@tricoteuses/senat 2.22.3 → 2.22.5

package/lib/model/texte.js

@@ -0,0 +1,256 @@
+import { JSDOM } from "jsdom";
+import { AKN_IDENTIFICATION_STRUCTURE_REGEXP, AKN_WORKFLOW_IDENTIFICATION_STRUCTURE_REGEXP } from "../scripts/datautil";
+import { DivisionType, } from "../types/texte";
+function buildWorklow(metaElement) {
+    const stepElements = metaElement.querySelectorAll("workflow step");
+    const steps = [];
+    for (const stepElement of stepElements) {
+        const identification = stepElement.getAttribute("href") ?? "";
+        const identificationParts = AKN_WORKFLOW_IDENTIFICATION_STRUCTURE_REGEXP.exec(identification)?.groups;
+        steps.push({
+            eId: stepElement.getAttribute("eId"),
+            date: stepElement.getAttribute("date")
+                ? new Date(stepElement.getAttribute("date") ?? "")
+                : null,
+            type: identificationParts?.["type"] || null,
+            session: identificationParts?.["session"] || null,
+            numero: identificationParts?.["numTexte"] || null,
+            version: identificationParts?.["version"]
+                ? identificationParts["version"]
+                : null,
+            outcome: stepElement.getAttribute("outcome"),
+        });
+    }
+    return steps;
+}
+function buildDivision(node, index) {
+    const eId = node.getAttribute("eId");
+    const tag = node.nodeName;
+    const level = DivisionType[tag];
+    const titleNode = node.querySelector("num");
+    const subtitleNode = node.querySelector("heading");
+    const headings = [
+        ...(titleNode
+            ? [
+                {
+                    text: titleNode.textContent?.trim() ?? null,
+                    html: titleNode.innerHTML?.trim() ?? null,
+                },
+            ]
+            : []),
+        ...(subtitleNode
+            ? [
+                {
+                    text: subtitleNode.textContent?.trim() ?? null,
+                    html: subtitleNode.innerHTML?.trim() ?? null,
+                },
+            ]
+            : []),
+    ];
+    const division = {
+        index,
+        eId,
+        tag,
+        level,
+        headings,
+    };
+    if (tag === "article") {
+        ;
+        division.alineas = [];
+    }
+    return division;
+}
+function buildAlinea(contentNode, alineaNode) {
+    const eId = alineaNode.getAttribute("eId");
+    const heading = {
+        text: alineaNode.querySelector("num")?.textContent ?? null,
+    };
+    const pastille = alineaNode.getAttribute("data:pastille") ?? null;
+    return {
+        eId,
+        heading,
+        text: contentNode.textContent?.trim() ?? null,
+        html: contentNode.innerHTML?.trim() ?? null,
+        pastille,
+    };
+}
+function buildEmptyArticle(index) {
+    return {
+        index: index,
+        eId: "",
+        tag: "article",
+        level: DivisionType["article"],
+        headings: [],
+        alineas: [],
+    };
+}
+function flattenTexte(texteContentRoot) {
+    const divisions = [];
+    let divisionIndex = 0;
+    const iter = (node) => {
+        if (node.nodeName === "content") {
+            return;
+        }
+        switch (node.nodeName) {
+            case "tome":
+            case "part":
+            case "book":
+            case "title":
+            case "subtitle":
+            case "chapter":
+            case "section":
+            case "subsection":
+            case "paragraph":
+            case "article":
+                divisions.push(buildDivision(node, divisionIndex++));
+                break;
+        }
+        if (node.nodeName === "alinea") {
+            Array.from(node.childNodes)
+                // Find direct content children programmatically
+                // because `:scope` selector does not work
+                // https://github.com/jsdom/jsdom/issues/2998
+                .filter((alineaChildNode) => alineaChildNode.nodeName === "content")
+                .forEach((alineaContentNode) => {
+                // Hypothesis: alineas should always be enclosed in articles
