fixes performance issues and output file

action.yml

@@ -1,5 +1,5 @@
-name: 'DeepL Translate Github Action'
-author: 'Estee Tey'
+name: 'DeepL API Translate Github Action'
+author: 'Tyler Haas'
 description: 'Translate any document using DeepL Translate API'
 
 inputs:
@@ -15,6 +15,9 @@ inputs:
   output_file_name_pattern:
     description: "Pattern of the output file name, including the folder name"
     required: true
+  base_ref:
+    description: "Optional git branch to diff the source file against before translating. Defaults to the pull request base branch when available."
+    required: false
   deepl_api_key:
     description: "API Key for DeepL API"
     required: true
@@ -29,25 +32,38 @@ inputs:
     description: "End tag to ignore when translating"
     required: false
     default: <!-- /notranslate -->
+  model_type:
+    description: "Model type to use for translation. Valid values: quality_optimized, prefer_quality_optimized, latency_optimized"
+    required: false
+  timeout:
+    description: "Connection timeout for each HTTP request retry, in milliseconds. If not provided, uses the default value (5000ms)"
+    required: false
 
 runs:
   using: composite
   steps:
   - name: Checkout
-    uses: actions/checkout@v3
+    uses: actions/checkout@v4
+    with:
+      repository: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.repo.full_name || github.repository }}
+      ref: ${{ github.event_name == 'pull_request' && github.head_ref || github.ref }}
+      fetch-depth: 0
   - name: Set up node
-    uses: actions/setup-node@v3
+    uses: actions/setup-node@v4
   - name: Run translation script
     shell: bash
     env:
       target_languages: ${{ inputs.target_languages }}
       input_file_path: ${{ inputs.input_file_path }}
       output_file_name_prefix: ${{ inputs.output_file_name_prefix }}
       output_file_name_pattern: ${{ inputs.output_file_name_pattern }}
+      base_ref: ${{ inputs.base_ref }}
       deepl_api_key: ${{ inputs.deepl_api_key }}
       ignore_terms: ${{ inputs.ignore_terms }}
       no_translate_start_tag: ${{ inputs.no_translate_start_tag }}
       no_translate_end_tag: ${{ inputs.no_translate_end_tag }}
+      model_type: ${{ inputs.model_type }}
+      timeout: ${{ inputs.timeout }}
     run: |
       cd ${{github.action_path}} && yarn install && yarn start
   - name: remove unused temp file if it exists

package.json

@@ -1,6 +1,6 @@
 {
   "name": "deepl-translate-github-action",
-  "version": "2.2.0",
+  "version": "2.2.1",
   "author": "Estee <estee.tsw@gmail.com>",
   "license": "MIT",
   "scripts": {
@@ -10,12 +10,12 @@
     "coverage": "vitest run --coverage"
   },
   "dependencies": {
-    "deepl-node": "^1.14.0"
+    "deepl-node": "^1.22.0"
   },
   "devDependencies": {
     "@types/node": "^22.7.6",
     "ts-node": "^10.9.2",
     "typescript": "^5.6.3",
     "vitest": "^2.1.3"
   }
-}
+}

playground/local.ts

@@ -12,7 +12,42 @@ const endTagForNoTranslate = "<!-- /keep -->";
 
 const tempFilePath = path.join(playgroundPath, "to_translate.txt");
 const fileExtensionsThatAllowForIgnoringBlocks = [".html", ".xml", ".md", ".txt"];
-const targetLanguages = ["ja"] as deepl.TargetLanguageCode[];
+
+// All supported DeepL target languages as of April 2025
+const targetLanguages: deepl.TargetLanguageCode[] = [
+  "ar",   // Arabic
+  "bg",   // Bulgarian
+  "cs",   // Czech
+  "da",   // Danish
+  "de",   // German
+  "el",   // Greek
+  "en-GB", // English (British)
+  "en-US", // English (American)
+  "es",   // Spanish
+  "et",   // Estonian
+  "fi",   // Finnish
+  "fr",   // French
+  "hu",   // Hungarian
+  "id",   // Indonesian
+  "it",   // Italian
+  "ja",   // Japanese
+  "ko",   // Korean
+  "lt",   // Lithuanian
+  "lv",   // Latvian
+  "nb",   // Norwegian Bokmål
+  "nl",   // Dutch
+  "pl",   // Polish
+  "pt-BR", // Portuguese (Brazilian)
+  "pt-PT", // Portuguese (all Portuguese variants excluding Brazilian Portuguese)
+  "ro",   // Romanian
+  "ru",   // Russian
+  "sk",   // Slovak
+  "sl",   // Slovenian
+  "sv",   // Swedish
+  "tr",   // Turkish
+  "uk",   // Ukrainian
+  "zh"    // Chinese (unspecified variant for backward compatibility)
+];
 
