add volcengine OSS support

sentence_api/repository.py

@@ -43,6 +43,7 @@ class VideoRepository:
                     error_message TEXT,
                     algorithm_version TEXT,
                     transcription TEXT,
+                    remote_url TEXT,
                     created_at TEXT NOT NULL,
                     updated_at TEXT NOT NULL
                 );
@@ -113,6 +114,14 @@ class VideoRepository:
                     "ALTER TABLE dub_share_segments ADD COLUMN score_details TEXT NOT NULL DEFAULT '{}'"
                 )
 
+        with self._connect() as connection:
+            video_columns = {
+                row["name"]
+                for row in connection.execute("PRAGMA table_info(videos)")
+            }
+            if "remote_url" not in video_columns:
+                connection.execute("ALTER TABLE videos ADD COLUMN remote_url TEXT")
+
     def upsert_upload(
         self,
         *,
@@ -137,8 +146,8 @@ class VideoRepository:
                 INSERT INTO videos (
                     video_hash, title, filename, stored_filename, content_type,
                     size_bytes, duration_ms, language, status, error_message,
-                    algorithm_version, transcription, created_at, updated_at
-                ) VALUES (?, ?, ?, ?, ?, ?, NULL, ?, 'uploaded', NULL, NULL, NULL, ?, ?)
+                    algorithm_version, transcription, remote_url, created_at, updated_at
+                ) VALUES (?, ?, ?, ?, ?, ?, NULL, ?, 'uploaded', NULL, NULL, NULL, NULL, ?, ?)
                 ON CONFLICT(video_hash) DO UPDATE SET
                     title = excluded.title,
                     filename = excluded.filename,
@@ -180,6 +189,7 @@ class VideoRepository:
         self,
         document: SentenceBoundaryDocument,
         transcription: Optional[str],
+        ready: bool = True,
     ) -> None:
         with self._connect() as connection:
             connection.execute(
@@ -208,12 +218,13 @@ class VideoRepository:
             )
             connection.execute(
                 """
-                UPDATE videos SET duration_ms = ?, status = 'ready', error_message = NULL,
+                UPDATE videos SET duration_ms = ?, status = ?, error_message = NULL,
                     algorithm_version = ?, transcription = ?, updated_at = ?
                 WHERE video_hash = ?
                 """,
                 (
                     document.duration_ms,
+                    "ready" if ready else "processing",
                     document.algorithm_version,
                     transcription,
                     utc_now(),
@@ -221,6 +232,17 @@ class VideoRepository:
                 ),
             )
 
+    def mark_oss_uploaded(self, video_hash: str, remote_url: str) -> None:
+        with self._connect() as connection:
+            connection.execute(
+                """
+                UPDATE videos
+                SET remote_url = ?, status = 'ready', error_message = NULL, updated_at = ?
+                WHERE video_hash = ?
+                """,
+                (remote_url, utc_now(), normalize_video_hash(video_hash)),
+            )
+
     def list_videos(self) -> List[Dict[str, Any]]:
         with self._connect() as connection:
             rows = connection.execute(