+                let lastArticle = divisions.findLast((division) => division.tag === "article");
+                if (!lastArticle) {
+                    lastArticle = buildEmptyArticle(divisionIndex++);
+                    divisions.push(lastArticle);
+                }
+                lastArticle.alineas.push(buildAlinea(alineaContentNode, node));
+            });
+        }
+        if (node.hasChildNodes()) {
+            node.childNodes.forEach((childNode) => iter(childNode));
+        }
+    };
+    iter(texteContentRoot);
+    return divisions;
+}
+export function transformTexte(document) {
+    const metaElement = document.querySelector("meta");
+    const preambleElement = document.querySelector("preamble");
+    const identification = metaElement
+        ?.querySelector("FRBRExpression FRBRuri")
+        ?.getAttribute("value") ?? "";
+    const identificationParts = AKN_IDENTIFICATION_STRUCTURE_REGEXP.exec(identification)?.groups;
+    const bodyElement = document.querySelector("body");
+    const sessionYears = identificationParts?.["session"]?.split("-") || null;
+    const datePresentation = metaElement
+        ?.querySelector("FRBRdate[name='#presentation']")
+        ?.getAttribute("date");
+    const dateDepot = metaElement
+        ?.querySelector("FRBRdate[name='#depot']")
+        ?.getAttribute("date");
+    const datePublicationXml = metaElement
+        ?.querySelector("FRBRdate[name='#publication-xml']")
+        ?.getAttribute("date");
+    return {
+        titre: preambleElement?.querySelector("docTitle")?.textContent || null,
+        titreCourt: metaElement
+            ?.querySelector("FRBRalias[name='intitule-court']")
+            ?.getAttribute("value") || null,
+        signetDossier: metaElement
+            ?.querySelector("FRBRalias[name='signet-dossier-legislatif-senat']")
+            ?.getAttribute("value") || null,
+        urlDossierSenat: metaElement
+            ?.querySelector("FRBRalias[name='url-senat']")
+            ?.getAttribute("value") || null,
+        urlDossierAssemblee: metaElement
+            ?.querySelector("FRBRalias[name='url-AN']")
+            ?.getAttribute("value") || null,
+        type: identificationParts?.["type"] || null,
+        session: sessionYears && sessionYears.length > 0 ? sessionYears[0] : null,
+        numero: identificationParts?.["numTexte"]
+            ? parseInt(identificationParts["numTexte"])
+            : null,
+        datePresentation: datePresentation ? new Date(datePresentation) : null,
+        dateDepot: dateDepot ? new Date(dateDepot) : null,
+        datePublicationXml: datePublicationXml
+            ? new Date(datePublicationXml)
+            : null,
+        version: identificationParts?.["version"]
+            ? identificationParts["version"]
+            : null,
+        workflow: metaElement ? buildWorklow(metaElement) : [],
+        divisions: bodyElement ? flattenTexte(bodyElement) : [],
+    };
+}
+export function transformExposeDesMotifs(document) {
+    const sectionElements = document.querySelectorAll("section");
+    const exposeDesMotifsRegexp = new RegExp("EXPOS.{1,2}[\\n\\s]DES[\\n\\s]MOTIFS");
+    for (const sectionElement of sectionElements) {
+        const firstParagraph = sectionElement.querySelector("p:first-of-type");
+        const secondParagraph = sectionElement.querySelector("p:nth-of-type(2)");
+        if (!firstParagraph) {
+            continue;
+        }
+        const firstParagraphContent = firstParagraph.textContent;
+        const secondParagraphContent = secondParagraph?.textContent;
+        if (!firstParagraphContent ||
+            !exposeDesMotifsRegexp.test(firstParagraphContent.toUpperCase())) {
+            if (!secondParagraphContent ||
+                !exposeDesMotifsRegexp.test(secondParagraphContent.toUpperCase())) {
+                continue;
+            }
+            else {
+                secondParagraph.remove();
+            }
+        }
+        firstParagraph.remove();
