fix: Title content after period sometimes stripped in import (#9712)

closes #9694

server/commands/documentImporter.test.ts

@@ -37,6 +37,27 @@ describe("documentImporter", () => {
     expect(response.title).toEqual("images");
   });
 
+  it("should not strip content after period in title", async () => {
+    const user = await buildUser();
+    const fileName = "01. test";
+    const content = await fs.readFile(
+      path.resolve(__dirname, "..", "test", "fixtures", "images.docx")
+    );
+
+    const response = await sequelize.transaction((transaction) =>
+      documentImporter({
+        user,
+        mimeType:
+          "application/vnd.openxmlformats-officedocument.wordprocessingml.document",
+        fileName,
+        content,
+        ctx: createContext({ user, transaction }),
+      })
+    );
+    expect(response.text).toContain("This is a test document for images");
+    expect(response.title).toEqual("01. test");
+  });
+
   it("should convert Word Document to markdown for application/octet-stream mimetype", async () => {
     const user = await buildUser();
     const fileName = "images.docx";

server/commands/documentImporter.ts

@@ -1,4 +1,5 @@
 import emojiRegex from "emoji-regex";
+import mime from "mime-types";
 import truncate from "lodash/truncate";
 import parseTitle from "@shared/utils/parseTitle";
 import { DocumentValidation } from "@shared/validations";
@@ -35,7 +36,19 @@ async function documentImporter({
     fileName,
     mimeType
   );
-  let title = fileName.replace(/\.[^/.]+$/, "");
+
+  // find valid extensions and remove them from the title
+  const extensions = [
+    "docx",
+    "md",
+    "markdown",
+    "html",
+    ...(mime.extensions[mimeType] ?? []),
+  ];
+  let title = fileName.replace(
+    new RegExp(`\\.(${extensions.join("|")})$`, "i"),
+    ""
+  );
 
   // find and extract emoji near the beginning of the document.
   const regex = emojiRegex();