v1.5.3

README.md

@@ -120,9 +120,9 @@ If you need to test SIO analytics service and don't have an available RTSP sourc
 Next, copy your test video file to the live555 mount path:
 
 ```bash
-mkdir -p "${SH_BASE}"/services/live555/test-data
+mkdir -p "${SH_BASE}"/media/input/video/live555
 # cp or scp
-cp <my-video>  "${SH_BASE}"/services/live555/test-data/my-video.mp4
+cp <my-video>  "${SH_BASE}"/media/input/video/live555/my-video.mp4
 ```
 
 You can also execute this by running:

RELEASE.md

@@ -1,5 +1,14 @@
 # Release Notes
 
+## v1.5.3
+- Add Aqueduct API and UI examples
+- Refine Aqueduct runner example
+- Add fakeRTSP example configuration
+- Update live555
+- Update MCP to 1.3.2
+- Update SIO to r231120
+- Fix samples' compatibility with DNNCam/DNNNode
+- Allow samples to define SIO image version via SIO_RELEASE define
 
 
 ## v1.5.2

VERSION

@@ -1 +1 @@
-v1.5.2
+v1.5.3

configurations/fakeRTSP.conf

@@ -0,0 +1,13 @@
+select_live555_video
+disable amqp-stats
+remove_orphans
+enable live555
+enable mcp
+enable rabbitmq
+enable sio
+select_example sio live555
+up live555
+up rabbitmq
+up mcp
+test_rtsp_stream rtsp://localhost:7554/data/my-video.mkv 5
+restart sio

deployment-examples/SighthoundRestApiGateway/docker-compose.yml

@@ -2,7 +2,7 @@ version: "3"
 services:
 
   analytics:
-    image: us-central1-docker.pkg.dev/ext-edge-analytics/docker/sio:r230828
+    image: us-central1-docker.pkg.dev/ext-edge-analytics/docker/sio:${SIO_RELEASE-r231120}${SIO_DOCKER_TAG_VARIANT}
     restart: unless-stopped
     environment:
       # Location where SIO will place generated model engine files

deployment-examples/StandaloneSIOWithExtension/README.md

@@ -1,15 +1,17 @@
-# Standalone Folder Watcher Deployment of SIO
+# Standalone SIO sample
 
 In this example, SIO is deployed in a stand-alone manner, without a RabbitMQ broker or client processing the results.
 Instead, a pipeline extension is used to process and report the pipeline output.
 
 Contact [[email protected]](mailto:[email protected]) with any questions, and visit our [Developer Portal](https://dev.sighthound.com) for more information.
 
 
-## Running the sample
+## General
 
 Before getting started, you must copy your `sighthound-license.json` into the `./StandaloneSIOWithExtension/config/` folder. If you do not have a license, please contact [[email protected]](mailto:[email protected]).
 
+## Running the folder watcher sample
+
 Ensure `./StandaloneSIOWithExtension/data/input` exists prior to starting the service.
 
 Next, open a terminal, `cd` into the `./StandaloneSIOWithExtension/` folder, and run the following command to start the services:
@@ -26,3 +28,18 @@ SIO_DOCKER_RUNTIME=nvidia docker compose up -d
 
 You can then deposit images and videos into `./StandaloneSIOWithExtension/data/input` and watch the output being printed as those are being procesed.
 