+        return {
+            text: sectionElement.textContent?.trim() ?? null,
+            html: sectionElement.innerHTML?.trim() ?? null,
+        };
+    }
+    return null;
+}
+export function parseTexte(texteXml) {
+    try {
+        const { document } = new JSDOM(texteXml, {
+            contentType: "text/xml",
+        }).window;
+        return transformTexte(document);
+    }
+    catch (error) {
+        console.error(`Could not parse texte with error ${error}`);
+    }
+    return null;
+}
+// Prevent from memory leak
+// https://github.com/jsdom/jsdom/issues/2583#issuecomment-559520814
+export async function parseTexteFromFile(xmlFilePath) {
+    try {
+        const { document } = (await JSDOM.fromFile(xmlFilePath, { contentType: "text/xml" })).window;
+        return transformTexte(document);
+    }
+    catch (error) {
+        console.error(`Could not parse texte with error ${error}`);
+    }
+    return null;
+}
+export function parseExposeDesMotifs(exposeDesMotifsHtml) {
+    try {
+        const { document } = new JSDOM(exposeDesMotifsHtml, {
+            contentType: "text/html",
+        }).window;
+        return transformExposeDesMotifs(document);
+    }
+    catch (error) {
+        console.error(`Could not parse exposé des motifs with error ${error}`);
+    }
+    return null;
+}
+// Prevent from memory leak
+// https://github.com/jsdom/jsdom/issues/2583#issuecomment-559520814
+export async function parseExposeDesMotifsFromFile(htmlFilePath) {
+    try {
+        const { document } = (await JSDOM.fromFile(htmlFilePath, { contentType: "text/html" })).window;
+        return transformExposeDesMotifs(document);
+    }
+    catch (error) {
+        console.error(`Could not parse exposé des motifs with error ${error}`);
+    }
+    return null;
+}

package/lib/model/texte.mjs

@@ -0,0 +1,208 @@
+import { JSDOM } from "jsdom";
+import { AKN_WORKFLOW_IDENTIFICATION_STRUCTURE_REGEXP, AKN_IDENTIFICATION_STRUCTURE_REGEXP } from "../scripts/datautil";
+import { DivisionType, } from "../types/texte";
+function buildWorklow(metaElement) {
+    const stepElements = metaElement.querySelectorAll("workflow step");
+    const steps = [];
+    for (const stepElement of stepElements) {
+        const identification = stepElement.getAttribute("href") ?? "";
+        const identificationParts = AKN_WORKFLOW_IDENTIFICATION_STRUCTURE_REGEXP.exec(identification)?.groups;
+        steps.push({
+            eId: stepElement.getAttribute("eId"),
+            date: stepElement.getAttribute("date") ? new Date(stepElement.getAttribute("date") ?? "") : null,
+            type: identificationParts?.type || null,
+            session: identificationParts?.session || null,
+            numero: identificationParts?.numTexte || null,
+            version: identificationParts?.version ? identificationParts.version : null,
+            outcome: stepElement.getAttribute("outcome"),
+        });
+    }
+    return steps;
+}
+function buildDivision(node, index) {
+    const eId = node.getAttribute("eId");
+    const tag = node.nodeName;
+    const level = DivisionType[tag];
+    const titleNode = node.querySelector("num");
+    const subtitleNode = node.querySelector("heading");
+    const headings = [
+        ...(titleNode ? [{
+                text: titleNode.textContent?.trim() ?? null,
+                html: titleNode.innerHTML?.trim() ?? null,
+            }] : []),
+        ...(subtitleNode ? [{
+                text: subtitleNode.textContent?.trim() ?? null,
+                html: subtitleNode.innerHTML?.trim() ?? null,
+            }] : []),
+    ];
+    const division = {
+        index,
+        eId,
+        tag,
+        level,
+        headings,
+    };
+    if (tag === "article") {
+        division.alineas = [];
+    }
+    return division;
+}
+function buildAlinea(contentNode, alineaNode) {
+    const eId = alineaNode.getAttribute("eId");
+    const heading = {
+        text: alineaNode.querySelector("num")?.textContent ?? null,
+    };
