Merge pull request #104 from mimotej/fix-thesis-scraping

fix: Fix scraping issues caused by recent change of dspace site
OqixDevs · Feb 14, 2024 · 9a44b5b · 9a44b5b
2 parents c0e359c + 2dadb5b
commit 9a44b5b
Show file tree

Hide file tree

Showing 3 changed files with 16 additions and 6 deletions.
diff --git a/src/commands/verify.ts b/src/commands/verify.ts
@@ -67,9 +67,13 @@ export async function execute(interaction: ChatInputCommandInteraction) {
         });
     }
     try {
+        let thesisId = bachelorThesisParsedUrl.pathname.split('/')[3];
+        if (thesisId === undefined) {
+            thesisId = bachelorThesisParsedUrl.pathname.split('/')[2];
+        }
         user = await prisma.users.findMany({
             where: {
-                idThesis: bachelorThesisParsedUrl.pathname.split('/')[3],
+                idThesis: thesisId,
             },
         });
     } catch (err) {

diff --git a/src/utils/assignRole.ts b/src/utils/assignRole.ts
@@ -52,11 +52,14 @@ export async function assignRole(
             );
             const member = interaction.member as GuildMember;
             try {
+                let idThesis = bachelorThesisParsedUrl.pathname.split('/')[3];
+                if (idThesis === undefined) {
+                    idThesis = bachelorThesisParsedUrl.pathname.split('/')[2];
+                }
                 await prisma.users.create({
                     data: {
                         discordId: interaction.user.id,
-                        idThesis:
-                            bachelorThesisParsedUrl.pathname.split('/')[3],
+                        idThesis: idThesis,
                         status: 'verified',
                     },
                 });

diff --git a/src/utils/scrapeThesis.ts b/src/utils/scrapeThesis.ts
@@ -15,8 +15,11 @@ export async function scrapeThesis(bachelorThesisPath: string) {
     }
     const $ = load(response.data);
     const authorName = $(
-        '#aspect_artifactbrowser_ItemViewer_div_item-view > div > div.row > div.col-sm-4 > div:nth-child(2) > div > a'
+        'ds-metadata-representation-list.ds-item-page-mixed-author-field:nth-child(4) > ds-metadata-field-wrapper:nth-child(1) > div:nth-child(1) > div:nth-child(2) > ds-metadata-representation-loader:nth-child(1) > ds-plain-text-metadata-list-element:nth-child(1) > div:nth-child(1) > a:nth-child(1)'
     ).text();
-
-    return authorName.split(',').reverse().join(' ');
+    return authorName
+        .split(',')
+        .reverse()
+        .map((name) => name.trim())
+        .join(' ');
 }