feat: add batch size parameter to prevent context leaking

.changeset/add-batch-size.md

@@ -0,0 +1,5 @@
+---
+"lingo.dev": minor
+---
+
+feat: add `--batch-size` parameter to `run` and `i18n` commands to prevent context leaking

packages/cli/src/cli/cmd/i18n.ts

@@ -90,6 +90,11 @@ export default new Command()
     "--strict",
     "Stop immediately on first error instead of continuing to process remaining buckets and locales (fail-fast mode)",
   )
+  .option(
+    "--batch-size <number>",
+    "Number of translations to process in a single batch",
+    parseInt,
+  )
   .action(async function (options) {
     updateGitignore();
 
@@ -429,13 +434,13 @@ export default new Command()
                 }
 
                 bucketOra.start(
-                  `[${sourceLocale} -> ${targetLocale}] [${
-                    Object.keys(processableData).length
+                  `[${sourceLocale} -> ${targetLocale}] [${Object.keys(processableData).length
                   } entries] (0%) AI localization in progress...`,
                 );
                 let processPayload = createProcessor(i18nConfig!.provider, {
                   apiKey: settings.auth.apiKey,
                   apiUrl: settings.auth.apiUrl,
+                  batchSize: flags.batchSize,
                 });
                 processPayload = withExponentialBackoff(
                   processPayload,
@@ -453,9 +458,8 @@ export default new Command()
                     targetData: flags.force ? {} : targetData,
                   },
                   (progress, sourceChunk, processedChunk) => {
-                    bucketOra.text = `[${sourceLocale} -> ${targetLocale}] [${
-                      Object.keys(processableData).length
-                    } entries] (${progress}%) AI localization in progress...`;
+                    bucketOra.text = `[${sourceLocale} -> ${targetLocale}] [${Object.keys(processableData).length
+                      } entries] (${progress}%) AI localization in progress...`;
                   },
                 );
 
@@ -657,6 +661,7 @@ function parseFlags(options: any) {
     file: Z.array(Z.string()).optional(),
     interactive: Z.boolean().prefault(false),
     debug: Z.boolean().prefault(false),
+    batchSize: Z.number().min(1).optional(),
   }).parse(options);
 }
 

packages/cli/src/cli/cmd/run/_types.ts

@@ -55,5 +55,6 @@ export const flagsSchema = z.object({
   debounce: z.number().positive().prefault(5000), // 5 seconds default
   sound: z.boolean().optional(),
   pseudo: z.boolean().optional(),
+  batchSize: z.number().min(1).optional(),
 });
 export type CmdRunFlags = z.infer<typeof flagsSchema>;

packages/cli/src/cli/cmd/run/index.ts

@@ -122,6 +122,11 @@ export default new Command()
     "--pseudo",
     "Enable pseudo-localization mode: automatically pseudo-translates all extracted strings with accented characters and visual markers without calling any external API. Useful for testing UI internationalization readiness",
   )
+  .option(
+    "--batch-size <number>",
+    "Number of translations to process in a single batch (not applicable when using lingo.dev provider)",
+    (val: string) => parseInt(val),
+  )
   .action(async (args) => {
     let email: string | null = null;
     try {

packages/cli/src/cli/cmd/run/setup.ts

@@ -52,15 +52,20 @@ export default async function setup(input: CmdRunContext) {
         task: async (ctx, task) => {
           const provider = ctx.flags.pseudo ? "pseudo" : ctx.config?.provider;
           const vNext = ctx.config?.vNext;
-          ctx.localizer = createLocalizer(provider, ctx.flags.apiKey, vNext);
+          ctx.localizer = createLocalizer(
+            provider,
+            ctx.flags.apiKey,
+            vNext,
+            ctx.flags.batchSize,
+          );
           if (!ctx.localizer) {
             throw new Error(
               "Could not create localization provider. Please check your i18n.json configuration.",
             );
           }
           task.title =
             ctx.localizer.id === "Lingo.dev" ||
-            ctx.localizer.id === "Lingo.dev vNext"
+              ctx.localizer.id === "Lingo.dev vNext"
               ? `Using ${chalk.hex(colors.green)(ctx.localizer.id)} provider`
               : ctx.localizer.id === "pseudo"
                 ? `Using ${chalk.hex(colors.blue)("pseudo")} mode for testing`
@@ -108,23 +113,23 @@ export default async function setup(input: CmdRunContext) {
 
           const subTasks = isLingoDotDev
             ? [
-                "Brand voice enabled",
-                "Translation memory connected",
-                "Glossary enabled",
-                "Quality assurance enabled",
-              ].map((title) => ({ title, task: () => {} }))
+              "Brand voice enabled",
+              "Translation memory connected",
+              "Glossary enabled",
+              "Quality assurance enabled",
+            ].map((title) => ({ title, task: () => { } }))
             : isPseudo
               ? [
-                  "Pseudo-localization mode active",
-                  "Character replacement configured",
-                  "No external API calls",
-                ].map((title) => ({ title, task: () => {} }))
+                "Pseudo-localization mode active",
+                "Character replacement configured",
+                "No external API calls",
+              ].map((title) => ({ title, task: () => { } }))
               : [
-                  "Skipping brand voice",
-                  "Skipping glossary",
-                  "Skipping translation memory",
-                  "Skipping quality assurance",
-                ].map((title) => ({ title, task: () => {}, skip: true }));
+                "Skipping brand voice",
+                "Skipping glossary",
+                "Skipping translation memory",
+                "Skipping quality assurance",
+              ].map((title) => ({ title, task: () => { }, skip: true }));
 
           return task.newListr(subTasks, {
             concurrent: true,