feat/add-call-tokenuri

[TimDaub]

@Kunkka0822 why can't we be using https://github.com/neume-network/strategies/tree/main/src/strategy-factories/call-tokenuri ?

[Kunkka0822]

@Kunkka0822 why can't we be using https://github.com/neume-network/strategies/tree/main/src/strategy-factories/call-tokenuri ?

As you can see, we need more functions and in my opinion, these are not designed to be extensible.

[il3ven]

Extractor is trying to fetch all tokenURIs for all tokenIDs. Good.

It relies on the nextTokenId function to get all tokenIds. If the nextTokenId is 100, the extractor will fetch all tokenURIs from 1 to 100.

My preferred way to get all tokenIds would be call-block-logs. We can track the Transfer event.

@Kunkka0822 also tries to transform the output from extractor using a write function. Such transforms can be done in the transformer file.

The output of the write function strips all the essential metadata such as contract address and block number. This information will be useful later. The final NFT output should be according to our defined schema and it has fields such as createdAt.

I didn't understand the need of the current transformer. Instead of writing two tokenURIs such as ar://b2dea../1 and ar://b2dea../2. It only writes ar://b2dea../[1-2]. It minimises the data to be written but I don't think it is enough reason to implement it. Such notation will increase our logic in subsequent sound-protocol strategies where we have to decode this first. An example of subsequent sound-protocol strategy is sound-protocol-get-tokenuri which fetches the ar:// URI.

---

Actionable advice from my side:

• Make sure we are able to track all tokenIDs using call-block-logs. @Kunkka0822 We can discuss more on this if you have doubts.
• Once we have the tokenIDs, write a strategy to call tokenURI function on the contract. By convention, this strategy will be called sound-protocol-call-tokenuri. If nothing goes wrong this strategy should be exactly like other call-tokenuri strategies and can also use the call-tokenuri factory.

[il3ven]

I talked with @Kunkka0822 over Discord and explained my above comment. We have decided to:

• Get all tokenIds from call-block-logs and not use nextTokenId contract function. Where the output after transforming it with logs-to-subgraph will be like.

  [
  {"address":"0x3bf96afe2291d76f2934350624080faefeec9a46","tokenId":"1","createdAtBlockNumber":"13761051","platform":{"name":"sound-protocol"}},
  {"address":"0x3bf96afe2291d76f2934350624080faefeec9a46","tokenId":"2","createdAtBlockNumber":"13761051","platform":{"name":"sound-protocol"}}
  ...
  ]

• Write a sound-protocol-call-tokenuri that takes the above tokenIds as input and fetches the tokenURI using the contract function. After transformation it's output will be like

  {"metadata":{"block":{"number":"13761051"},"contract":{"address":"0x3bf96afe2291d76f2934350624080faefeec9a46"},"tokenId":"1"},"results":"ar://-bLbFg-u3BraSe7oQ90IbW3603C7VCSkS4LomfFR5V0/1"}
  {"metadata":{"block":{"number":"13761051"},"contract":{"address":"0x3bf96afe2291d76f2934350624080faefeec9a46"},"tokenId":"2"},"results":"ar://-bLbFg-u3BraSe7oQ90IbW3603C7VCSkS4LomfFR5V0/2"}

[TimDaub]

ok good makes sense