Integrate with external Guardian policies as data sources

[anvabr]

Problem description

Executing Guardian Policy workflow results in a production of artefacts, namely VCs and VPs, and ultimately minted tokens. Each of these artefacts is currently populated with relevant information as directed by the active Policy on the Guardian instance. Currently there is no way to use the output of one policy as an input into another policy, in particular in the automated fashion.

Note: #1515 covers the basic data enrichment functionality where additional info can be inserted into VCs as 'external evidence'. In the contrast, this ticket covers the 'Policy chaining' functionality.

Requirements

• Support the use, from inside the policy, of VCs/VPs produced by other Guardian instances published in IPFS.
• Enable users to manually input the link or other ID of the document to be used by the policy as input.
• Allow users to setup their 'secondary' guardian instance[s] in such a way that their continuation of the active policy execution gets automatically triggered (unlocked) by the appearance of the 'target' artefact which is produced by another policy/guardian instance.
• Enable unlimited depth of Guardian policy chains, mixed with auto-execution and manual specification of the 'input' artefacts.
• Input documents' trust chain should be shown (to the extend possible) in the policy trust chain as well as the identity/credentials of the source, time/date and other identifying information as appropriate.

Definition of done

• Functionality implemented as per the requirements above
• Documentation created
• A functioning example of policy chaining is added to the built-in policies

Acceptance criteria

• Out to the box Guardian supports integration with other Guardian instances, policy authors should be able to add this to the relevant stages of their policy workflow.
• Users at the policy execution time should be able to specify the source (input) documents based on various criteria, as well there should be an option for policy author to 'hard code' the policy input sources to be used by default or as mandatory.

[anvabr]

In the minimal design there are two policies executed in two separate Guardian instances: 1st policy - the producer of the artefacts (VCs) and the 2nd policy - the consumer of these artefacts and the producer of tokens. The 1st policy could be producing multiple types of VCs, the 2nd policy will be consuming a subset of these (all documents of a particular type), and use them as per (2nd) policy flow to produce tokens.

In general case the policies can be developed independently, the 2nd policy can allow users (at run time) to select which document (from what source) matching which schemas should be picked up by it and processed further. The 2nd policy engine is constantly 'listening' for the corresponding messages from the source (in its topic) and pick up documents as they appear.

This is displayed diagrammatically below.