0,0 @@
-import path from "path";
-import * as cheerio from "cheerio";
-import { AGENDA_FOLDER, DATA_TRANSFORMED_FOLDER } from "../loaders";
-import fs from "fs-extra";
-import { sessionStartYearFromDate } from "../model/seance";
-import { frDateToISO, hourShortToStartTime } from "./date";
-import { normalizeSpaces } from "./string_cleaning";
-// Convert "quinze heures trente", "15 heures 30", "dix-sept heures moins le quart", etc. en "HHMM"
-function parseFrenchClockToHHMM(input) {
-    const s = (input || "")
-        .toLowerCase()
-        .normalize("NFKD")
-        .replace(/[\u0300-\u036f]/g, "")
-        .trim();
-    if (!s)
-        return undefined;
-    const digitMatch = s.match(/(\d{1,2})\s*heures?(?:\s*(\d{1,2}))?/);
-    if (digitMatch) {
-        const h = Math.min(24, Math.max(0, parseInt(digitMatch[1], 10)));
-        const m = digitMatch[2] ? Math.min(59, Math.max(0, parseInt(digitMatch[2], 10))) : 0;
-        return `${String(h).padStart(2, "0")}${String(m).padStart(2, "0")}`;
-    }
-    const NUM = new Map([
-        ["zero", 0],
-        ["une", 1],
-        ["un", 1],
-        ["deux", 2],
-        ["trois", 3],
-        ["quatre", 4],
-        ["cinq", 5],
-        ["six", 6],
-        ["sept", 7],
-        ["huit", 8],
-        ["neuf", 9],
-        ["dix", 10],
-        ["onze", 11],
-        ["douze", 12],
-        ["treize", 13],
-        ["quatorze", 14],
-        ["quinze", 15],
-        ["seize", 16],
-        ["dix-sept", 17],
-        ["dix sept", 17],
-        ["dix-huit", 18],
-        ["dix huit", 18],
-        ["dix-neuf", 19],
-        ["dix neuf", 19],
-        ["vingt", 20],
-        ["vingt et une", 21],
-        ["vingt-et-une", 21],
-        ["vingt et un", 21],
-        ["vingt-et-un", 21],
-        ["vingt-deux", 22],
-        ["vingt deux", 22],
-        ["vingt-trois", 23],
-        ["vingt trois", 23],
-        ["vingt-quatre", 24],
-        ["vingt quatre", 24],
-    ]);
-    const hourWordMatch = s.match(/([a-z\- ]+?)\s*heures?/);
-    if (!hourWordMatch)
-        return undefined;
-    const hourWord = hourWordMatch[1].trim();
-    let hour = NUM.get(hourWord);
-    if (hour == null) {
-        const cleaned = hourWord.replace(/\s+/g, " ");
-        hour = NUM.get(cleaned);
-    }
-    if (hour == null)
-        return undefined;
-    let minutes = 0;
-    if (/\bet (demie|demi)\b/.test(s))
-        minutes = 30;
-    else if (/\bet quart\b/.test(s))
-        minutes = 15;
-    else if (/\bmoins le quart\b/.test(s)) {
-        hour = (hour + 23) % 24;
-        minutes = 45;
-    }
-    else {
-        const MIN = new Map([
-            ["cinq", 5],
-            ["dix", 10],
-            ["quinze", 15],
-            ["vingt", 20],
-            ["vingt-cinq", 25],
-            ["vingt cinq", 25],
-            ["trente", 30],
-            ["trente-cinq", 35],
-            ["trente cinq", 35],
-            ["quarante", 40],
-            ["quarante-cinq", 45],
-            ["quarante cinq", 45],
-            ["cinquante", 50],
-            ["cinquante-cinq", 55],
-            ["cinquante cinq", 55],
-        ]);
-        const minWordMatch = s.match(/heures?\s+([a-z\- ]+?)(?:[).,;]|$)/);
