How exactly does ipfs cat method find and display contents of files using a CID by making use of DHT?

I have done a lot of research on the internet to learn how exactly ipfs cat and get methods work find and download files from other peers using a CID. I want to fully understand how this process works: "The cat method first searches your own node for the file requested, and if it can't find it there, it will attempt to find it on the broader IPFS network(https://proto.school/regular-files-api/04)".

This is the ipfs source code for cat:

async function * cat (ipfsPath, options = {}) {
    ipfsPath = normalizeCidPath(ipfsPath)

    if (options.preload !== false) {
      const pathComponents = ipfsPath.split('/')
      preload(CID.parse(pathComponents[0]))
    }

    const file = await exporter(ipfsPath, repo.blocks, options)

    // File may not have unixfs prop if small & imported with rawLeaves true
    if (file.type === 'directory') {
      throw new Error('this dag node is a directory')
    }

    if (!file.content) {
      throw new Error('this dag node has no content')
    }

    yield * file.content(options)
  }

I deduce that two important arguments that allow for peer routing and file fetching are repo.blocks and preload. repo.blocks is created during ipfs.create() and then passed as a parameter to ipfs.createCat() which is the method that actually creates the cat method. preload is also created by ipfs.create() and passed as an argument to ipfs.createCat() so that it can be used in ipfs.cat(). What confuses me the most is which one of preload or repo.blocks is actually responsible for CID querying. I analyzed the underlying methods for this part of cat:

const pathComponents = ipfsPath.split('/')
      preload(CID.parse(pathComponents[0]))

and learned that this is the part of ipfs.cat that makes http connections to other peers. However, this part:

const file = await exporter(ipfsPath, repo.blocks, options)

includes sub-methods like

const block = await blockstore.get(cid, options);
const node = dagPb.decode(block);

which also seem to be related to CID querying through use of distributed hash tables. blockstore.get did not make use of any methods that seemed to connect to other peers or search for peers that have a CID, but I am still very confused on whether these methods have any relation to CID querying. I highly appreciate any help on how the cat method works under the hood from someone who is an expert in ipfs or at least resources I can use to learn the material myself.