runtime: remove some Emerald references

[pro-wh]

constructing a runtime analyzer now takes an analyzer.Runtime parameter

cross reference #305

last (?) step of https://app.clickup.com/t/3u9y285

[mitjat]

Hah I have an almost identical commit on my machine :)
Thank you!
LGTM pending the unresolved comments.

[mitjat]

As per the comment https://github.com/oasisprotocol/oasis-indexer/pull/306/files#r1095643489, here are the two tests that I ran to verify that this PR basically resolves https://app.clickup.com/t/3u9y285 and makes us ready to use multiple analyzers.

In both tests, I ran reindex_and_run.sh regression test with the config posted below, once with sapphire commented out and once present.

Test 1 - config:
Both analyzers scan the same blocks

  analyzers:
    emerald:
      from: 1_003_298  # round at Damask genesis
      to: 1_003_308  # 10 blocks
    sapphire:
      from: 1_003_298  # round at Damask genesis
      to: 1_003_308  # 10 blocks

No differences in emerald queries on the resulting DB. This strongly suggests no erroneous double dead-reckoning, or similar overlap between analyzers' output.

Test 2 - config:
Analyzers overlap in half of the blocks

  analyzers:
    emerald:
      from: 1_003_298  # round at Damask genesis
      to: 1_003_318  # 20 blocks
    sapphire:
      from: 1_003_308  # emerald + 10
      to: 1_003_328  # emerald + 10

No difference in any queries, except /consensus/accounts which presents a union of all true consensus accounts and accounts encountered in runtimes (via the address_preimages table). This makes sense, as the "sapphire" analyzer scanned 10 extra blocks, which involved not-seen-before eth addresses.

[pro-wh]

thanks for the testing!