-        if (minWordMatch) {
-            const mw = minWordMatch[1].trim();
-            const m1 = MIN.get(mw);
-            if (m1 != null)
-                minutes = m1;
-        }
-    }
-    return `${String(hour).padStart(2, "0")}${String(minutes).padStart(2, "0")}`;
-}
-function extractWeekStartFromHead($) {
-    const og = $('meta[property="og:title"]').attr("content") || $("title").text();
-    const m = (og ?? "").toLowerCase().match(/semaine du\s+(\d{1,2}\s+\w+\s+\d{4})/i);
-    if (m)
-        return frDateToISO(m[1]);
-    return undefined;
-}
-function detectOrganeFromTitle(s) {
-    const t = (s ?? "").trim();
-    if (!t)
-        return { organeTitleRaw: undefined, organeDetected: undefined };
-    const lower = t.toLowerCase();
-    const m = lower.match(/commission(?:\s+des|\s+de|)\s+([^:]+)$/i);
-    let organeDetected;
-    if (m && m[1]) {
-        organeDetected = ("Commission " + m[1])
-            .replace(/\s+/g, " ")
-            .replace(/\s+:? comptes? rendus?$/i, "")
-            .trim();
-        organeDetected = organeDetected[0].toUpperCase() + organeDetected.slice(1);
-    }
-    return { organeTitleRaw: t, organeDetected };
-}
-function extractDaysAndOpenings($) {
-    const days = [];
-    const h2s = $("h2").toArray();
-    for (let i = 0; i < h2s.length; i++) {
-        const h = h2s[i];
-        const txt = normalizeSpaces($(h).text());
-        const m = txt.match(/^(?:Lundi|Mardi|Mercredi|Jeudi|Vendredi|Samedi|Dimanche)\s+(.+?)$/i);
-        if (!m)
-            continue;
-        const iso = frDateToISO(m[1]);
-        if (!iso)
-            continue;
-        let openTime;
-        let cur = $(h).next();
-        while (cur.length && cur[0].tagName !== "h2") {
-            const t = normalizeSpaces(cur.text());
-            const mt = t.match(/La réunion est ouverte à\s+(\d{1,2})\s*h(?:\s*(\d{2}))?/i);
-            if (mt) {
-                openTime = `${mt[1].padStart(2, "0")}:${(mt[2] ?? "00").padStart(2, "0")}`;
-                break;
-            }
-            cur = cur.next();
-        }
-        days.push({ date: iso, openTime, h2Index: i });
-    }
-    return days;
-}
-function extractOrganeCode($) {
-    const names = $("a[name]")
-        .toArray()
-        .map((a) => ($(a).attr("name") || "").trim());
-    return names.find((n) => /^[A-Z]{3,6}$/.test(n));
-}
-export function parseCommissionMetadataFromHtml(html, sourceFileName) {
-    const $ = cheerio.load(html);
-    const h1 = $("h1.page-title").first().text().trim() || undefined;
-    const headTitle = $('meta[property="og:title"]').attr("content") || $("title").text() || undefined;
-    const { organeTitleRaw, organeDetected } = detectOrganeFromTitle(h1 || headTitle);
-    let weekStart = extractWeekStartFromHead($);
-    const days = extractDaysAndOpenings($);
-    if (!weekStart && days.length > 0)
-        weekStart = days[0].date;
-    const organeCode = extractOrganeCode($);
-    return {
-        sourceFile: sourceFileName ?? null,
-        organeTitleRaw: organeTitleRaw ?? null,
-        organeDetected: organeDetected ?? null,
-        organeCode: organeCode ?? null,
