Hey, I am using Roboflow Workflows and am currently trying to save my video data using data aggregation. The problem is that there is no save option in the block itself and when using other blocks like local save etc. data aggregation cant be chosen / is invalid to be chosen as the source. How do I save the data from this block / where is it going?
Hi @Moon,
When using data aggregator you can place csv formatter behind it and then locsl file sink
I have built simple example:
Here is JSON:
{
"version": "1.0",
"inputs": [
{
"type": "InferenceImage",
"name": "image"
}
],
"steps": [
{
"type": "YoloWorldModel",
"name": "model_1",
"images": "$inputs.image",
"class_names": [
"person"
]
},
{
"type": "roboflow_core/bounding_box_visualization@v1",
"name": "bounding_box_visualization",
"image": "$inputs.image",
"predictions": "$steps.model_1.predictions"
},
{
"type": "roboflow_core/data_aggregator@v1",
"name": "data_aggregator",
"data": {
"foo": "$steps.model_1.predictions"
},
"interval": 1,
"data_operations": {
"foo": []
},
"interval_unit": "minutes",
"aggregation_mode": {
"foo": [
"count"
]
}
},
{
"type": "roboflow_core/csv_formatter@v1",
"name": "csv_formatter",
"columns_data": {
"foo": "$steps.data_aggregator.foo_count"
},
"columns_operations": {
"foo": [
{
"type": "ToString"
}
]
}
},
{
"type": "roboflow_core/local_file_sink@v1",
"name": "local_file_sink",
"content": "$steps.csv_formatter.csv_content",
"output_mode": "append_log",
"target_directory": "/tmp/",
"file_name_prefix": "my_workflow_output",
"max_entries_per_file": 10000
}
],
"outputs": [
{
"type": "JsonField",
"name": "output",
"coordinates_system": "own",
"selector": "$steps.bounding_box_visualization.image"
}
]
}
In this example files will be stored in files named like this: /tmp/my_workflow_output_*
