Integration tests creating existing DID's

[jamshale]

The integration tests are recreating the same did's they previously created causing fails because the schema/cred_def's they are trying to create already exist on the ledger. Should be able to make the seed/did creation more random to prevent this from happening.

[PatStLouis]

have we found in the code where the seed is generated? I was thinking in here but I couldn't see it

[ianco]

The aca-py seed is created here: https://github.com/hyperledger/aries-cloudagent-python/blob/main/demo/runners/support/agent.py#L211

It's kind of hokey and not really robust enough for running integration tests, where we may create hundreds of DID's in a single run of the tests. (It was initially implemented for a simple demo but should be updated ...)

[PatStLouis]

I can suggest to switch to python secrets.token_urlsafe to create a random seed with large entropy instead of randint.

[ianco]

I can suggest to switch to python secrets.token_urlsafe to create a random seed with large entropy instead of randint.

I'll take a look, thanks for the suggestion.

[PatStLouis]

@ianco looks like @dbluhm is addressing this in another pr with a similar approach, using token_hex instead. Less entropy but good enough!