-        weekStart: weekStart ?? null,
-        days, // [{date, openTime?, h2Index}]
-    };
-}
-function isGroupedReunion(o) {
-    return o && typeof o === "object" && typeof o.uid === "string" && typeof o.date === "string";
-}
-export async function loadAgendaForDate(dataDir, yyyymmdd, session) {
-    const baseDir = path.join(dataDir, AGENDA_FOLDER, DATA_TRANSFORMED_FOLDER, String(session));
-    if (!(await fs.pathExists(baseDir)))
-        return [];
-    const files = (await fs.readdir(baseDir)).filter((f) => f.startsWith(`RUSN${yyyymmdd}IDC`) && f.toLowerCase().endsWith(".json"));
-    const out = [];
-    for (const f of files) {
-        const p = path.join(baseDir, f);
-        try {
-            const raw = await fs.readFile(p, "utf8");
-            const obj = JSON.parse(raw);
-            if (!isGroupedReunion(obj)) {
-                continue;
-            }
-            if (!obj.uid.startsWith(`RUSN${yyyymmdd}IDC`)) {
-                continue;
-            }
-            out.push(obj);
-        }
-        catch {
-            // ignore
-        }
-    }
-    return out;
-}
-export async function linkCRtoCommissionGroup(opts) {
-    const { dataDir, dateISO, organeDetected, hourShort, crUid, titreGuess, groupUid } = opts;
-    const computedUid = crUid.replace(/^CRC/, "RU");
-    const uid = groupUid ?? computedUid;
-    const session = sessionStartYearFromDate(new Date(dateISO));
-    const groupedDir = path.join(dataDir, AGENDA_FOLDER, DATA_TRANSFORMED_FOLDER, String(session));
-    await fs.ensureDir(groupedDir);
-    const filePath = path.join(groupedDir, `${uid}.json`);
-    let group = null;
-    let created = false;
-    let updated = false;
-    try {
-        if (await fs.pathExists(filePath)) {
-            group = await fs.readJSON(filePath);
-        }
-    }
-    catch (e) {
-        console.warn(`[AGENDA][COM] Unreadable JSON → ${filePath} (${e?.message}) → will recreate`);
-    }
-    if (!group) {
-        group = {
-            uid,
-            chambre: "SN",
-            date: dateISO,
-            type: "Commission",
-            startTime: hourShortToStartTime(hourShort),
-            endTime: null,
-            captationVideo: false,
-            titre: titreGuess ?? "",
-            organe: organeDetected ?? "Commission",
-            objet: titreGuess ?? "",
-            events: [],
-            compteRenduRefUid: crUid,
-        };
-        created = true;
-        console.log(`[AGENDA][COM] Created new group uid=${uid} for CR uid=${crUid}`);
-    }
-    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
-    // const sommaire = cr?.metadonnees?.sommaire as Sommaire | undefined;
-    // if (sommaire) {
-    //   const { titre: dTitre, objet: dObjet } = deriveTitreObjetFromSommaire(sommaire, undefined);
-    //   if (!group.titre && dTitre) group.titre = dTitre;
-    //   if ((!group.objet || !group.objet.trim()) && dObjet) group.objet = dObjet;
-    // } else if (!group.titre && titreGuess) {
-    //   group.titre = titreGuess;
-    // }
-    await fs.writeJSON(filePath, group, { spaces: 2 });
-    return { uid, filePath, created, updated };
-}

