QA: pool creation issue for txs with many transfer logs

[jmulq]

During QA, the subgraph failed at ~91% synced.

This was caused by the following:

• A deployment event contained may other event logs besides the usual unique transfer event we see relating to a pool from other deployments https://etherscan.io/tx/0xc4f29301d1db7ffaa35f9432f93965f5c5f8fc410503d77f62cadf9f8b51b9aa#eventlog
• The substream did not have granular enough criteria for extracting the specific transfer event relating to a pool. The prior logic worked for most pool deployment txs, however this edge case above was not handled. The new logic is robust enough to handle all future deployments.