Support source_lang param for translation recipe

[prasoonvarshney]

Currently, the source language is is hardcoded to English.

Adding a quick fix to read source_lang from config, just like target_lang and defaulting it to English to preserve current default behavior.

Summary by CodeRabbit

• New Features
  • Source language for translation tasks is now configurable (default: English), so users can set the input language for translations without changing core behavior. This update makes translating from languages other than English easier to configure and use across existing translation workflows.

[coderabbitai]

📝 Walkthrough

Walkthrough

Added a configurable source_lang field (default "en") to TranslationConfig. Replaced hardcoded source-language strings with self.cfg.source_lang in data preprocessing and example prompt logging; no changes to target language, translation key, or core translation logic.

Changes

Cohort / File(s)	Summary
Configuration Enhancement recipes/translation/translate_jsonl.py	Added source_lang: str = "en" to TranslationConfig (nested dataclass). Updated preprocess_data to use self.cfg.source_lang instead of hardcoded "en", and updated log_example_prompt to use self.cfg.source_lang instead of hardcoded "English".

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~5 minutes

🚥 Pre-merge checks | ✅ 3

✅ Passed checks (3 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title 'Support source_lang param for translation recipe' directly and clearly summarizes the main change: adding a configurable source_lang parameter to the translation recipe, which aligns with the PR's core objective of making source language configurable instead of hardcoded.
Docstring Coverage	✅ Passed	Docstring coverage is 100.00% which is sufficient. The required threshold is 80.00%.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

• 📝 Generate docstrings (stacked PR)
• 📝 Generate docstrings (commit on current branch)

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Post copyable unit tests in a comment
• Commit unit tests in branch feature/translation_source_lang

Tip

Try Coding Plans. Let us write the prompt for your AI agent so you can ship faster (with fewer bugs).
Share your feedback on Discord.

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[coderabbitai]

🧹 Nitpick comments (1)

recipes/translation/translate_jsonl.py (1)

225-233: Avoid repeated language-name lookups inside the per-line loop.

full_language_name(self.cfg.source_lang) and full_language_name(self.cfg.target_lang) are invariant across all items; precompute once before iterating to keep preprocessing simpler and cheaper.

Refactor suggestion

         translation_data_points = []
         async_position = 0
+        source_lang_name = full_language_name(self.cfg.source_lang)
+        target_lang_name = full_language_name(self.cfg.target_lang)
         for text in all_translatable_lines:
             translation_data_point = {
-                "source_lang": full_language_name(self.cfg.source_lang),
-                "target_lang": full_language_name(self.cfg.target_lang),
+                "source_lang": source_lang_name,
+                "target_lang": target_lang_name,
                 "src": text,
                 self.cfg.async_position_key: async_position,
             }

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@recipes/translation/translate_jsonl.py` around lines 225 - 233, Precompute
the invariant language names before the per-line loop instead of calling
full_language_name(self.cfg.source_lang) and
full_language_name(self.cfg.target_lang) for every item; assign them to local
variables (e.g., src_lang_name, tgt_lang_name) and then use those in
constructing translation_data_point inside the loop that iterates over
all_translatable_lines, keeping use of self.cfg.async_position_key and
async_position unchanged.

🤖 Prompt for all review comments with AI agents

Verify each finding against the current code and only fix it if needed.

Nitpick comments:
In `@recipes/translation/translate_jsonl.py`:
- Around line 225-233: Precompute the invariant language names before the
per-line loop instead of calling full_language_name(self.cfg.source_lang) and
full_language_name(self.cfg.target_lang) for every item; assign them to local
variables (e.g., src_lang_name, tgt_lang_name) and then use those in
constructing translation_data_point inside the loop that iterates over
all_translatable_lines, keeping use of self.cfg.async_position_key and
async_position unchanged.

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

Run ID: 0e337559-44d4-4e0f-b297-076c06ffef1e

📥 Commits

Reviewing files that changed from the base of the PR and between 5ac8609 and b377d29.

📒 Files selected for processing (1)

• recipes/translation/translate_jsonl.py

[shuoyangd]

LGTM

[coderabbitai]

🧹 Nitpick comments (1)

recipes/translation/translate_jsonl.py (1)

225-233: Optional: precompute language display names outside the per-text loop.

This loop repeatedly resolves the same source_lang/target_lang names for every segment. You can compute them once before iterating.

♻️ Proposed refactor

         translation_data_points = []
         async_position = 0
+        source_lang_name = full_language_name(self.cfg.source_lang)
+        target_lang_name = full_language_name(self.cfg.target_lang)
         for text in all_translatable_lines:
             translation_data_point = {
-                "source_lang": full_language_name(self.cfg.source_lang),
-                "target_lang": full_language_name(self.cfg.target_lang),
+                "source_lang": source_lang_name,
+                "target_lang": target_lang_name,
                 "src": text,
                 self.cfg.async_position_key: async_position,
             }

As per coding guidelines, "Keep code simple and elegant; reuse/extend existing functionality when possible."

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@recipes/translation/translate_jsonl.py` around lines 225 - 233, The loop in
which translation_data_point objects are created repeatedly calls
full_language_name(self.cfg.source_lang) and
full_language_name(self.cfg.target_lang) for each text; compute source and
target display names once before iterating (e.g., store in local variables like
src_display and tgt_display) and use those variables inside the loop that builds
translation_data_point (which includes self.cfg.async_position_key,
async_position, and appends to translation_data_points) to avoid redundant work
and improve clarity.

🤖 Prompt for all review comments with AI agents

Verify each finding against the current code and only fix it if needed.

Nitpick comments:
In `@recipes/translation/translate_jsonl.py`:
- Around line 225-233: The loop in which translation_data_point objects are
created repeatedly calls full_language_name(self.cfg.source_lang) and
full_language_name(self.cfg.target_lang) for each text; compute source and
target display names once before iterating (e.g., store in local variables like
src_display and tgt_display) and use those variables inside the loop that builds
translation_data_point (which includes self.cfg.async_position_key,
async_position, and appends to translation_data_points) to avoid redundant work
and improve clarity.

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

Run ID: 7c6cc3e3-a5b3-45ea-a69d-29afe8713254

📥 Commits

Reviewing files that changed from the base of the PR and between b377d29 and b7a3792.

📒 Files selected for processing (1)

• recipes/translation/translate_jsonl.py