Spark Piece Indexer: find piece info in ContextID

[bajtos]

Rework spark-piece-indexer as follows:

• Ingest only advertisements where ContextID contains Piece Info
• The REST API for querying should accept ContextID instead of PieceCID

How does Curio map Piece Info to ContextID:

• The PieceInfo is a structure defined as {PieceCID, PieceSize}
• The piece info is serialised using DAG-CBOR and base64 encoded

Example IPNI advertisement:
https://http.f02620.devtty.eu/ipni-provider/12D3KooWBpYWsD4DuzTTVQdS6N5mt6rc5HtY11p7KLfhyVLw83d6/ipni/v1/ad/baguqeera6g3t7nianlk2kplejqyzl5wvbaq6h4vdqrhhgm2ctid7q4sgcx4a

{
  "Addresses":["/dns/cid.contact/https/http-path/%2Fingest%2Fannounce"],
  "ContextID":{"/":{"bytes":"ghsAAAAIAAAAANgqWCgAAYHiA5IgIPxmN381s3gKZTaNDaP+GGLv+fs22x1BbOe5TC4QYugO"}},
  // etc.
}

Decoded ContextID (via https://cbor.me):

82                                      # array(2)
   1B 0000000800000000                  # unsigned(34359738368)
   D8 2A                                # tag(42)
      58 28                             # bytes(40)
         000181E203922020FC66377F35B3780A65368D0DA3FE1862EFF9FB36DB1D416CE7B94C2E1062E80E # 

Node.js script to inspect and reconstruct the ContextID:

import { encode, decode } from '@ipld/dag-cbor'
import { CID } from 'multiformats'

const contextId = 'ghsAAAAIAAAAANgqWCgAAYHiA5IgIPxmN381s3gKZTaNDaP+GGLv+fs22x1BbOe5TC4QYugO'
console.log('Input ContextID', contextId)

const contextIdBytes = Buffer.from(contextId, 'base64')
const data = decode(contextIdBytes)
console.log('decoded', data)

console.log('PieceSize', data[0].toString())
console.log('PieceCID', data[1].toString())

const contextId2Bytes = encode([
  Number(data[0].toString()),
  CID.parse(data[1].toString())
])
const contextId2 = Buffer.from(contextId2Bytes).toString('base64')
console.log('Reconstructed ContextID', contextId2)
console.log('Match?', contextId === contextId2)