+    const pastille = alineaNode.getAttribute("data:pastille") ?? null;
+    return {
+        eId,
+        heading,
+        text: contentNode.textContent?.trim() ?? null,
+        html: contentNode.innerHTML?.trim() ?? null,
+        pastille,
+    };
+}
+function buildEmptyArticle(index) {
+    return {
+        index: index,
+        eId: "",
+        tag: "article",
+        level: DivisionType["article"],
+        headings: [],
+        alineas: [],
+    };
+}
+function flattenTexte(texteContentRoot) {
+    const divisions = [];
+    let divisionIndex = 0;
+    const iter = (node) => {
+        if (node.nodeName === "content") {
+            return;
+        }
+        switch (node.nodeName) {
+            case "tome":
+            case "part":
+            case "book":
+            case "title":
+            case "subtitle":
+            case "chapter":
+            case "section":
+            case "subsection":
+            case "paragraph":
+            case "article":
+                divisions.push(buildDivision(node, divisionIndex++));
+                break;
+        }
+        if (node.nodeName === "alinea") {
+            Array.from(node.childNodes)
+                // Find direct content children programmatically
+                // because `:scope` selector does not work
+                // https://github.com/jsdom/jsdom/issues/2998
+                .filter((alineaChildNode) => alineaChildNode.nodeName === "content")
+                .forEach((alineaContentNode) => {
+                // Hypothesis: alineas should always be enclosed in articles
+                let lastArticle = divisions.findLast(division => division.tag === "article");
+                if (!lastArticle) {
+                    lastArticle = buildEmptyArticle(divisionIndex++);
+                    divisions.push(lastArticle);
+                }
+                lastArticle.alineas.push(buildAlinea(alineaContentNode, node));
+            });
+        }
+        if (node.hasChildNodes()) {
+            node.childNodes.forEach((childNode) => iter(childNode));
+        }
+    };
+    iter(texteContentRoot);
+    return divisions;
+}
+export function transformTexte(document) {
+    const metaElement = document.querySelector("meta");
+    const preambleElement = document.querySelector("preamble");
+    const identification = metaElement?.querySelector("FRBRExpression FRBRuri")?.getAttribute("value") ?? "";
+    const identificationParts = AKN_IDENTIFICATION_STRUCTURE_REGEXP.exec(identification)?.groups;
+    const bodyElement = document.querySelector("body");
+    const datePresentation = metaElement?.querySelector("FRBRdate[name='#presentation']")?.getAttribute("date");
+    const dateDepot = metaElement?.querySelector("FRBRdate[name='#depot']")?.getAttribute("date");
+    const datePublicationXml = metaElement?.querySelector("FRBRdate[name='#publication-xml']")?.getAttribute("date");
+    return {
+        titre: preambleElement?.querySelector("docTitle")?.textContent || null,
+        titreCourt: metaElement?.querySelector("FRBRalias[name='intitule-court']")
+            ?.getAttribute("value") || null,
+        signetDossier: metaElement?.querySelector("FRBRalias[name='signet-dossier-legislatif-senat']")
+            ?.getAttribute("value") || null,
+        urlDossierSenat: metaElement?.querySelector("FRBRalias[name='url-senat']")?.getAttribute("value") || null,
+        urlDossierAssemblee: metaElement?.querySelector("FRBRalias[name='url-AN']")?.getAttribute("value") || null,
+        type: identificationParts?.type || null,
+        session: identificationParts?.session || null,
+        numero: identificationParts?.numTexte ? parseInt(identificationParts.numTexte) : null,
+        datePresentation: datePresentation ? new Date(datePresentation) : null,
+        dateDepot: dateDepot ? new Date(dateDepot) : null,
+        datePublicationXml: datePublicationXml ? new Date(datePublicationXml) : null,
+        version: identificationParts?.version ? identificationParts.version : null,
