Refactor Gradio Application (#328)

* Show Live Telemetry while Benchmark is running
* Use Coming Soon for pipelines under development
* Use Intel Digital Library Images for Thumbnails
* Update SmartNVR thumbnail to Asset ID a1187747 
* Update Transportation2 thumbnail to Asset ID a1427620
* Add Pipeline information on Run tab
* Suppress warning and logs
* Apply black format to python files

Co-authored-by: oommensy <[email protected]>

Author: dnoliver

tools/visual-pipeline-and-platform-evaluation-tool/app.py

@@ -33,10 +33,10 @@ services:
     environment:
       - MODELS_PATH=/output/
     user: root
-    command: bash -xc '
+    command: bash -c '
         cp /tmp/models.sh . &&
         chmod +x models.sh &&
-        bash -cx ./models.sh &&
+        bash -c ./models.sh &&
         tail -f /dev/null
       '
     restart: on-failure:5

tools/visual-pipeline-and-platform-evaluation-tool/compose.yml

@@ -1,6 +1,6 @@
 #/bin/bash
 
-set -euxo pipefail
+set -euo pipefail
 
 # Remove healthcheck.txt if it exists
 rm -f healthcheck.txt

tools/visual-pipeline-and-platform-evaluation-tool/models.sh

@@ -3,4 +3,5 @@ definition: >
   The Smart Network Video Recorder (NVR) Proxy Pipeline builds on top of the
   Simple NVR Proxy Pipeline to add support for video analytics on a subset of input channels.
 metadata:
-  classname: SmartNVRPipeline
+  classname: SmartNVRPipeline
+  enabled: true

tools/visual-pipeline-and-platform-evaluation-tool/pipelines/smartnvr/config.yaml

@@ -3,4 +3,5 @@ definition: >
   The Transportation Proxy Pipeline 2 (D-C) collects road data to understand road usage,
   and collects additional data for tracking vehicles.
 metadata:
-  classname: Transportation2Pipeline
+  classname: Transportation2Pipeline
+  enabled: false

tools/visual-pipeline-and-platform-evaluation-tool/pipelines/smartnvr/thumbnail.png

@@ -18,18 +18,7 @@
 
 
 def prepare_video_and_constants(
-    input_video_player,
-    object_detection_model,
-    object_detection_device,
-    object_detection_batch_size,
-    object_detection_nireq,
-    object_detection_inference_interval,
-    object_classification_model,
-    object_classification_device,
-    object_classification_batch_size,
-    object_classification_nireq,
-    object_classification_inference_interval,
-    object_classification_reclassify_interval,
+    **kwargs: dict[str, any],
 ):
     """
     Prepares the video output path, constants, and parameter grid for the pipeline.
@@ -42,6 +31,27 @@ def prepare_video_and_constants(
     Returns:
         tuple: A tuple containing video_output_path, constants, and param_grid.
     """
+
+    # Collect parameters from kwargs
+    input_video_player = kwargs.get("input_video_player", "")
+    object_detection_model = kwargs.get("object_detection_model", "")
+    object_detection_device = kwargs.get("object_detection_device", "")
+    object_detection_batch_size = kwargs.get("object_detection_batch_size", 1)
+    object_detection_inference_interval = kwargs.get(
+        "object_detection_inference_interval", 0.0
+    )
+    object_detection_nireq = kwargs.get("object_detection_nireq", 1)
+    object_classification_model = kwargs.get("object_classification_model", "")
+    object_classification_device = kwargs.get("object_classification_device", "")
+    object_classification_batch_size = kwargs.get("object_classification_batch_size", 1)
+    object_classification_inference_interval = kwargs.get(
+        "object_classification_inference_interval", 0.0
+    )
+    object_classification_reclassify_interval = kwargs.get(
+        "object_classification_reclassify_interval", 0.0
+    )
+    object_classification_nireq = kwargs.get("object_classification_nireq", 1)
+
     random_string = "".join(random.choices(string.ascii_lowercase + string.digits, k=6))
     video_output_path = input_video_player.replace(
         ".mp4", f"-output-{random_string}.mp4"
@@ -52,6 +62,10 @@ def prepare_video_and_constants(
     if os.path.exists(video_output_path):
         os.remove(video_output_path)
 
+    # Reset the FPS file
+    with open("/home/dlstreamer/vippet/.collector-signals/fps.txt", "w") as f:
+        f.write(f"0.0\n")
+
     param_grid = {
         "object_detection_device": object_detection_device.split(", "),
         "object_detection_batch_size": [object_detection_batch_size],
@@ -202,7 +216,7 @@ def run_pipeline_and_extract_metrics(
         )
 
         # Log the command
-        logger.info(f"Pipeline Command: {_pipeline}")
+        logger.debug(f"Pipeline Command: {_pipeline}")
 
         try:
             # Set the environment variable to enable all drivers