🐍 Remove .py and .md from linkify domains (#940)

jupyter-book · Feb 28, 2024 · b289f03 · b289f03
1 parent cff47b1
commit b289f03
Show file tree

Hide file tree

Showing 6 changed files with 37 additions and 1 deletion.
diff --git a/.changeset/breezy-plants-act.md b/.changeset/breezy-plants-act.md
@@ -0,0 +1,5 @@
+---
+'myst-parser': patch
+---
+
+Remove .py from linkify domains
diff --git a/package-lock.json b/package-lock.json
diff --git a/packages/myst-parser/package.json b/packages/myst-parser/package.json
@@ -55,6 +55,7 @@
     "myst-directives": "^1.0.22",
     "myst-roles": "^1.0.22",
     "myst-spec": "^0.0.5",
+    "tlds": "^1.250.0",
     "unified": "^10.1.1",
     "unist-builder": "^3.0.0",
     "unist-util-remove": "^3.1.0",

diff --git a/packages/myst-parser/src/config.ts b/packages/myst-parser/src/config.ts
@@ -67,3 +67,6 @@ export const MARKDOWN_IT_CONFIG = {
     },
   },
 };
+
+// List of valid TLDs to exclude from linkify
+export const EXCLUDE_TLDS = ['py', 'md'];
diff --git a/packages/myst-parser/src/myst.ts b/packages/myst-parser/src/myst.ts
@@ -3,7 +3,8 @@ import { defaultDirectives } from 'myst-directives';
 import { defaultRoles } from 'myst-roles';
 import type { Plugin } from 'unified';
 import { VFile } from 'vfile';
-import { MARKDOWN_IT_CONFIG } from './config.js';
+import tlds from 'tlds';
+import { EXCLUDE_TLDS, MARKDOWN_IT_CONFIG } from './config.js';
 import { tokensToMyst } from './tokensToMyst.js';
 import {
   mathPlugin,
@@ -70,6 +71,9 @@ export function createTokenizer(opts?: Options) {
     } as any,
     markdownit,
   );
+  if (markdownit.linkify) {
+    tokenizer.linkify.tlds(tlds.filter((tld) => !EXCLUDE_TLDS.includes(tld)));
+  }
   if (extensions.smartquotes) tokenizer.enable('smartquotes');
   if (extensions.tables) tokenizer.enable('table');
   if (extensions.colonFences) tokenizer.use(colonFencePlugin);

diff --git a/packages/myst-parser/tests/linkify.spec.ts b/packages/myst-parser/tests/linkify.spec.ts
@@ -104,4 +104,18 @@ describe('linkify', () => {
       ],
     });
   });
+  it('dont linkify .py', () => {
+    const content = 'Link in paragraph: example.py';
+    const expected = {
+      type: 'root',
+      children: [
+        {
+          type: 'paragraph',
+          children: [{ type: 'text', value: 'Link in paragraph: example.py' }],
+        },
+      ],
+    };
+    expect(stripPositions(mystParse(content))).toEqual(expected);
+    expect(stripPositions(mystParse(content, { markdownit: { linkify: true } }))).toEqual(expected);
+  });
 });