+        workflow: metaElement ? buildWorklow(metaElement) : [],
+        divisions: bodyElement ? flattenTexte(bodyElement) : [],
+    };
+}
+export function transformExposeDesMotifs(document) {
+    const sectionElements = document.querySelectorAll("section");
+    const exposeDesMotifsRegexp = new RegExp("EXPOS.{1,2}[\\n\\s]DES[\\n\\s]MOTIFS");
+    for (const sectionElement of sectionElements) {
+        const firstParagraph = sectionElement.querySelector("p:first-of-type");
+        const secondParagraph = sectionElement.querySelector("p:nth-of-type(2)");
+        if (!firstParagraph) {
+            continue;
+        }
+        const firstParagraphContent = firstParagraph.textContent;
+        const secondParagraphContent = secondParagraph?.textContent;
+        if (!firstParagraphContent || !exposeDesMotifsRegexp.test(firstParagraphContent.toUpperCase())) {
+            if (!secondParagraphContent || !exposeDesMotifsRegexp.test(secondParagraphContent.toUpperCase())) {
+                continue;
+            }
+            else {
+                sectionElement.removeChild(secondParagraph);
+            }
+        }
+        sectionElement.removeChild(firstParagraph);
+        return {
+            text: sectionElement.textContent?.trim() ?? null,
+            html: sectionElement.innerHTML?.trim() ?? null,
+        };
+    }
+    return null;
+}
+export function parseTexte(texteXml) {
+    try {
+        const { document } = (new JSDOM(texteXml, {
+            contentType: "text/xml",
+        })).window;
+        return transformTexte(document);
+    }
+    catch (error) {
+        console.error(`Could not parse texte with error ${error}`);
+    }
+    return null;
+}
+// Prevent from memory leak
+// https://github.com/jsdom/jsdom/issues/2583#issuecomment-559520814
+export async function parseTexteFromFile(xmlFilePath) {
+    try {
+        const { document } = (await JSDOM.fromFile(xmlFilePath, { contentType: "text/xml" })).window;
+        return transformTexte(document);
+    }
+    catch (error) {
+        console.error(`Could not parse texte with error ${error}`);
+    }
+    return null;
+}
+export function parseExposeDesMotifs(exposeDesMotifsHtml) {
+    const { document } = (new JSDOM(exposeDesMotifsHtml, {
+        contentType: "text/html",
+    })).window;
+    return transformExposeDesMotifs(document);
+}
+// Prevent from memory leak
+// https://github.com/jsdom/jsdom/issues/2583#issuecomment-559520814
+export async function parseExposeDesMotifsFromFile(htmlFilePath) {
+    const { document } = (await JSDOM.fromFile(htmlFilePath, { contentType: "text/html" })).window;
+    return transformExposeDesMotifs(document);
+}

package/lib/model/texte.ts

@@ -0,0 +1,229 @@
+import { JSDOM } from "jsdom"
+
+import { AKN_WORKFLOW_IDENTIFICATION_STRUCTURE_REGEXP, AKN_IDENTIFICATION_STRUCTURE_REGEXP } from "../scripts/datautil"
+import {
+  Alinea,
+  Article,
+  Division,
+  DivisionContent,
+  DivisionTag,
+  DivisionType, ExposeDesMotifs,
+  FlatTexte, Step,
+  Version,
+} from "../types/texte"
+
+function buildWorklow (metaElement: HTMLMetaElement): Step[] {
+  const stepElements = metaElement.querySelectorAll("workflow step")
+  const steps: Step[] = []
+  for (const stepElement of stepElements) {
+    const identification = stepElement.getAttribute("href") ?? ""
+    const identificationParts = AKN_WORKFLOW_IDENTIFICATION_STRUCTURE_REGEXP.exec(identification)?.groups
+    steps.push({
+      eId: stepElement.getAttribute("eId")!,
+      date: stepElement.getAttribute("date") ? new Date(stepElement.getAttribute("date") ?? "") : null,
+      type: identificationParts?.type || null,
+      session: identificationParts?.session || null,
+      numero: identificationParts?.numTexte || null,
+      version: identificationParts?.version ? identificationParts.version as Version : null,
+      outcome: stepElement.getAttribute("outcome"),
+    })
+  }
+  return steps
+}
+
+function buildDivision (node: Node, index: number): Division {
+  const eId = (node as Element).getAttribute("eId")!
+  const tag = node.nodeName as DivisionTag
+  const level = DivisionType[tag]
+  const titleNode = (node as Element).querySelector("num")
+  const subtitleNode = (node as Element).querySelector("heading")
+  const headings: DivisionContent[] = [
+    ...(titleNode ? [{
+      text: titleNode.textContent?.trim() ?? null,
+      html: titleNode.innerHTML?.trim() ?? null,
+    }] : []),
+    ...(subtitleNode ? [{
+      text: subtitleNode.textContent?.trim() ?? null,
+      html: subtitleNode.innerHTML?.trim() ?? null,
+    }] : []),
+  ]
+  const division = {
+    index,
+    eId,
+    tag,
+    level,
+    headings,
+  }
+  if (tag === "article") {
+    (division as Article).alineas = []
+  }
+  return division
+}
+
+function buildAlinea (contentNode: Node, alineaNode: Node): Alinea {
+  const eId = (alineaNode as Element).getAttribute("eId")!
+  const heading = {
+    text: (alineaNode as Element).querySelector("num")?.textContent ?? null,
+  }
+  const pastille = (alineaNode as Element).getAttribute("data:pastille") ?? null
+  return {
+    eId,
+    heading,
+    text: (contentNode as Element).textContent?.trim() ?? null,
+    html: (contentNode as Element).innerHTML?.trim() ?? null,
+    pastille,
+  }
+}
+
+function buildEmptyArticle (index: number): Article {
+  return {
+    index: index,
+    eId: "",
+    tag: "article",
+    level: DivisionType["article"],
+    headings: [],
+    alineas: [],
+  }
+}
+
+function flattenTexte (texteContentRoot: Node): Division[] {
+  const divisions: Division[] = []
+  let divisionIndex = 0
+  const iter = (node: Node) => {
+    if (node.nodeName === "content") {
+      return
+    }
+    switch (node.nodeName) {
+      case "tome":
+      case "part":
+      case "book":
+      case "title":
+      case "subtitle":
+      case "chapter":
+      case "section":
+      case "subsection":
+      case "paragraph":
+      case "article":
+        divisions.push(buildDivision(node, divisionIndex++))
+        break
+    }
+    if (node.nodeName === "alinea") {
+      Array.from(node.childNodes)
+        // Find direct content children programmatically
+        // because `:scope` selector does not work
+        // https://github.com/jsdom/jsdom/issues/2998
+        .filter((alineaChildNode: Node) => alineaChildNode.nodeName === "content")
+        .forEach((alineaContentNode: Node) => {
+          // Hypothesis: alineas should always be enclosed in articles
+          let lastArticle = divisions.findLast(division => division.tag === "article") as Article
+          if (!lastArticle) {
+            lastArticle = buildEmptyArticle(divisionIndex++)
+            divisions.push(lastArticle)
+          }
+          lastArticle.alineas.push(buildAlinea(alineaContentNode, node))
+        })
+    }
+    if (node.hasChildNodes()) {
+      node.childNodes.forEach((childNode: Node) => iter(childNode))
+    }
+  }
+  iter(texteContentRoot)
+  return divisions
+}
+
+export function transformTexte (document: Document): FlatTexte | null {
+  const metaElement = document.querySelector("meta")
+  const preambleElement = document.querySelector("preamble")
+  const identification = metaElement?.querySelector("FRBRExpression FRBRuri")?.getAttribute("value") ?? ""
+  const identificationParts = AKN_IDENTIFICATION_STRUCTURE_REGEXP.exec(identification)?.groups
+  const bodyElement = document.querySelector("body")
+  const datePresentation = metaElement?.querySelector("FRBRdate[name='#presentation']")?.getAttribute("date")
+  const dateDepot = metaElement?.querySelector("FRBRdate[name='#depot']")?.getAttribute("date")
+  const datePublicationXml = metaElement?.querySelector("FRBRdate[name='#publication-xml']")?.getAttribute("date")
+  return {
