Merge branch 'main' into fix/1733-batch-size

Author: vrcprl

packages/cli/CHANGELOG.md

@@ -1,5 +1,17 @@
 # lingo.dev
 
+## 0.133.7
+
+### Patch Changes
+
+- [#2059](https://github.com/lingodotdev/lingo.dev/pull/2059) [`4344399`](https://github.com/lingodotdev/lingo.dev/commit/43443996d4e6b6e4d4979bbe2aa7d5039b672d41) Thanks [@AndreyHirsa](https://github.com/AndreyHirsa)! - Skip undefined cues during VTT incremental batch push
+
+## 0.133.6
+
+### Patch Changes
+
+- [#2057](https://github.com/lingodotdev/lingo.dev/pull/2057) [`abe3d2f`](https://github.com/lingodotdev/lingo.dev/commit/abe3d2fb8b0f3356dc63f8f9d4b861de66be0da6) Thanks [@AndreyHirsa](https://github.com/AndreyHirsa)! - Fix VTT parser crash on files with STYLE/REGION blocks
+
 ## 0.133.5
 
 ### Patch Changes

packages/cli/package.json

@@ -1,6 +1,6 @@
 {
   "name": "lingo.dev",
-  "version": "0.133.5",
+  "version": "0.133.7",
   "description": "Lingo.dev CLI",
   "private": false,
   "repository": {

packages/cli/src/cli/loaders/index.spec.ts

@@ -2645,6 +2645,128 @@ Another cue
       expect(Object.values(data)).toContain("Another cue");
       expect(Object.values(data)).not.toContain("Hello world!");
     });
+
+    it("should handle vtt files with STYLE blocks", async () => {
+      setupFileMocks();
+
+      const input = `
+WEBVTT
+
+STYLE
+::cue(.heading) {
+  font-size: 150%;
+  font-weight: bold;
+}
+
+00:00:00.000 --> 00:00:01.000
+Hello world!
+
+00:00:30.000 --> 00:00:31.000
+Another cue
+      `.trim();
+
+      const expectedOutput = {
+        "0#0-1#": "Hello world!",
+        "1#30-31#": "Another cue",
+      };
+
+      mockFileOperations(input);
+
+      const vttLoader = createBucketLoader("vtt", "i18n/[locale].vtt", {
+        defaultLocale: "en",
+      });
+      vttLoader.setDefaultLocale("en");
+      const data = await vttLoader.pull("en");
+
+      expect(data).toEqual(expectedOutput);
+    });
+
+    it("should preserve STYLE blocks in push output", async () => {
+      setupFileMocks();
+
+      const input = `
+WEBVTT
+
+STYLE
+::cue(.heading) {
+  font-size: 150%;
+  font-weight: bold;
+}
+
+00:00:00.000 --> 00:00:01.000
+Hello world!
+      `.trim();
+
+      mockFileOperations(input);
+
+      const vttLoader = createBucketLoader("vtt", "i18n/[locale].vtt", {
+        defaultLocale: "en",
+      });
+      vttLoader.setDefaultLocale("en");
+      await vttLoader.pull("en");
+      await vttLoader.push("es", { "0#0-1#": "¡Hola mundo!" });
+
+      const written = (fs.writeFile as any).mock.calls[0][1] as string;
+      expect(written).toContain("STYLE");
+      expect(written).toContain("::cue(.heading)");
+      expect(written).toContain("¡Hola mundo!");
+    });
+
+    it("should handle vtt files with REGION blocks", async () => {
+      setupFileMocks();
+
+      const input = `
+WEBVTT
+
+REGION
+id:sidebar
+width:30%
+lines:3
+
+00:00:00.000 --> 00:00:01.000
+Hello world!
+      `.trim();
+
+      mockFileOperations(input);
+
+      const vttLoader = createBucketLoader("vtt", "i18n/[locale].vtt", {
+        defaultLocale: "en",
+      });
+      vttLoader.setDefaultLocale("en");
+      const data = await vttLoader.pull("en");
+
+      expect(data).toEqual({ "0#0-1#": "Hello world!" });
+    });
+
+    it("should preserve REGION blocks in push output", async () => {
+      setupFileMocks();
+
+      const input = `
+WEBVTT
+
+REGION
+id:sidebar
+width:30%
+lines:3
+
+00:00:00.000 --> 00:00:01.000
+Hello world!
+      `.trim();
+
+      mockFileOperations(input);
+
+      const vttLoader = createBucketLoader("vtt", "i18n/[locale].vtt", {
+        defaultLocale: "en",
+      });
+      vttLoader.setDefaultLocale("en");
+      await vttLoader.pull("en");
+      await vttLoader.push("es", { "0#0-1#": "¡Hola mundo!" });
+
+      const written = (fs.writeFile as any).mock.calls[0][1] as string;
+      expect(written).toContain("REGION");
+      expect(written).toContain("id:sidebar");
+      expect(written).toContain("¡Hola mundo!");
+    });
   });
 