+## Running the RTSP sample
+
+RTSP sample will consume the video feed via specified URL. The sample is packaged with a video file streamed from `live555` service container.
+You can remove that service from `docker-compose-rtsp` if pointing the configuration in `config/analytics/pipelines-rtsp.json` to a different RTSP URL.
+
+You start the sample in a manner similar to folder watcher sample, using `docker compose -f docker-compose-rtsp.yml up -d` command to start the service.
+
+This configuration also demonstrates how to adapt the pipeline to a particular camera view. In this case, we see cars driving on the street,
+cars making the right turn from the right lane and cars making the left turn from the left lane.
+For the cars on the street we almost never see their license plates; thus it makes most sense to exclude those entirely. This is done by setting `lptSkipCarsWithoutLPs`
+to `true`. The cars in the left lane may be seen or not, depending on whether there are cars in the right lane to occlude them. We expect that no one would want a
+camera deployed in such non-deterministic way - thus assuming that monitoring the right turn lane is the purpose of this camera. Setting filters in `boxFilter-rtsp.json`
+ensures we're not going to analyze objects outside of the bottom right quadrant ROI (`lp_roiFilter` accomplishes that), and won't analyze objects below size threshold (`vehicle_SizeFilter`
+and `lp_SizeFilter` filters accomplish that).
+

deployment-examples/StandaloneSIOWithExtension/config/analytics/boxFilter-rtsp.json

@@ -0,0 +1,28 @@
+[
+  {
+    "name": "lp_roiFilter",
+    "type" : "roi",
+    "region" : [ 0.5, 0.5, 0.99999, 0.99999 ],
+    "behavior": "boxInRoi",
+    "classes": ["licenseplate"],
+    "debug": false
+  },
+  {
+    "name": "lp_SizeFilter",
+    "type": "size",
+    "subtype": "dimension",
+    "max": 0,
+    "min": 30,
+    "classes": ["licenseplate", "motorbike"],
+    "debug": false
+  },
+  {
+    "name": "vehicle_SizeFilter",
+    "type": "size",
+    "subtype": "dimension",
+    "max": 0,
+    "min": 40,
+    "classes": ["car", "bus", "truck"],
+    "debug": false
+  }
+]

deployment-examples/StandaloneSIOWithExtension/config/analytics/pipelines-rtsp.json

@@ -0,0 +1,26 @@
+{
+    "unifiedEU-US" : {
+        "pipeline"      : "./share/pipelines/VehicleAnalytics/VehicleAnalyticsRTSP.yaml",
+        "restartPolicy" : "restart",
+        "parameters" : {
+            "VIDEO_IN"  : "rtsp://live555_svc:554/Turn-02.mkv",
+            "boxFilterConfig" : "/config/analytics/boxFilter-rtsp.json",
+            "detectionModel" : "gen6es",
+            "lptFilter" : "['us']",
+            "mmcFilter" :  "['us']",
+            "lptMinConfidence" : "0.5",
+            "sourceId" : "fw-1",
+            "lptPreferAccuracyToSpeed" : "true",
+            "extensionModules" : "/config/analytics/extension.py",
+            "extensionConfigurations" : "/config/analytics/extensionConfig1.json",
+            "useTracker" : "true",
+            "updateOnlyOnChange" : "true",
+            "lptStabilizationDelay" : "5",
+            "updateOnImprovedScore" : "true",
+            "debugSettings" : "log,json",
+            "lptSkipCarsWithoutLPs" : "true"
+        },
+        "disabledParameters" : {
+        }
+    }
+}