 (async () => {
 	await main({

src/git.ts

@@ -0,0 +1,108 @@
+import { execFile } from 'child_process'
+import { promisify } from 'util'
+
+const execFileAsync = promisify(execFile)
+
+interface BaseFileContentParams {
+  workspacePath: string
+  inputFileRelativePath: string
+  baseRef?: string
+}
+
+function isMissingPathError(stderr: string) {
+  return (
+    stderr.includes('exists on disk, but not in') ||
+    stderr.includes('pathspec') ||
+    stderr.includes('fatal: path') ||
+    stderr.includes('does not exist')
+  )
+}
+
+function getWorkflowRepositoryUrl(): string | null {
+  const serverUrl = process.env.GITHUB_SERVER_URL?.replace(/\/$/, '')
+  const repository = process.env.GITHUB_REPOSITORY
+
+  if (!serverUrl || !repository) {
+    return null
+  }
+
+  return `${serverUrl}/${repository}.git`
+}
+
+export async function getBaseFileContent({
+  workspacePath,
+  inputFileRelativePath,
+  baseRef,
+}: BaseFileContentParams): Promise<string | null> {
+  if (!baseRef) {
+    return null
+  }
+
+  const baseRepoUrl = getWorkflowRepositoryUrl()
+  const baseBranchRef = baseRepoUrl
+    ? `refs/remotes/base/${baseRef}`
+    : `refs/remotes/origin/${baseRef}`
+  const fetchSource = baseRepoUrl ?? 'origin'
+
+  try {
+    await execFileAsync(
+      'git',
+      [
+        'fetch',
+        '--no-tags',
+        fetchSource,
+        `+refs/heads/${baseRef}:${baseBranchRef}`,
+      ],
+      { cwd: workspacePath, maxBuffer: 10 * 1024 * 1024 },
+    )
+  } catch (error) {
+    console.warn(`Failed to fetch base branch ${baseBranchRef}, falling back to local refs if available.`, error)
+  }
+
+  let baseCommitRef: string | null = null
+
+  try {
+    const { stdout } = await execFileAsync(
+      'git',
+      ['merge-base', 'HEAD', baseBranchRef],
+      { cwd: workspacePath, maxBuffer: 10 * 1024 * 1024 },
+    )
+
+    const mergeBaseSha = stdout.trim()
+    if (mergeBaseSha) {
+      baseCommitRef = mergeBaseSha
+    }
+  } catch (error) {
+    console.warn(
+      `Failed to determine merge-base with ${baseBranchRef}, falling back to full translation.`,
+      error,
+    )
+  }
+
+  if (!baseCommitRef) {
+    return null
+  }
+
+  try {
+    const normalizedInputFileRelativePath = inputFileRelativePath.replace(/\\/g, '/')
+    const { stdout } = await execFileAsync(
+      'git',
+      ['show', `${baseCommitRef}:${normalizedInputFileRelativePath}`],
+      { cwd: workspacePath, maxBuffer: 10 * 1024 * 1024 },
+    )
+
+    return stdout
+  } catch (error) {
+    const stderr = error instanceof Error && 'stderr' in error ? String(error.stderr) : ''
+
+    if (isMissingPathError(stderr)) {
+      return null
+    }
+
+    console.warn(
+      `Failed to read ${inputFileRelativePath} from ${baseCommitRef}, falling back to full translation.`,
+      error,
+    )
+    return null
+  }
+}

src/index.ts

@@ -1,23 +1,30 @@
 import type { TargetLanguageCode } from "deepl-node";
 import { Translator } from 'deepl-node';
 import path from "path";
-import { main } from "./main";
+import { getBaseFileContent } from "./git";
+import { main, type ModelType } from "./main";
+import { createTranslatorOptions } from "./utils";
 
 const authKey = process.env.deepl_api_key as string;
-const translator = new Translator(authKey);
+const translatorOptions = createTranslatorOptions(process.env.timeout);
+
+const translator = new Translator(authKey, translatorOptions);
+const workspacePath = process.env.GITHUB_WORKSPACE as string;
+const inputFileRelativePath = process.env.input_file_path as string;
+const baseRef = process.env.base_ref || process.env.GITHUB_BASE_REF;
 const inputFilePath = path.join(
-	process.env.GITHUB_WORKSPACE as string,
-	process.env.input_file_path as string,
+	workspacePath,
+	inputFileRelativePath,
 );
 const outputFileNamePattern = path.join(
-	process.env.GITHUB_WORKSPACE as string,
+	workspacePath,
 	process.env.output_file_name_pattern as string,
 )
 const startTagForNoTranslate = process.env.no_translate_start_tag as string;
 const endTagForNoTranslate = process.env.no_translate_end_tag as string;
 
 const tempFilePath = path.join(
-	process.env.GITHUB_WORKSPACE as string,
+	workspacePath,
 	"to_translate.txt",
 );
 const fileExtensionsThatAllowForIgnoringBlocks = [".html", ".xml", ".md", ".txt"];
@@ -34,14 +41,35 @@ const fileExtensionsThatAllowForIgnoringBlocks = [".html", ".xml", ".md", ".txt"
 
 	targetLanguages = targetLanguages.filter(lang => !excludedLanguages.includes(lang));
 
+	const modelTypeEnv = process.env.model_type;
+	let modelType: ModelType | undefined;
+	if (modelTypeEnv) {
+		const validModelTypes: ModelType[] = ['quality_optimized', 'prefer_quality_optimized', 'latency_optimized'];
+		if (validModelTypes.includes(modelTypeEnv as ModelType)) {
+			modelType = modelTypeEnv as ModelType;
+		} else {
+			console.warn(`Invalid model_type value: ${modelTypeEnv}. Valid values are: ${validModelTypes.join(', ')}. Ignoring model_type parameter.`);
+		}
+	}
+
+	const baseFileContent = await getBaseFileContent({
+		workspacePath,
+		inputFileRelativePath,
+		baseRef,
+	});
+
 	await main({
 		translator,
+		workspacePath,
+		inputFileRelativePath,
 		inputFilePath,
 		outputFileNamePattern,
 		startTagForNoTranslate,
 		endTagForNoTranslate,
 		tempFilePath,
 		fileExtensionsThatAllowForIgnoringBlocks,
 		targetLanguages,
+		modelType,
+		baseFileContent,
 	});
 })();