stream

Stream utilities

This directory contains companion utilities for a running Plate Recognizer Stream installation. For Stream setup and configuration, see the Stream documentation.

Batch video processing

video_upload.py uploads every file in a directory to Stream's file-upload API and appends successful responses to output.jsonl.

Install the directory's dependencies, then start the uploader:

python -m pip install -r requirements.txt
python video_upload.py /path/to/videos

The file-upload server defaults to http://localhost:8081. Use --sdk-url to change it and --mask to select a camera mask.

Stream monitor

stream_monitor.py reads a Stream container's Docker logs and exposes its status over HTTP.

python stream_monitor.py --help
python stream_monitor.py --container stream --interval 30 --duration 120

The server listens on localhost:8001 by default. Query it from another terminal:

curl http://localhost:8001

Example response:

{
  "active": true,
  "cameras": {
    "camera-1": {"status": "running"},
    "camera-2": {"status": "offline"}
  }
}

active reports whether the Stream container is running. A camera is marked offline when it has not produced a recognized log entry within --duration seconds.

Remove Stream images

remove_images.py recursively deletes .jpg files and then removes empty directories. It supports two mutually exclusive cleanup modes:

• Age mode deletes images older than 1–23 hours.
• Disk-usage mode deletes the oldest images first when disk usage reaches a trigger percentage, stopping when the target free-space percentage is met.

This utility permanently deletes files. Test the command against a disposable directory before scheduling it against Stream data.

Delete by age

Delete images older than 23 hours:

python remove_images.py --age --threshold 23 --directory /path/to/stream

Delete by disk usage

When disk usage reaches 80%, delete the oldest images until at least 20% of the disk is free:

python remove_images.py \
  --usage \
  --trigger_usage 80 \
  --target_free 20 \
  --directory /path/to/stream

Run python remove_images.py --help for the complete argument reference. If you do not want Stream to save images at all, use the appropriate Stream configuration instead of periodically deleting them.

Schedule cleanup

On Linux, add one selected cleanup command to cron. For example, this runs age cleanup every night at midnight:

0 0 * * * /usr/bin/python3 /path/to/remove_images.py --age --threshold 23 --directory /path/to/stream

This example checks disk usage every ten minutes:

*/10 * * * * /usr/bin/python3 /path/to/remove_images.py --usage --trigger_usage 80 --target_free 20 --directory /path/to/stream

Use crontab -e for the account that owns the Stream files. If you edit /etc/crontab instead, include the account name between the schedule and the command.

On Windows, create a scheduled task in Command Prompt or PowerShell. The following example runs age cleanup daily at midnight:

schtasks /create /sc daily /tn RemoveStreamImages /tr "python C:\path\to\remove_images.py --age --threshold 23 --directory C:\PlateRecognizer\Stream" /st 00:00

For disk-usage cleanup, use a distinct task name and command:

schtasks /create /sc minute /mo 10 /tn FreeUpStreamDiskSpace /tr "python C:\path\to\remove_images.py --usage --trigger_usage 80 --target_free 20 --directory C:\PlateRecognizer\Stream"

To remove these tasks later:

schtasks /delete /tn RemoveStreamImages /f
schtasks /delete /tn FreeUpStreamDiskSpace /f