add test module

sentence_api/config.py

@@ -0,0 +1,75 @@
+import os
+from dataclasses import dataclass
+from pathlib import Path
+
+
+def _bool_env(name: str, default: bool) -> bool:
+    value = os.getenv(name)
+    if value is None:
+        return default
+    return value.strip().lower() in {"1", "true", "yes", "on"}
+
+
+@dataclass(frozen=True)
+class Settings:
+    data_dir: Path
+    legacy_boundaries_file: Path
+    admin_api_key: str
+    client_api_key: str
+    public_base_url: str
+    max_upload_bytes: int
+    max_attempt_bytes: int
+    keep_attempt_audio: bool
+    moss_transcribe_url: str
+    moss_model: str
+    moss_timeout_seconds: float
+    moss_max_new_tokens: int
+    pass_score: float
+
+    @classmethod
+    def from_env(cls) -> "Settings":
+        package_dir = Path(__file__).resolve().parent
+        data_dir = Path(os.getenv("ORAL_TRAINER_DATA_DIR", str(package_dir / "data")))
+        return cls(
+            data_dir=data_dir,
+            legacy_boundaries_file=Path(
+                os.getenv(
+                    "SENTENCE_BOUNDARIES_FILE",
+                    str(package_dir / "data" / "sentence_boundaries.json"),
+                )
+            ),
+            admin_api_key=os.getenv("ADMIN_API_KEY", ""),
+            client_api_key=os.getenv("CLIENT_API_KEY", ""),
+            public_base_url=os.getenv("PUBLIC_BASE_URL", "").rstrip("/"),
+            max_upload_bytes=int(os.getenv("MAX_VIDEO_UPLOAD_BYTES", str(12 * 1024**3))),
+            max_attempt_bytes=int(os.getenv("MAX_ATTEMPT_UPLOAD_BYTES", str(50 * 1024**2))),
+            keep_attempt_audio=_bool_env("KEEP_ATTEMPT_AUDIO", False),
+            moss_transcribe_url=os.getenv("MOSS_TRANSCRIBE_URL", "").strip(),
+            moss_model=os.getenv(
+                "MOSS_MODEL",
+                "OpenMOSS-Team/MOSS-Transcribe-Diarize",
+            ),
+            moss_timeout_seconds=float(os.getenv("MOSS_TIMEOUT_SECONDS", "1800")),
+            moss_max_new_tokens=int(os.getenv("MOSS_MAX_NEW_TOKENS", "65536")),
+            pass_score=float(os.getenv("ASSESSMENT_PASS_SCORE", "70")),
+        )
+
+    @property
+    def videos_dir(self) -> Path:
+        return self.data_dir / "v"
+
+    @property
+    def work_dir(self) -> Path:
+        return self.data_dir / "work"
+
+    @property
+    def attempts_dir(self) -> Path:
+        return self.data_dir / "attempts"
+
+    @property
+    def database_path(self) -> Path:
+        return self.data_dir / "oral_trainer.sqlite3"
+
+    def ensure_directories(self) -> None:
+        for path in (self.data_dir, self.videos_dir, self.work_dir, self.attempts_dir):
+            path.mkdir(parents=True, exist_ok=True)