baseline: initial working version

tests/conftest.py

@@ -0,0 +1,11 @@
+"""Pytest configuration for local imports."""
+
+from __future__ import annotations
+
+import sys
+from pathlib import Path
+
+
+ROOT = Path(__file__).resolve().parents[1]
+if str(ROOT) not in sys.path:
+    sys.path.insert(0, str(ROOT))

tests/test_main_cli.py

@@ -0,0 +1,61 @@
+"""Tests for CLI parser and translation config wiring."""
+
+from __future__ import annotations
+
+from main import _build_translation_config, build_parser
+
+
+def test_parser_accepts_lmstudio_flags():
+    parser = build_parser()
+
+    args = parser.parse_args(
+        [
+            "https://youtube.com/watch?v=demo",
+            "--translation-backend",
+            "lmstudio",
+            "--lmstudio-base-url",
+            "http://localhost:1234/v1",
+            "--lmstudio-model",
+            "gemma-custom",
+        ]
+    )
+
+    assert args.translation_backend == "lmstudio"
+    assert args.lmstudio_base_url == "http://localhost:1234/v1"
+    assert args.lmstudio_model == "gemma-custom"
+
+
+def test_translation_config_prefers_cli_over_env(monkeypatch):
+    monkeypatch.setenv("LM_STUDIO_BASE_URL", "http://env-host:1234/v1")
+    monkeypatch.setenv("LM_STUDIO_MODEL", "env-model")
+
+    parser = build_parser()
+    args = parser.parse_args(
+        [
+            "https://youtube.com/watch?v=demo",
+            "--lmstudio-base-url",
+            "http://cli-host:1234/v1",
+            "--lmstudio-model",
+            "cli-model",
+        ]
+    )
+
+    config = _build_translation_config(args)
+
+    assert config.base_url == "http://cli-host:1234/v1"
+    assert config.model == "cli-model"
+
+
+def test_translation_config_uses_env_defaults(monkeypatch):
+    monkeypatch.setenv("LM_STUDIO_BASE_URL", "http://env-host:1234/v1")
+    monkeypatch.setenv("LM_STUDIO_MODEL", "env-model")
+    monkeypatch.setenv("LM_STUDIO_API_KEY", "env-key")
+
+    parser = build_parser()
+    args = parser.parse_args(["https://youtube.com/watch?v=demo"])
+
+    config = _build_translation_config(args)
+
+    assert config.base_url == "http://env-host:1234/v1"
+    assert config.model == "env-model"
+    assert config.api_key == "env-key"

