argahsuknesib
commented
9 months ago Update on this.
The type index spec (https://solid.github.io/type-indexes/#introduction) describes the dataset (such as the LDES) with the data it contains. The flow to solve the issue can be, the replayer replays the dataset and then does a POST request to the public or private type index of the solid pod with metadata of the data replayed so that another agent (such as the aggregator) can use it.
What do you think? @svrstich I will implement this in the replayer.
In the current implementation, we replay the sensor events to a particular LDP location which contains the LDES. As we are the ones replaying, we know the location of the LDES already.
However, a client would not know by default where the LDES is stored, and what kind of data it contains. As it would query the pod directly rather than querying the LDES, there is a need for a description of the re-player somewhere on the Solid Pod such that a client when querying the pod can navigate and choose the relevant LDES stream required to compute the query with.
For example, The client does not know where (i.e which LDP Container) contains the heart rate of the user. In my implementation of the Solid Stream Aggregator's Bench-marking, I am assuming the information to be present at the public /profile/#card document for now (which can be changed, as it has it's own problems regarding privacy of the data)
How do you think we should approach the problem?