logiciels/tricoteuses-legifrance
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Normalize HTML & Remove <br> from texts

Browse source
This commit is contained in:
Emmanuel 2024-12-13 23:26:29 +01:00
parent 05a6704a13
commit 1576eac0ff
1 changed files with 23 additions and 7 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

30
src/scripts/export_consolidated_text_to_git.ts
View file

@ -3,6 +3,7 @@ import dedent from "dedent-js"
import fs from "fs-extra" import fs from "fs-extra"
import git from "isomorphic-git" import git from "isomorphic-git"
import path from "path" import path from "path"
import * as prettier from "prettier"
import sade from "sade" import sade from "sade"
import type { import type {
@ -319,6 +320,18 @@ async function addModifyingTextId(
} }
} }
async function cleanHtmlFragment(
fragment: string | undefined,
): Promise<string | undefined> {
return fragment === undefined
? undefined
: await prettier.format(
fragment.replaceAll("<<", "«").replaceAll(">>", "»"),
{
parser: "html",
},
)
}
async function exportConsolidatedTextToGit( async function exportConsolidatedTextToGit(
consolidatedTextId: string, consolidatedTextId: string,
targetDir: string, targetDir: string,
@ -803,13 +816,16 @@ async function exportConsolidatedTextToGit(
"Ancien identifiant", "Ancien identifiant",
jorfModifyingTexteVersion.META.META_COMMUN.ANCIEN_ID, jorfModifyingTexteVersion.META.META_COMMUN.ANCIEN_ID,
], ],
// TODO: Mettre l'URL dans le Git Tricoteuses // TODO: Mettre l'URL dans Légifrance et(?) le Git Tricoteuses
["URL", jorfModifyingTexteVersion.META.META_COMMUN.URL], ["URL", jorfModifyingTexteVersion.META.META_COMMUN.URL],
] ]
.filter(([, value]) => value !== undefined) .filter(([, value]) => value !== undefined)
.map(([key, value]) => `${key}: ${value}`) .map(([key, value]) => `${key}: ${value}`)
.join("\n") .join("\n")
summary = jorfModifyingTexteVersion.SM?.CONTENU summary = jorfModifyingTexteVersion.SM?.CONTENU?.replace(
/<br\s*\/>/gi,
"\n",
)
} else if (modifyingTextId.startsWith("LEGITEXT")) { } else if (modifyingTextId.startsWith("LEGITEXT")) {
const legiModifyingTexteVersion = const legiModifyingTexteVersion =
modifyingTexteVersion as LegiTexteVersion modifyingTexteVersion as LegiTexteVersion
@ -945,7 +961,7 @@ async function generateSectionTaGitDirectory(
###### ${articleTitle} ###### ${articleTitle}
${article.BLOC_TEXTUEL?.CONTENU} ${await cleanHtmlFragment(article.BLOC_TEXTUEL?.CONTENU)}
` + "\n", ` + "\n",
) )
await git.add({ await git.add({
@ -1091,7 +1107,7 @@ async function generateTextGitDirectory(
###### ${articleTitle} ###### ${articleTitle}
${article.BLOC_TEXTUEL?.CONTENU.replaceAll("<<", "«").replaceAll(">>", "»")} ${await cleanHtmlFragment(article.BLOC_TEXTUEL?.CONTENU)}
` + "\n", ` + "\n",
) )
await git.add({ await git.add({
@ -1138,9 +1154,9 @@ async function generateTextGitDirectory(
const readmeBlocks = [ const readmeBlocks = [
`${"#".repeat(Math.min(depth, 6))} ${texteTitle}`, `${"#".repeat(Math.min(depth, 6))} ${texteTitle}`,
texteVersion.VISAS?.CONTENU, await cleanHtmlFragment(texteVersion.VISAS?.CONTENU),
readmeLinks.map(({ href, title }) => `- [${title}](${href})`).join("\n"), readmeLinks.map(({ href, title }) => `- [${title}](${href})`).join("\n"),
texteVersion.SIGNATAIRES?.CONTENU, await cleanHtmlFragment(texteVersion.SIGNATAIRES?.CONTENU),
].filter((block) => block != null) ].filter((block) => block != null)
const readmeRepositoryRelativeFilePath = path.join( const readmeRepositoryRelativeFilePath = path.join(
repositoryRelativeDir, repositoryRelativeDir,
@ -1164,7 +1180,7 @@ async function generateTextGitDirectory(
["Identifiant", texteVersion.META.META_COMMUN.ID], ["Identifiant", texteVersion.META.META_COMMUN.ID],
["NOR", texteVersion.META.META_SPEC.META_TEXTE_CHRONICLE.NOR], ["NOR", texteVersion.META.META_SPEC.META_TEXTE_CHRONICLE.NOR],
["Ancien identifiant", texteVersion.META.META_COMMUN.ANCIEN_ID], ["Ancien identifiant", texteVersion.META.META_COMMUN.ANCIEN_ID],
// TODO: Mettre l'URL dans le Git Tricoteuses // TODO: Mettre l'URL dans Légifrance et(?) le Git Tricoteuses
["URL", texteVersion.META.META_COMMUN.URL], ["URL", texteVersion.META.META_COMMUN.URL],
] ]
.filter(([, value]) => value !== undefined) .filter(([, value]) => value !== undefined)