polygon/bridge: fix snapshot stores, logging, replay blocks if needed on startup

[taratorio]

1. fixes a bug with snapshot store which was causing uint64 underflow (due to 0-1) because it was not filtering out snapshot segment indices that were empty (so index.BaseDataId()+index.KeyCount()-1 = 0+0-1) in LastFrozenEntityId - this caused the milestone scrapper to be stuck forever in SynchronizeMilestone
2. fixes a bug with bridge block replay on startup if it is behind current execution tip - this relates to situations that can happen after initial sync and snapshot download and/or during regeneration of bridge/heimdall snapshots from 0 as described here astrid: create integration commands to manually regenerate checkpoints, spans, events, etc. #12737 (comment) (hence also closes astrid: create integration commands to manually regenerate checkpoints, spans, events, etc. #12737)
3. improves our Info logging by introducing periodic progress logging so it doesn't look like the node is stuck/not doing anything upon initial sync on Info level