Implement a way to generate a derived dataset from an already-discovered survey-related record.
Having already been discovered, the survey-related record will have at least one asset. The idea is to use this asset and convert it to a low-resolution version of itself, one which would be easy to show in the catalog UI and which can provide users with a quick overview of the data - the focus here is really on having a visual representation of the data that is fast to load.
Investigate on the most appropriate way to store this derived dataset and how it can be consumed by the frontend.
As an example:
Starting form a kmall file with raw bathymetry data, that has thousands of vector points, perform a random sample and keep just 10% of the points. Then use an IDW interpolation to produce a raster surface of the area and keep that as the low-res dataset.
This derived dataset is to then be saved as an additional asset of the survey-related record.
Let's implement a prototype of this in order to understand how it can be made to work with a single file type first, and then think about extending to other types.
Important points to keep in mind:
- A single file can potentially be used to generate more than one low-res dataset (ex: generate a rough raster surface from sampled vector points, generate a chart that shows an histogram of the vector points, etc)
- Users control when this derived asset is generated - this is not to be generated automatically
- The generation of derived assets must be idempotent - it should generate the same result if called multiple times, without causing any additional side-effects
- Generation is always going to be called as a background task
- The resulting dataset is to be bound to the original survey-related record as an additional asset.
Implement a way to generate a derived dataset from an already-discovered survey-related record.
Having already been discovered, the survey-related record will have at least one asset. The idea is to use this asset and convert it to a low-resolution version of itself, one which would be easy to show in the catalog UI and which can provide users with a quick overview of the data - the focus here is really on having a visual representation of the data that is fast to load.
Investigate on the most appropriate way to store this derived dataset and how it can be consumed by the frontend.
As an example:
Starting form a kmall file with raw bathymetry data, that has thousands of vector points, perform a random sample and keep just 10% of the points. Then use an IDW interpolation to produce a raster surface of the area and keep that as the low-res dataset.
This derived dataset is to then be saved as an additional asset of the survey-related record.
Let's implement a prototype of this in order to understand how it can be made to work with a single file type first, and then think about extending to other types.
Important points to keep in mind: