Blank Canvas When Importing Local-Files TimeSeries Tasks via Docker

**Describe the bug**
When importing a task JSON that points at locally-mounted video and CSV files via the /data/local-files/ connector, Label Studio shows a blank labeling screen. No video or graph loads, and the browser console reports:

[2025-06-25 14:02:50,537] [django.server::log_message::213] [INFO] "GET /data/local-files/?d=task-2/time-series.csv HTTP/1.1" 200 5562
[2025-06-25 14:02:50,537] [django.server::log_message::213] [INFO] "GET /data/local-files/?d=task-2/time-series.csv HTTP/1.1" 200 5562
[2025-06-25 14:02:50,564] [django.server::log_message::213] [INFO] "GET /data/local-files/?d=task-2/time-series.csv HTTP/1.1" 200 5562
[2025-06-25 14:02:50,564] [django.server::log_message::213] [INFO] "GET /data/local-files/?d=task-2/time-series.csv HTTP/1.1" 200 5562
[2025-06-25 14:02:50,602] [django.server::log_message::213] [INFO] "GET /data/local-files/?d=task-2/opossum_snow.mp4 HTTP/1.1" 206 1
[2025-06-25 14:02:50,602] [django.server::log_message::213] [INFO] "GET /data/local-files/?d=task-2/opossum_snow.mp4 HTTP/1.1" 206 1

**To Reproduce**
Steps to reproduce the behavior:
Start Label Studio with Docker (nightly develop tag) and these env vars:

1.
docker run -it \
-p 8080:8080 \
-v /home/hostpath/labelstudio_data:/label-studio/data \
-v /home/hostpath:/label-studio/files \
-e LABEL_STUDIO_LOCAL_FILES_SERVING_ENABLED=true \
-e LABEL_STUDIO_LOCAL_FILES_DOCUMENT_ROOT=/label-studio/files \
-e fflag_feat_front_bros58_timeseries_multichannel_short=1 \
-e FF_MULTICHANNEL_TS=true \
heartexlabs/label-studio:20250624.145640-develop-629af7ffb

2. Place opossum_snow.mp4 and time-series.csv under your host path

3. Import this JSON via the UI (Project → Import):

[
{
"id": 9,
"data": {
"video": "/data/local-files/?d=task-2/opossum_snow.mp4",
"accel_data": "/data/local-files/?d=task-2/time-series.csv",
"gyro_data": "/data/local-files/?d=task-2/time-series.csv"
},
"annotations": [],
"predictions": []
}
]

4. Open the newly created task and observe that the labeling canvas is completely blank—no video or chart

**Expected behavior**
The video player should load and play back opossum_snow.mp4 via the local-files connector.

The two TimeSeries panels (accelerometer and gyroscope) should fetch and render time-series.csv according to the XML template’s timeColumn, timeFormat, and settings.

No CORS or undefined-property errors should occur because all files are served locally.

**Screenshots**
If applicable, add screenshots to help explain your problem.

**Environment (please complete the following information):**
-Label Studio Docker image: heartexlabs/label-studio:20250624.145640-develop-629af7ffb

Running via Docker on: Rocky Linux
Docker flags:

-e LABEL_STUDIO_LOCAL_FILES_SERVING_ENABLED=true

-e LABEL_STUDIO_LOCAL_FILES_DOCUMENT_ROOT=/label-studio/files

Data mount:

Host: /home/hostpath/task-2

Container: /label-studio/files/task-2

Label config XML:

```

:contentReference[oaicite:1]{index=1}
:contentReference[oaicite:2]{index=2}

:contentReference[oaicite:3]{index=3}
:contentReference[oaicite:4]{index=4}

```

**Additional context**



External-URL example tasks (using https://app.heartex.ai/static/samples/opossum_snow.mp4, https://app.humansignal.com/samples/time-series.csv?time=time&values=accel_x%2Caccel_y&sep=%2C&tf=%H:%M:%S.%f, and https://app.humansignal.com/samples/time-series.csv?time=time&values=gyro_x%2Cgyro_y&sep=%2C&tf=%H:%M:%S.%f) render perfectly with the same XML.

Local-files tasks always blank or error out, despite confirming via docker exec that files exist and are readable (-rw-r--r--).