deployment-examples/StandaloneSIOWithExtension/docker-compose-rtsp.yml

@@ -0,0 +1,43 @@
+version: "2.3"
+services:
+
+  # By default pipelines.json will point to streams served by this container.
+  # If you point to your own cameras or streams, this container has no other
+  # function and can be disabled.
+  live555_svc:
+    image: us-central1-docker.pkg.dev/ext-edge-analytics/docker/live555:2.0.4-examples
+    container_name: sample-live555
+    restart: unless-stopped
+    ports:
+      - "8554:554"
+
+
+  analytics:
+    image: us-central1-docker.pkg.dev/ext-edge-analytics/docker/sio:${SIO_RELEASE-r231120}${SIO_DOCKER_TAG_VARIANT}
+    restart: unless-stopped
+    environment:
+      # Location where SIO will place generated model engine files
+      - SIO_DATA_DIR=/data/sio-cache
+      # We need this to see output from Python extension module
+      - PYTHONUNBUFFERED=1
+    # Container runtime defaults to `runc` if SIO_DOCKER_RUNTIME not set. Use `nvidia` if GPU is installed.
+    runtime: ${SIO_DOCKER_RUNTIME-runc}
+    volumes:
+      # Read-only shared folder for data exchange with host / other containers.
+      # We'll use it for license, config files, etc.
+      - ./config:/config:ro
+      # Writable shared folder for data exchange with host
+      # We'll use it for storing the generated model files, data exchange folder, etc.
+      - ./data:/data
+    entrypoint:
+      - /sighthound/sio/bin/runPipelineSet
+      # Pipeline configuration file
+      - /config/analytics/pipelines-rtsp.json
+      # License at the path accessible in the container
+      - --license-path=/config/sighthound-license.json
+      # Log level (info, debug, trace)
+      - --log=${SIO_LOG_LEVEL-info}
+    depends_on:
+      # This dependency can be removed with live555 if no longer necessary.
+      - live555_svc
+

deployment-examples/StandaloneSIOWithExtension/docker-compose.yml

@@ -1,8 +1,8 @@
-version: "3"
+version: "2.3"
 services:
 
   analytics:
-    image: us-central1-docker.pkg.dev/ext-edge-analytics/docker/sio:r231024
+    image: us-central1-docker.pkg.dev/ext-edge-analytics/docker/sio:${SIO_RELEASE-r231120}${SIO_DOCKER_TAG_VARIANT}
     restart: unless-stopped
     environment:
       # Location where SIO will place generated model engine files

