multiformats
This library is for building an interface for working with various
inter-related multiformat technologies (multicodec, multihash, multibase,
and CID).
The interface contains all you need for encoding and decoding the basic
structures with no codec information, codec encoder/decoders, base encodings
or hashing functions. You can then add codec info, codec encoders/decoders,
base encodings, and hashing functions to the interface.
This allows you to pass around an interface containing only the code you need
which can greatly reduce dependencies and bundle size.
import { create } from 'multiformats'
import sha2 from 'multiformats/hashes/sha2'
import dagcbor from '@ipld/dag-cbor'
const { multihash, multicodec, CID } = create()
multihash.add(sha2)
multicodec.add(dagcbor)
const buffer = multicodec.encode({ hello, 'world' }, 'dag-cbor')
const hash = await multihash.hash(buffer, 'sha2-256')
const cid = new CID(1, 'raw', hash)
However, if you're doing this much you should probably use multiformats
with the Block
API.
import multiformats from 'multiformats/basics'
import dagcbor from '@ipld/dag-cbor'
import { create } from '@ipld/block'
multiformats.multicodec.add(dagcbor)
const Block = create(multiformats)
const block = Block.encoder({ hello: world }, 'dag-cbor')
const cid = await block.cid()
Plugins
By default, no base encodings, hash functions, or codec implementations are included with multiformats
.
However, you can import the following bundles to get a multiformats
interface with them already configured.
bundle | bases | hashes | codecs |
---|
multiformats/basics | base32 , base64 | sha2-256 , sha2-512 | json , raw |
Base Encodings (multibase)
Hash Functions (multihash)
hashes | import | repo |
---|
sha2-256 , sha2-512 | 'multiformats/hashes/sha2' | multiformats/js-multiformats |
sha3-224 , sha3-256 , sha3-384 ,sha3-512 , shake-128 , shake-256 , keccak-224 , keccak-256 , keccak-384 , keccak-512 | '@multiformats/sha3' | multiformats/js-sha3 |
Codec Implementations (multicodec)
API
multiformats([table])
Returns a new multiformats interface.
Can optionally pass in a table of multiformat entries.
multihash
multihash.encode
multihash.decode
multihash.validate
multihash.add
multihash.hash
multicodec
multicodec.encode
multicodec.decode
multicodec.add
multibase
multibase.encode
multibase.decode
multibase.add
CID
Changes from cids
:
- All base encodings are cached indefinitely.
- CID's can be created without any multiformat data.
- The new API is entirely based on parsing the varints
so it doesn't need the table metadata in order to associate
string names.
There are also numerous deprecations. These deprecations all stem from the
fact that we no longer know the full set of available multicodec information.
It's actually quite possible to provide a CID interface without this, you can
still do everything you used to do, you just need to use ints instead of strings
and do some of the fancier V0 coercions outside this library.
Deprecation List:
- the multibase encoding is no longer cached during instantiation.
- this being indeterministic was causing some nasty problems downstream
since
toString()
needs to be used as a cache key and it's not possible
to encode V0 into anything but base58btc. this means that you can't have
deterministic hash keys without also requiring base58btc support, so we
removed this feature.
- no longer accept bare multihash buffers as V0, only base58btc encoded strings
or the complete set of parts with the multihash as the third argument.
- The logic for treating these as V0 was problematic and could lead to
unwanted errors. In practice, the places we need to handle V0 are known
and we can move this logic there if necessary and then instantiate a CID
with all the necessary args.
- Some of the prior code relied on knowing the full set of possible base
prefixes, which we can no longer do.
- no more .toBaseEncodedString(), just toString()
- no more .multibaseName
- no more .prefix()
- no more .codec
- new property ".code" is the multiformat integer.
- this is going to be a painful transition but we have to get off of using
the string if we ever want to drop the full table. while the DX for this is
nice it forces into bloating the bundle and makes using new codecs very
painful.