tests/test_translation.py

@@ -0,0 +1,136 @@
+"""Tests for the LM Studio translation layer."""
+
+from __future__ import annotations
+
+import httpx
+import pytest
+
+from src.core_utils import TranslationError
+from src.translation import LMStudioTranslator, TranslationConfig
+
+
+def _mock_client(handler):
+    return httpx.Client(transport=httpx.MockTransport(handler))
+
+
+def test_translation_config_normalizes_base_url():
+    config = TranslationConfig.from_env(base_url="http://127.0.0.1:1234")
+
+    assert config.base_url == "http://127.0.0.1:1234/v1"
+    assert config.chat_completions_url == "http://127.0.0.1:1234/v1/chat/completions"
+    assert config.model == "gemma-3-4b-it"
+
+
+def test_build_payload_includes_model_and_prompt():
+    translator = LMStudioTranslator(TranslationConfig(), client=_mock_client(lambda request: None))
+
+    payload = translator.build_payload("Hello world", "en", "es")
+
+    assert payload["model"] == "gemma-3-4b-it"
+    assert payload["messages"][0]["role"] == "system"
+    assert "Translate the user-provided text from en to es." in payload["messages"][0]["content"]
+    assert payload["messages"][1]["content"] == "Hello world"
+
+
+def test_translate_segments_preserves_order_and_blank_segments():
+    def handler(request: httpx.Request) -> httpx.Response:
+        text = request.read().decode("utf-8")
+        if "first" in text:
+            content = "primero"
+        elif "third" in text:
+            content = "tercero"
+        else:
+            content = "desconocido"
+        return httpx.Response(200, json={"choices": [{"message": {"content": content}}]})
+
+    translator = LMStudioTranslator(TranslationConfig(), client=_mock_client(handler))
+
+    translated = translator.translate_segments(["first", "", "third"], target_language="es", source_language="en")
+
+    assert translated == ["primero", "", "tercero"]
+
+
+def test_retry_on_transient_http_error_then_succeeds():
+    attempts = {"count": 0}
+
+    def handler(request: httpx.Request) -> httpx.Response:
+        attempts["count"] += 1
+        if attempts["count"] == 1:
+            return httpx.Response(503, json={"error": {"message": "busy"}})
+        return httpx.Response(200, json={"choices": [{"message": {"content": "hola"}}]})
+
+    translator = LMStudioTranslator(
+        TranslationConfig(max_retries=2),
+        client=_mock_client(handler),
+        sleeper=lambda _: None,
+    )
+
+    translated = translator.translate_text("hello", target_language="es", source_language="en")
+
+    assert translated == "hola"
+    assert attempts["count"] == 2
+
+
+def test_parse_response_content_rejects_empty_content():
+    with pytest.raises(TranslationError, match="empty translation"):
+        LMStudioTranslator.parse_response_content({"choices": [{"message": {"content": "  "}}]})
+
+
+def test_translate_text_raises_on_malformed_response():
+    def handler(request: httpx.Request) -> httpx.Response:
+        return httpx.Response(200, json={"choices": []})
+
+    translator = LMStudioTranslator(TranslationConfig(), client=_mock_client(handler))
+
+    with pytest.raises(TranslationError, match="did not contain a chat completion message"):
+        translator.translate_text("hello", target_language="es", source_language="en")
+
+
+def test_translate_text_falls_back_to_user_only_prompt_for_template_error():
+    attempts = {"count": 0}
+
+    def handler(request: httpx.Request) -> httpx.Response:
+        attempts["count"] += 1
+        body = request.read().decode("utf-8")
+        if attempts["count"] == 1:
+            return httpx.Response(
+                400,
+                text='{"error":"Error rendering prompt with jinja template: \\"Conversations must start with a user prompt.\\""}',
+            )
+        assert '"role":"user"' in body
+        return httpx.Response(200, json={"choices": [{"message": {"content": "hola"}}]})
+
+    translator = LMStudioTranslator(TranslationConfig(), client=_mock_client(handler))
+
+    translated = translator.translate_text("hello", target_language="es", source_language="en")
+
+    assert translated == "hola"
+    assert attempts["count"] == 2
+
+
+def test_translate_text_falls_back_to_structured_prompt_for_custom_template():
+    attempts = {"count": 0}
+
+    def handler(request: httpx.Request) -> httpx.Response:
+        attempts["count"] += 1
+        body = request.read().decode("utf-8")
+        if attempts["count"] == 1:
+            return httpx.Response(
+                400,
+                text='{"error":"Error rendering prompt with jinja template: \\"Conversations must start with a user prompt.\\""}',
+            )
+        if attempts["count"] == 2:
+            return httpx.Response(
+                400,
+                text='{"error":"Error rendering prompt with jinja template: \\"User role must provide `content` as an iterable with exactly one item. That item must be a mapping(type:\'text\' | \'image\', source_lang_code:string, target_lang_code:string, text:string | none, image:string | none).\\""}',
+            )
+        assert '"source_lang_code":"en"' in body
+        assert '"target_lang_code":"es"' in body
+        return httpx.Response(200, json={"choices": [{"message": {"content": "hola"}}]})
+
+    translator = LMStudioTranslator(TranslationConfig(), client=_mock_client(handler))
+
+    translated = translator.translate_text("hello", target_language="es", source_language="en")
+
+    assert translated == "hola"
+    assert attempts["count"] == 3