deployment-examples/VideoStreamsConsumer/README.md

@@ -28,6 +28,12 @@ If you have an NVIDIA GPU installed and properly configured, you can run the fol
 SIO_DOCKER_RUNTIME=nvidia docker compose up -d
 ```
 
+Or if you have an NVIDIA Tegra device properly configured (e/g DNNcam), you can run the following command instead to enable GPU acceleration:
+
+```bash
+SIO_DOCKER_RUNTIME=nvidia SIO_DOCKER_TAG_VARIANT="-r32.7.3-arm64v8" docker-compose up -d
+```
+
 ## Running the client sample
 
 Once the services are up and running, start the client sample with a `docker compose up` command while inside the relevant sample's folder in `./clients/python/`.

deployment-examples/VideoStreamsConsumer/docker-compose.yml

@@ -1,11 +1,11 @@
-version: "3"
+version: "2.3"
 services:
 
   # By default pipelines.json will point to streams served by this container.
   # If you point to your own cameras or streams, this container has no other
   # function and can be disabled.
   live555_svc:
-    image: us-central1-docker.pkg.dev/ext-edge-analytics/docker/live555:latest
+    image: us-central1-docker.pkg.dev/ext-edge-analytics/docker/live555:2.0.4-examples
     container_name: sample-live555
     restart: unless-stopped
     ports:
@@ -31,7 +31,7 @@ services:
 
   # The SIO analytics container, consuming the streams and analyzing them
   analytics_svc:
-    image: us-central1-docker.pkg.dev/ext-edge-analytics/docker/sio:r230908
+    image: us-central1-docker.pkg.dev/ext-edge-analytics/docker/sio:${SIO_RELEASE-r231120}${SIO_DOCKER_TAG_VARIANT}
     container_name: sample-sio
     restart: unless-stopped
     environment:

deployment-examples/VideoStreamsRecorder/docker-compose.yml

@@ -5,7 +5,7 @@ services:
   # If you point to your own cameras or streams, this container has no other
   # function and can be disabled.
   live555_svc:
-    image: us-central1-docker.pkg.dev/ext-edge-analytics/docker/live555:2.0.3-examples
+    image: us-central1-docker.pkg.dev/ext-edge-analytics/docker/live555:2.0.4-examples
     container_name: sample-live555
     restart: unless-stopped
     ports:
@@ -52,7 +52,7 @@ services:
 
   # The SIO analytics container, consuming the streams and analyzing them
   analytics_svc:
-    image: us-central1-docker.pkg.dev/ext-edge-analytics/docker/sio:r230908${SIO_DOCKER_TAG_VARIANT}
+    image: us-central1-docker.pkg.dev/ext-edge-analytics/docker/sio:${SIO_RELEASE-r231120}${SIO_DOCKER_TAG_VARIANT}
     container_name: sample-sio
     restart: unless-stopped
     environment:

examples/AqueductAPI/Dockerfile

@@ -0,0 +1,22 @@
+# Use an official Python runtime as a parent image
+FROM python:3.8-slim-buster
+
+# Working directory
+WORKDIR /app
+
+# Install dependencies
+COPY requirements.txt /app/
+RUN pip install --no-cache-dir -r requirements.txt
+
+# Copy the application into the container
+# COPY ../lib/ /app/lib
+COPY ./src/ /app/
+
+# Make port 5000 available
+EXPOSE 5000
+ENV PYTHONPATH=${PYTHONPATH}:/app/:/app/lib/
+# TERM environment variable not set.
+ENV TERM=xterm-256color
+
+# Run app.py when the container launches
+CMD ["python", "app.py"]

examples/AqueductAPI/README.md

@@ -0,0 +1,32 @@
+# Aqueduct API
+
+This is a Flask-based API to manage video analytics pipelines. The API provides functionalities to start, stop, delete, and fetch the status of pipelines.
+
+## Features
+- Start Pipeline: Initialize a new pipeline with specified parameters.
+- Stop Pipeline: Terminate an existing pipeline.
+- Delete Pipeline: Remove an existing pipeline.
+- Fetch Status: Retrieve the status of all pipelines or a single pipeline.
+
+## How to run
+First start all the dependencies:
+```bash
+./scripts sh-services apply ./examples/AqueductAPI/aqueduct.conf
+cd ./examples/AqueductAPI
+```
+
+And then build and run the Docker Compose:
+```bash
+docker compose build
+docker compose up
+```
+
+Your API should now be accessible at http://localhost:8888.
+
+## API Endpoints
+- POST `/pipelines/start`: Starts a new pipeline.
+- POST `/pipelines/stop`: Stops an existing pipeline.
+- POST `/pipelines/delete`: Deletes an existing pipeline.
+- GET `/pipelines/status`: Gets the status of all pipelines.
+- GET `/pipelines/status/<string:sourceId>`: Gets the status of a specific pipeline.
+- GET `/health`: Health check.

examples/AqueductAPI/aqueduct.conf

@@ -0,0 +1,13 @@
+select_live555_video
+disable amqp-stats
+remove_orphans
+enable live555
+enable mcp
+enable rabbitmq
+enable sio
+select_example sio aqueduct
+up live555
+up rabbitmq
+up mcp
+test_rtsp_stream rtsp://localhost:7554/data/my-video.mkv 5
+restart sio

examples/AqueductAPI/docker-compose.yml

@@ -0,0 +1,20 @@
+version: "2"
+
+services:
+  aqueduct-api:
+    build: .
+    restart: unless-stopped
+    volumes:
+      - ../lib:/app/lib
+      - ../../..:/data/sighthound
+      - ./src:/app
+    ports:
+      - "8888:8888"
+    networks:
+      core_sighthound:
+        aliases:
+          - aqueduct-api
+
+networks:
+  core_sighthound:
+    external: true

examples/AqueductAPI/requirements.txt

@@ -0,0 +1,9 @@
+Flask==1.1.2
+flask-restx==1.1.0
+flask-cors==3.0.10
+pika==1.2.0
+jinja2==2.11.2
+markupsafe==1.1.1
+itsdangerous==1.1.0
+werkzeug==2.0.3
+tabulate==0.8.9