The goal of this issue is to identify all the things we need to do to properly provide a "Data Use Policy" (which I'm going to abbreviate as DUP) to users who would download data from our system, keeping in mind that each "data source" may have their own policy and may require us to have a user acknowledge and agree to their policy prior to downloading data from our system.

During the meeting with kooper and brock we talked about the DUP that Ameriflux requires. We have a task for IMLCZO to prompt the user to agree to IMLCZO's DUP. That issue addresses IMLCZO's DUP, but we started to see that it doesn't take into account requirements that the original source of the data might also require, and we asked, "What if a user is downloading data from multiple sources, each with their own DUP?"

We can imagine an easy way to prompt the user about the IMLCZO DUP, but when a user clicks "Download as CSV" that queries the "datapoints" endpoint. We asked, "Where in that process will we even know that Ameriflux has a DUP requirement?"

We first talked about where to store a link to the DUP. "Should it be stored in a stream or a sensor?" Rob posed this scenario that I think makes for a good example:

Imagine I have an EPA instrument out in the field and a USGS instrument right next to it. Even if they are side by side, at the same location, we would still create these as two separate sensors from the two separate data sources in the API, so storing the "data_use_policy_url" in the sensor seems the most logical.

That's a good example, because I can imagine a scenario where I'd have the same example, but two instruments by EPA and one by USGS. In that case, I'd create one sensor for EPA, and one stream for each instrument. The sensor would have the "data_use_policy_url" and the streams could each have a "source_url" that may or may not be unique.

Afterwards, I am looking at how we have "source" in our API's at the moment, and there are cases where the "source" URL is on each datapoint. I'd suggest that moving forward that should be on the stream, but I am pretty sure that the graphs in the Time Series View look for the "source" value in the datapoints right now. It should instead be checking for this in the stream.