package/lib/utils/date.d.ts

@@ -1,10 +0,0 @@
-export declare function yyyymmddFromPath(xmlFilePath: string): string;
-export declare function parseYYYYMMDD(yyyymmdd: string): Date | null;
-export declare function frDateToISO(s?: string): string | undefined;
-export declare function hourShortToStartTime(hourShort: string | null): string | null;
-export declare function epochToParisDateTime(epochSec: number): {
-    date: string;
-    startTime: string;
-} | null;
-export declare function toTargetEpoch(time: string | null, date?: string | null): number | null;
-export declare function toFRDate(dateYYYYMMDD: string): string;

package/lib/utils/date.js

@@ -1,100 +0,0 @@
-import { DateTime } from "luxon";
-export function yyyymmddFromPath(xmlFilePath) {
-    return xmlFilePath.replace(/^.*?(\d{8}).*$/i, "$1");
-}
-export function parseYYYYMMDD(yyyymmdd) {
-    if (!/^\d{8}$/.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 frDateToISO(s) {
-    if (!s)
-        return;
-    const months = {
-        janvier: 1,
-        février: 2,
-        fevrier: 2,
-        mars: 3,
-        avril: 4,
-        mai: 5,
-        juin: 6,
-        juillet: 7,
-        août: 8,
-        aout: 8,
-        septembre: 9,
-        octobre: 10,
-        novembre: 11,
-        décembre: 12,
-        decembre: 12,
-    };
-    const cleaned = s
-        .trim()
-        .replace(/\u00A0/g, " ")
-        .replace(/ +/g, " ");
-    const m = cleaned.match(/^(\d{1,2})(?:er)?\s+([a-zéèêîïôûùç]+)\s+(\d{4})$/i);
-    if (!m)
-        return;
-    const d = String(parseInt(m[1], 10)).padStart(2, "0");
-    const mon = months[m[2].toLowerCase()];
-    if (!mon)
-        return;
-    const y = m[3];
-    return `${y}-${String(mon).padStart(2, "0")}-${d}`;
-}
-export function hourShortToStartTime(hourShort) {
-    if (!hourShort || hourShort === "NA")
-        return null;
-    if (!/^\d{4}$/.test(hourShort))
-        return null;
-    const hh = hourShort.slice(0, 2);
-    const mm = hourShort.slice(2, 4);
-    return `${hh}:${mm}`;
-}
-export function epochToParisDateTime(epochSec) {
-    if (!Number.isFinite(epochSec))
-        return null;
-    const dUtc = new Date(epochSec * 1000);
-    // Offset heuristic (same logique que parisOffsetForDate)
-    const m = dUtc.getUTCMonth() + 1; // 1..12
-    const offsetHours = m >= 4 && m <= 10 ? 2 : 1;
-    const offsetStr = offsetHours === 2 ? "+02:00" : "+01:00";
-    // Applique l'offset pour obtenir la date/heure locales Paris
-    const localMs = dUtc.getTime() + offsetHours * 3600 * 1000;
-    const dl = new Date(localMs);
-    const yyyy = String(dl.getUTCFullYear());
-    const mm = String(dl.getUTCMonth() + 1).padStart(2, "0");
-    const dd = String(dl.getUTCDate()).padStart(2, "0");
-    const hh = String(dl.getUTCHours()).padStart(2, "0");
-    const mi = String(dl.getUTCMinutes()).padStart(2, "0");
-    const ss = String(dl.getUTCSeconds()).padStart(2, "0");
-    const ms = String(dl.getUTCMilliseconds()).padStart(3, "0");
-    return {
-        date: `${yyyy}-${mm}-${dd}`,
-        startTime: `${hh}:${mi}:${ss}.${ms}${offsetStr}`,
-    };
-}
-export function toTargetEpoch(time, date) {
-    if (!time)
-        return null;
-    let dtLocal;
-    if (time.includes("T")) {
-        dtLocal = DateTime.fromISO(time, { zone: "Europe/Paris" });
-    }
-    else if (date) {
-        dtLocal = DateTime.fromISO(`${date}T${time}`, { zone: "Europe/Paris" });
-    }
-    else {
-        return null;
-    }
-    if (!dtLocal.isValid)
-        return null;
-    return Math.floor(dtLocal.toUTC().toSeconds());
-}
-export function toFRDate(dateYYYYMMDD) {
-    const [y, m, d] = dateYYYYMMDD.split("-");
-    return `${d}/${m}/${y}`; // DD/MM/YYYY
-}

package/lib/utils/nvs-timecode.d.ts

@@ -1,7 +0,0 @@
-export declare function getAgendaSegmentTimecodes(dataNvs: string, finalPlayerNvs: string, agendaTitleOrObjet: string): {
-    start: number;
-    end: number | null;
-    chapterId: string;
-    nextChapterId: string | null;
-    score: number;
-} | null;

package/lib/utils/nvs-timecode.js

@@ -1,79 +0,0 @@
-import { XMLParser } from "fast-xml-parser";
-import { dice, normalize } from "./scoring";
-import { decodeHtmlEntities } from "./string_cleaning";
-const CHAPTER_MATCH_THRESHOLD = 0.5;
-const xmlParser = new XMLParser({
-    ignoreAttributes: false,
-    attributeNamePrefix: "@_",
-});
-function getTimecodeForChapterId(finalPlayerNvs, chapterId) {
-    const xml = xmlParser.parse(finalPlayerNvs);
-    const synchros = xml?.player?.synchro;
-    if (!synchros)
-        return null;
-    const synchsArray = Array.isArray(synchros) ? synchros : [synchros];
-    const match = synchsArray.find((s) => String(s["@_id"]) === String(chapterId));
-    if (!match)
-        return null;
-    const rawTimecode = match["@_timecode"];
-    if (rawTimecode == null)
-        return null;
-    const ms = Number(rawTimecode);
-    if (Number.isNaN(ms))
-        return null;
-    return Math.floor(ms / 1000);
-}
-function toArray(v) {
-    if (!v)
-        return [];
-    return Array.isArray(v) ? v : [v];
-}
-function getLevel1Chapters(dataNvs) {
-    const xml = xmlParser.parse(dataNvs);
-    const root = xml?.data?.chapters?.chapter ?? xml?.chapters?.chapter;
-    const roots = toArray(root);
-    return roots
-        .map((ch, i) => {
-        const id = ch?.id ?? ch?.["@_id"];
-        const labelRaw = ch?.label ?? ch?.["@_label"] ?? "";
-        return {
-            id: String(id),
-            label: decodeHtmlEntities(String(labelRaw)).trim(),
-            index: i,
-        };
-    })
-        .filter((c) => c.id && c.label);
-}
-function pickBestLevel1ChapterForAgenda(chapters, agendaTitle) {
-    const q = normalize(agendaTitle);
-    let best = null;
-    for (const ch of chapters) {
-        const s = dice(q, ch.label);
-        if (!best || s > best.score)
-            best = { chapter: ch, score: s };
-    }
-    if (!best || best.score < CHAPTER_MATCH_THRESHOLD)
-        return { chapter: chapters[0], score: 0 };
-    return best;
-}
-export function getAgendaSegmentTimecodes(dataNvs, finalPlayerNvs, agendaTitleOrObjet) {
-    const l1 = getLevel1Chapters(dataNvs);
-    if (!l1.length)
-        return null;
-    const best = pickBestLevel1ChapterForAgenda(l1, agendaTitleOrObjet);
-    if (!best)
-        return null;
-    const chapter = best.chapter;
-    const next = l1[chapter.index + 1] ?? null;
-    const start = getTimecodeForChapterId(finalPlayerNvs, chapter.id);
-    if (start == null)
-        return null;
-    const end = next ? getTimecodeForChapterId(finalPlayerNvs, next.id) : null;
-    return {
-        start,
-        end,
-        chapterId: chapter.id,
-        nextChapterId: next?.id ?? null,
-        score: best.score,
-    };
-}

package/lib/utils/reunion_grouping.d.ts

@@ -1,11 +0,0 @@
-import { AgendaEvent, GroupedReunion, TimeSlot } from "../types/agenda";
-import { DossierLegislatifResult } from "../model/dosleg";
-type KnownType = "SP" | "COM" | "MC" | "OD" | "ID";
-type DossierBySenatUrl = Record<string, DossierLegislatifResult>;
-export declare function groupNonSPByTypeOrganeHour(events: AgendaEvent[], DossierBySenatUrl: DossierBySenatUrl): Record<"IDC" | "IDM" | "IDO" | "IDI", GroupedReunion[]>;
-export declare function groupSeancePubliqueBySlot(events: AgendaEvent[], dossierBySenatUrl: DossierBySenatUrl): Record<TimeSlot, GroupedReunion[]>;
-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;
-export {};