   describe("XML bucket loader", () => {

packages/cli/src/cli/loaders/vtt.ts

@@ -2,6 +2,26 @@ import webvtt from "node-webvtt";
 import { ILoader } from "./_types";
 import { createLoader } from "./_utils";
 
+const UNSUPPORTED_BLOCK_REGEX = /^(?:STYLE|REGION)\s*$/;
+
+function isUnsupportedBlock(block: string): boolean {
+  const firstLine = block.trimStart().split("\n", 1)[0];
+  return UNSUPPORTED_BLOCK_REGEX.test(firstLine);
+}
+
+// node-webvtt doesn't handle STYLE/REGION blocks — strip them before parsing
+function stripUnsupportedBlocks(input: string): string {
+  return input
+    .replace(/\r\n/g, "\n")
+    .split("\n\n")
+    .filter((part) => !isUnsupportedBlock(part))
+    .join("\n\n");
+}
+
+function getUnsupportedBlocks(input: string): string[] {
+  return input.replace(/\r\n/g, "\n").split("\n\n").filter(isUnsupportedBlock);
+}
+
 export default function createVttLoader(): ILoader<
   string,
   Record<string, any>
@@ -11,7 +31,7 @@ export default function createVttLoader(): ILoader<
       if (!input) {
         return ""; // if VTT file does not exist yet we can not parse it - return empty string
       }
-      const vtt = webvtt.parse(input)?.cues;
+      const vtt = webvtt.parse(stripUnsupportedBlocks(input))?.cues;
       if (Object.keys(vtt).length === 0) {
         return {};
       } else {
@@ -22,27 +42,40 @@ export default function createVttLoader(): ILoader<
         }, {});
       }
     },
-    async push(locale, payload) {
-      const output = Object.entries(payload).map(([key, text]) => {
-        const [id, timeRange, identifier] = key.split("#");
-        const [startTime, endTime] = timeRange.split("-");
-
-        return {
-          end: Number(endTime),
-          identifier: identifier,
-          start: Number(startTime),
-          styles: "",
-          text: text,
-        };
-      });
+    async push(locale, payload, originalInput) {
+      const output = Object.entries(payload).reduce(
+        (cues: any[], [key, text]) => {
+          if (!text) return cues;
+
+          const [, timeRange, identifier] = key.split("#");
+          const [startTime, endTime] = timeRange.split("-");
+
+          cues.push({
+            end: Number(endTime),
+            identifier,
+            start: Number(startTime),
+            styles: "",
+            text,
+          });
+          return cues;
+        },
+        [],
+      );
 
       const input = {
         valid: true,
         strict: true,
         cues: output,
       };
 
-      return webvtt.compile(input);
+      const compiled = webvtt.compile(input);
+
+      // Re-insert STYLE/REGION blocks after the WEBVTT header
+      const blocks = getUnsupportedBlocks(originalInput ?? "");
+      if (blocks.length === 0) return compiled;
+
+      const [header, ...rest] = compiled.split("\n\n");
+      return [header, ...blocks, ...rest].join("\n\n");
     },
   });
 }