Single source of truth for token metadata

[gkaracha]

In principle, all configuration should come from checker.yaml. Currently we have a mix of hard-wired and configurable information, which can easily get out of sync. Taken from https://github.com/tezos-checker/checker/pull/290#issuecomment-964113677:

Also, while developing this I came across some issues due to the way we currently deal with a couple token ids:

• tokens.collateral.token_id (in checker.yaml) and TezWrapper.tez_token_id must be the same when collateral = TEZ.
• tokens.cfmm_token.token_id (in checker.yaml) and Wctez.wctez_token_id must be the same when collateral = TEZ. Currently TezWrapper.tez_token_id and Wctez.wctez_token_id are set manually which is a problem (when they get out of sync Checker gets into an inconsistent state and the e2e tests fail).

Let's have all token information come from checker.yaml, even for those tokens that are conditionally generated (both wctez and tezWrapper are needed when collateral = TEZ, which is the only case at the moment).

[gkaracha]

mockFA2 (#288) might also fall into this category, even if it is to be used only for testing purposes.

[gkaracha]

After having another look at this issue I think we need the following to be able to close the issue:

• [x] Remove TezWrapper.tez_token_id, Wctez.wctez_token_id, and MockFA2.token_id and instead have them generated at build time from the checker.yaml file (tokens.issued.wtez.token_id, tokens.issued.wctez.token_id, and tokens.issued.mock_fa2.token_id, respectively).
• [x] We don't want to have the corresponding files be turned into jinja templates, so probably we should introduce something like src/tokenMetadata.ml which is generated at build time and have them all be available there.
• [x] The e2e tests also seem to have some remaining hardcoded token_ids; these should also be taken from the yaml file.