Mathijs de Bruin

Due to large amount of no-ops in backend, it seems not to be the case.

Publish metadata through HAMT on IPFS

3

For direct acccess. Possibly strategies: a) Use built-in experimental IPFS HAMT, which works transparently. b) Use [IPLD-HAMT](https://ipld.io/specs/advanced-data-layouts/hamt/spec/), which is in progress. Ref: #183

https://docs.ipfs.io/how-to/best-practices-for-nft-data/#metadata

The more timeouts, the more parallel fetches we want to do (up until a certain maximum, of course).

so we can fine-tune the kind of requests that IPFS gets of various types - e.g. we can handle a lot of Stat() calls, but much less Ls() calls and...

Everything else is a server error and is very unlikely to spontaneously succeed the next time. Later: re-enable all retries and max retry limit through nsq.

https://godoc.org/github.com/ipfs/go-ipfs-http-api Offers streaming, context closing and full IPFS API compatibility (eventually).

``` Apr 14 18:17:04 boole ipfs-crawler[32158]: [119B blob data] Apr 14 18:17:04 boole ipfs-crawler[32158]: [126B blob data] Apr 14 18:17:04 boole ipfs-crawler[32158]: [139B blob data] Apr 14 18:17:04 boole ipfs-crawler[32158]:...

Separate for IPFS/Tika and Elasticsearch.