Make sure the static visualizer is working with the Python app.

Update the README and mention that Python dependencies need to be
installed before launching the visualizer.

Signed-off-by: Andreas Heumann <[email protected]>

Author: AndreasHeumann

applications/pipeline_visualization/README.md

@@ -33,7 +33,7 @@ This will start a NATS server listening on `0.0.0.0:4222`.
 
 ### Step 3: Start the Visualizer
 
-In a second terminal, start one of the Python visualizers:
+In a second terminal, make sure the [visualizer Python dependencies](#visualizer-python-dependencies) are installed and start one of the Python visualizers:
 
 ```bash
 cd applications/pipeline_visualization/visualizer
@@ -44,7 +44,7 @@ The web interface will be available at: **[http://localhost:8050](http://localho
 
 ### Step 4: Run the Holoscan Application
 
-In a third terminal, run the C++ application:
+In a third terminal, run the application:
 
 ```bash
 # Run the Python version (default when --language is not specified)
@@ -185,9 +185,11 @@ The data format is defined using FlatBuffers for efficient serialization:
 
 ## Prerequisites
 
-All dependencies are installed automatically when using the `holohub run` command inside the Holohub container.
+All dependencies to run the application are installed automatically when using the `holohub run` command inside
+the Holohub container. Vince the visualizer is run outside of the Holohub container, its dependencies have
+to be installed separately. See next section for details.
 
-### Python Dependencies
+### Visualizer Python Dependencies
 
 Install the required Python packages:
 

applications/pipeline_visualization/visualizer/visualizer_static.py

@@ -137,9 +137,13 @@ def update_source(n_intervals, subject):  # noqa: ARG001
 
                 # Identify which graph this data belongs to
                 unique_id = message.UniqueId().decode()
-                try:
-                    index = self._unique_ids.index(unique_id)
-                except ValueError:
+                for i, id in enumerate(self._unique_ids):
+                    # the python application is using tensor maps where the unique_id also contains
+                    # the tensor name, so we need to check if the unique_id starts with the id
+                    if unique_id.startswith(id):
+                        index = i
+                        break
+                else:
                     logger.warning(f"Unknown unique_id: {unique_id}")
                     continue