The mappings that you write will perform transformations on the Ethereum data you are sourcing, and it will dictate how this data is loaded into the Graph Node. Mappings can be very simple but can become complex. It depends on how much abstraction you want between the data and the underlying Ethereum contract.

I do not think anything is limiting the complexity. It seems like anything is possible. Or?

From what I have found there are definitely things that aren’t possible.

I’ll copy/paste these questions into a doc and we can go into it over a call. In order to give a good response I’d need more time and to write some pseudo code to show the holes in the flow. I’ll write something up as soon as I can (gotta finish a task when this meeting ends at work so can’t right now)

Okay so here a quick response while i have a few minutes freed up:

• I do not think that walking/correlating data within two different The Graph subgraphs is possible without an added semantic definition layer on top. For example, you couldn’t query for “all votes by Fred in all Aragon & DAOstack DAOs”. In order to understand both, you’d need to “hard code” in support for both. If DAOstack & Aragon both implement the same DAO semantic definition, you could do this.

In reply to @ontologymachine:matrix.org
so you’re saying it’s incompatible with emergent/generative definitions bc it would require top-down definitions c

“Wyre is committed to helping decentralized applications to maintain on-chain compliance, enabling your users to interact frictionlessly with the fiat world.”

https://www.sendwyre.com/partner/