+    titre: preambleElement?.querySelector("docTitle")?.textContent || null,
+    titreCourt: metaElement?.querySelector("FRBRalias[name='intitule-court']")
+      ?.getAttribute("value") || null,
+    signetDossier:
+      metaElement?.querySelector("FRBRalias[name='signet-dossier-legislatif-senat']")
+        ?.getAttribute("value") || null,
+    urlDossierSenat: metaElement?.querySelector("FRBRalias[name='url-senat']")?.getAttribute("value") || null,
+    urlDossierAssemblee: metaElement?.querySelector("FRBRalias[name='url-AN']")?.getAttribute("value") || null,
+    type: identificationParts?.type || null,
+    session: identificationParts?.session || null,
+    numero: identificationParts?.numTexte ? parseInt(identificationParts.numTexte) : null,
+    datePresentation: datePresentation ? new Date(datePresentation) : null,
+    dateDepot: dateDepot ? new Date(dateDepot) : null,
+    datePublicationXml: datePublicationXml ? new Date(datePublicationXml) : null,
+    version: identificationParts?.version ? identificationParts.version as Version : null,
+    workflow: metaElement ? buildWorklow(metaElement) : [],
+    divisions: bodyElement ? flattenTexte(bodyElement) : [],
+  }
+}
+
+export function transformExposeDesMotifs (document: Document): ExposeDesMotifs | null {
+  const sectionElements = document.querySelectorAll("section")
+  const exposeDesMotifsRegexp = new RegExp("EXPOS.{1,2}[\\n\\s]DES[\\n\\s]MOTIFS")
+  for (const sectionElement of sectionElements) {
+    const firstParagraph = sectionElement.querySelector("p:first-of-type")
+    const secondParagraph = sectionElement.querySelector("p:nth-of-type(2)")
+    if (!firstParagraph) {
+      continue
+    }
+
+    const firstParagraphContent = firstParagraph.textContent
+    const secondParagraphContent = secondParagraph?.textContent
+    if (!firstParagraphContent || !exposeDesMotifsRegexp.test(firstParagraphContent.toUpperCase())) {
+      if (!secondParagraphContent || !exposeDesMotifsRegexp.test(secondParagraphContent.toUpperCase())) {
+        continue
+      } else {
+        sectionElement.removeChild(secondParagraph)
+      }
+    }
+
+    sectionElement.removeChild(firstParagraph)
+    return {
+      text: sectionElement.textContent?.trim() ?? null,
+      html: sectionElement.innerHTML?.trim() ?? null,
+    }
+  }
+  return null
+}
+
+export function parseTexte (texteXml: string): FlatTexte | null {
+  try {
+    const { document } = (new JSDOM(texteXml, {
+      contentType: "text/xml",
+    })).window
+    return transformTexte(document)
+  } catch (error: any) {
+    console.error(`Could not parse texte with error ${error}`)
+  }
+  return null
+}
+
+// Prevent from memory leak
+// https://github.com/jsdom/jsdom/issues/2583#issuecomment-559520814
+export async function parseTexteFromFile (xmlFilePath: string): Promise<FlatTexte | null> {
+  try {
+    const { document } = (await JSDOM.fromFile(xmlFilePath, { contentType: "text/xml" })).window
+    return transformTexte(document)
+  } catch (error: any) {
+    console.error(`Could not parse texte with error ${error}`)
+  }
+  return null
+}
+
+export function parseExposeDesMotifs (exposeDesMotifsHtml: string): ExposeDesMotifs | null {
+  const { document } = (new JSDOM(exposeDesMotifsHtml, {
+    contentType: "text/html",
+  })).window
+  return transformExposeDesMotifs(document)
+}
+
+// Prevent from memory leak
+// https://github.com/jsdom/jsdom/issues/2583#issuecomment-559520814
+export async function parseExposeDesMotifsFromFile (htmlFilePath: string): Promise<ExposeDesMotifs | null> {
+  const { document } = (await JSDOM.fromFile(htmlFilePath, { contentType: "text/html" })).window
+  return transformExposeDesMotifs(document)
+}