feat: URL parser matches bare domains (domain.tld)

Author: ClaudiuCeia

README.md

@@ -369,7 +369,7 @@ const entities = Duckling([Email.parser, Hashtag.parser]).extract(
 | **Temperature** | `temperature` | `72°F`, `20 celsius`, `-5°C`                             |
 | **Quantity**    | `quantity`    | `5 kg`, `100 miles`, `3,500.00`                          |
 | **Location**    | `location`    | `United States`, `Germany`, `Japan`                      |
-| **URL**         | `url`         | `https://example.com/path?q=1`                           |
+| **URL**         | `url`         | `https://example.com/path?q=1`, `docs.example.org`       |
 | **Institution** | `institution` | `University of Oxford`, `New York City Hall`             |
 | **Language**    | `language`    | `English`, `Japanese`, `Portuguese`                      |
 

src/URL.ts

@@ -46,6 +46,7 @@ type URLLanguage = {
   Suffix: Parser<string>;
   Domain: Parser<string>;
   Full: Parser<URLEntity>;
+  Bare: Parser<URLEntity>;
   parser: Parser<URLEntity>;
 };
 
@@ -108,7 +109,26 @@ export const URL: URLLanguage = createLanguage<URLLanguage>({
         ),
     );
   },
+  Bare: (s): Parser<URLEntity> => {
+    return map(
+      seq(
+        s.Domain,
+        optional(seq(str(":"), s.Port)),
+        optional(s.Suffix),
+      ),
+      ([domain, maybePort, maybeSuffix], b, a) =>
+        url(
+          {
+            url: `${domain}${
+              maybePort ? `:${maybePort[1]}` : ""
+            }${maybeSuffix ?? ""}`,
+          },
+          b,
+          a,
+        ),
+    );
+  },
   parser: (s): Parser<URLEntity> => {
-    return dot(any(s.Full));
+    return dot(any(s.Full, s.Bare));
   },
 });

tests/URL.test.ts

@@ -86,3 +86,55 @@ Deno.test("URL with query params (Wikipedia create account)", () => {
     "https://en.wikipedia.org/w/index.php?title=Special:CreateAccount&returnto=Master+Juba",
   );
 });
+
+Deno.test("URL bare domain", () => {
+  const res = Duckling().extract("Visit google.com for more");
+
+  assertEquals(res, [
+    {
+      start: 6,
+      end: 16,
+      kind: "url",
+      text: "google.com",
+      value: { url: "google.com" },
+    },
+  ]);
+});
+
+Deno.test("URL bare domain with subdomain", () => {
+  const res = Duckling().extract("Check docs.example.org please");
+
+  assertEquals(res, [
+    {
+      start: 6,
+      end: 22,
+      kind: "url",
+      text: "docs.example.org",
+      value: { url: "docs.example.org" },
+    },
+  ]);
+});
+
+Deno.test("URL bare domain with path", () => {
+  const res = Duckling().extract("See example.com/about for info");
+
+  assertEquals(res[0].kind, "url");
+  assertEquals(res[0].text, "example.com/about");
+  assertEquals(res[0].value, { url: "example.com/about" });
+});
+
+Deno.test("URL bare domain with port", () => {
+  const res = Duckling().extract("Running at localhost.com:3000 now");
+
+  assertEquals(res[0].kind, "url");
+  assertEquals(res[0].text, "localhost.com:3000");
+  assertEquals(res[0].value, { url: "localhost.com:3000" });
+});
+
+Deno.test("URL prefers full URL over bare domain", () => {
+  const res = Duckling().extract("Go to https://example.com/path please");
+
+  assertEquals(res.length, 1);
+  assertEquals(res[0].text, "https://example.com/path");
+  assertEquals(res[0].value, { url: "https://example.com/path" });
+});