fix: use new retrieval code path for indexers

[jacobheun]

Summary

This change ensures that requests made via the network indexers, like cid.contact, are routed through the simplified graphsync retrieval pathways. Currently ingestion will fallback to legacy retrievals which is prone to stalling, and is less performant.

Verification with a local STI indexer

I verified the ingestion pathways succeeded locally by installing a local instance of StoreTheIndex.

# Install STI
go install github.com/ipni/storetheindex@latest
# Initialize the daemon
storetheindex init
# Remove the bootstrap nodes in the config to avoid connecting to mainnet
contents="$(jq '.Bootstrap.Peers = []' ~/.storetheindex/config)" && echo -E "${contents}" > ~/.storetheindex/config
# Start the daemon
storetheindex daemon
# Trigger a sync against the target Boostd instance
storetheindex admin sync --addr /ip4/127.0.0.1/tcp/50000/p2p/$PEER_ID --pubid $PEER_ID

You should then see logs in the daemon. Once ingested you can via records in the indexer:

• Ingested Providers: http://localhost:3000/providers
• Specific CIDs: http://localhost:3000/cid/$CID

Example output

$ provider ls ad --provider-addr-info=/ip4/127.0.0.1/tcp/50000/p2p/12D3KooWPGvdyREpGXC7BLj1idgLNFb8hLjHnw56hxp7nKQkcJQd --print-entries --topic=/indexer/ingest/localnet-0ac799ad-8fe5-484e-b345-2403e966897d
ID:           baguqeerax65mruyywqevq5czkzlv3ftef3gtnlxje62i2h4qmbkizzc5haia
PreviousID:   baguqeera52qpcvtzqmaenpf3qje4z5aefxaytrt63v3pw4wsw4cws2nhg4ca
ProviderID:   12D3KooWPGvdyREpGXC7BLj1idgLNFb8hLjHnw56hxp7nKQkcJQd
ContextID:    
Addresses:    [/ip6/::1/tcp/50000 /ip4/192.168.1.17/tcp/50000 /ip4/127.0.0.1/tcp/50000]
Is Remove:    false
Metadata :    
Extended Providers:
   Override: false
   Providers:
      None
Entries:
  ---------------------
  Chunk Count: 0
  Total Count: 0

Refs

Fixes #1432