fix(tap-agent): resolve V2 actor system and collection ID parsing issues

This commit addresses multiple issues that were preventing the V2 TAP actor
system from functioning properly:

1. **Actor Race Condition Prevention**: Added existence check in SenderAccount
   to prevent duplicate actor creation attempts that caused "Actor already
   registered" errors during concurrent receipt processing.

2. **Collection ID Parsing**: Fixed parsing of collection_id strings from
   database during migration period:
   - Handle both 20-byte addresses and 32-byte collection IDs
   - Proper whitespace trimming and error reporting
   - Consistent 0x prefix handling for both Address and CollectionId parsing

3. **Attestation Error Messages**: Updated error messages to be factual about
   what is observed rather than speculative about causes when key derivation
   fails after testing multiple combinations.

4. **BIP32 Path Length Handling**: Maintained existing v1/v2 fallback system
   for handling long deployment IDs that exceed BIP32 path limits.

The V2 TAP actor system now starts successfully and can process receipts.
RAV generation is pending allocation closure in the test environment.

🤖 Generated with [Claude Code](https://claude.ai/code)

Co-Authored-By: Claude <[email protected]>

Author: suchapalaver

Cargo.lock

@@ -11,7 +11,6 @@ status_url = "http://graph-node:8030/graphql"
 
 [subgraphs.network]
 query_url = "http://graph-node:8000/subgraphs/name/graph-network"
-deployment_id = "NETWORK_DEPLOYMENT_PLACEHOLDER"
 syncing_interval_secs = 5
 recently_closed_allocation_buffer_secs = 10
 

contrib/indexer-service/config.toml

@@ -6,7 +6,9 @@ edition = "2021"
 [dependencies]
 anyhow.workspace = true
 indexer-allocation = { path = "../allocation" }
+sha2 = "0.10"
 thegraph-core.workspace = true
+tracing.workspace = true
 
 
 [dev-dependencies]