The websocket included script

src/paperless_migration/scripts/transform.py

@@ -3,6 +3,7 @@
 #   "rich",
 #   "ijson",
 #   "typer-slim",
+#   "websockets",
 # ]
 # ///
 
@@ -14,6 +15,7 @@ from pathlib import Path
 from typing import Any
 from typing import TypedDict
 
+import ijson
 import typer
 from rich.console import Console
 from rich.progress import BarColumn
@@ -22,14 +24,8 @@ from rich.progress import SpinnerColumn
 from rich.progress import TextColumn
 from rich.progress import TimeElapsedColumn
 from rich.table import Table
-
+from websockets.sync.client import ClientConnection
-try:
+from websockets.sync.client import connect
-    import ijson  # type: ignore
-except ImportError as exc:  # pragma: no cover - handled at runtime
-    raise SystemExit(
-        "ijson is required for migration transform. "
-        "Install dependencies (e.g., `uv pip install ijson`).",
-    ) from exc
 
 app = typer.Typer(add_completion=False)
 console = Console()
@@ -80,6 +76,8 @@ def migrate(
         "-o",
         callback=validate_output,
     ),
+    ws_url: str | None = typer.Option(None, "--ws"),
+    update_frequency: int = typer.Option(100, "--freq"),
 ) -> None:
     """
     Process JSON fixtures with detailed summary and timing.
@@ -94,6 +92,15 @@ def migrate(
     total_processed: int = 0
     start_time: float = time.perf_counter()
 
+    ws: ClientConnection | None = None
+    if ws_url:
+        try:
+            ws = connect(ws_url)
+        except Exception as e:
+            console.print(
+                f"[yellow]Warning: Could not connect to WebSocket: {e}[/yellow]",
+            )
+
     progress = Progress(
         SpinnerColumn(),
         TextColumn("[bold blue]{task.description}"),
@@ -103,6 +110,7 @@ def migrate(
         console=console,
     )
 
+    try:
         with (
             progress,
             input_path.open("rb") as infile,
@@ -129,8 +137,23 @@ def migrate(
                 json.dump(fixture, outfile, ensure_ascii=False)
                 progress.advance(task, 1)
 
+                if ws and (i % update_frequency == 0):
+                    ws.send(
+                        json.dumps(
+                            {
+                                "task": "processing",
+                                "completed": total_processed,
+                                "stats": dict(stats),
+                            },
+                        ),
+                    )
+
             outfile.write("\n]\n")
 
+    finally:
+        if ws:
+            ws.close()
+
     end_time: float = time.perf_counter()
     duration: float = end_time - start_time