@ethereumjs/trie
The modified Merkle Patricia tree (trie) provides a persistent data structure to map between arbitrary-length binary data (byte arrays). It is defined in terms of a mutable data structure to map between 256-bit binary fragments and arbitrary-length binary data. The core of the trie, and its sole requirement in terms of the protocol specification, is to provide a single 32-byte value that identifies a given set of key-value pairs.
Installation
To obtain the latest version, simply require the project using npm
:
npm install @ethereumjs/trie
Usage
This class implements the basic Modified Merkle Patricia Trie in the Trie
base class, which you can use with the useKeyHashing
option set to true
to create a trie which stores values under the keccak256
hash of its keys (this is the Trie flavor which is used in Ethereum production systems).
Checkpointing functionality to Trie
through the methods checkpoint
, commit
and revert
.
It is best to select the variant that is most appropriate for your unique use case.
Initialization and Basic Usage
import { Trie } from '@ethereumjs/trie'
import { bytesToUtf8, MapDB, utf8ToBytes } from '@ethereumjs/util'
const trie = new Trie({ db: new MapDB() })
async function test() {
await trie.put(utf8ToBytes('test'), utf8ToBytes('one'))
const value = await trie.get(utf8ToBytes('test'))
console.log(value ? bytesToUtf8(value) : 'not found')
}
test()
Use with static constructors
.create()
import { Trie } from '@ethereumjs/trie'
import { bytesToUtf8, utf8ToBytes } from '@ethereumjs/util'
const trie = await Trie.create()
async function test() {
await trie.put(utf8ToBytes('test'), utf8ToBytes('one'))
const value = await trie.get(utf8ToBytes('test'))
console.log(value ? bytesToUtf8(value) : 'not found')
}
test()
When the static Trie.create
constructor is used without any options, the trie
object is instantiated with defaults configured to match the Ethereum production spec (i.e. keys are hashed using SHA256). It also persists the state root of the tree on each write operation, ensuring that your trie remains in the state you left it when you start your application the next time.
.createTrieFromProof()
import { Trie } from '@ethereumjs/trie'
async function test() {
const proof = await someOtherTrie.createProof(someKey)
const trie = await Trie.createTrieFromProof(proof, { useKeyHashing: true })
const otherProof = await someOtherTrie.createProof(someOtherKey)
await trie.updateTrieFromProof(otherProof)
const value = await trie.get(someKey)
console.log(value)
const otherValue = await trie.get(someOtherKey)
console.log(otherValue)
}
test()
When the Trie.createTrieFromProof
constructor is used, it instantiates a new partial trie based only on the branch of the trie contained in the provided proof.
Walking a Trie
Starting with the v6 release there is a new API for walking and iterating a trie by using an async walk generator, which now enables to walk tries without altering the walk controller and also now enables to walk a sparse (not completely filled) trie.
The new walk functionality can be used like the following:
import { Trie } from '@ethereumjs/trie'
const trie = await Trie.create()
const walk = trie.walkTrieIterable(trie.root())
for await (const { node, currentKey } of walk) {
}
Trie
Configuration Options
Database Options
The DB
opt in the TrieOpts
allows you to use any database that conforms to the DB
interface to store the trie data in. We provide several examples for database implementations. The level.js example is used in the ethereumjs client
while lmdb.js is an alternative implementation that uses the popular LMDB as its underlying database.
If no db
option is provided, an in-memory database powered by a Javascript Map will fulfill this role (imported from @ethereumjs/util
, see mapDB module).
If you want to use an alternative database, you can integrate your own by writing a DB wrapper that conforms to the DB
interface (in @ethereumjs/util
). The DB
interface defines the methods get
, put
, del
, batch
and copy
that a concrete implementation of the DB
interface will need to implement.
LevelDB
As an example, to leverage LevelDB
for all operations then you should create a file with the following implementation from our recipes in your project. Then instantiate your DB and trie as below:
import { Trie } from '@ethereumjs/trie'
import { Level } from 'level'
import { LevelDB } from './your-level-implementation'
const trie = new Trie({ db: new LevelDB(new Level('MY_TRIE_DB_LOCATION')) })
Node Deletion (Pruning)
By default, the deletion of trie nodes from the underlying database does not occur in order to avoid corrupting older trie states (as of v4.2.0
). Should you only wish to work with the latest state of a trie, you can switch to a delete behavior (for example, if you wish to save disk space) by using the useNodePruning
constructor option (see related release notes in the changelog for further details).
Root Persistence
You can enable persistence by setting the useRootPersistence
option to true
when constructing a trie through the Trie.create
function. As such, this value is preserved when creating copies of the trie and is incapable of being modified once a trie is instantiated.
import { Trie } from '@ethereumjs/trie'
const trie = await Trie.create({
useRootPersistence: true,
})
Proofs
Merkle Proofs
The createProof
and verifyProof
functions allow you to verify that a certain value does or does not exist within a Merkle Patricia Tree with a given root.
Proof-of-Inclusion
The following code demonstrates how to construct and subsequently verify a proof that confirms the existence of the key test
(which corresponds with the value one
) within the given trie. This is also known as inclusion, hence the name 'Proof-of-Inclusion.'
import { Trie } from '@ethereumjs/trie'
import { bytesToUtf8, utf8ToBytes } from '@ethereumjs/util'
const trie = new Trie()
async function test() {
await trie.put(utf8ToBytes('test'), utf8ToBytes('one'))
const proof = await trie.createProof(utf8ToBytes('test'))
const value = await trie.verifyProof(trie.root(), utf8ToBytes('test'), proof)
console.log(value ? bytesToUtf8(value) : 'not found')
}
test()
Proof-of-Exclusion
The following code demonstrates how to construct and subsequently verify a proof that confirms that the key test3
does not exist within the given trie. This is also known as exclusion, hence the name 'Proof-of-Exclusion.'
import { Trie } from '@ethereumjs/trie'
import { bytesToUtf8, utf8ToBytes } from '@ethereumjs/util'
const trie = new Trie()
async function test() {
await trie.put(utf8ToBytes('test'), utf8ToBytes('one'))
await trie.put(utf8ToBytes('test2'), utf8ToBytes('two'))
const proof = await trie.createProof(utf8ToBytes('test3'))
const value = await trie.verifyProof(trie.root(), utf8ToBytes('test3'), proof)
console.log(value ? bytesToUtf8(value) : 'null')
}
test()
Invalid Proofs
If verifyProof
detects an invalid proof, it will throw an error. While contrived, the below example illustrates the resulting error condition in the event a prover tampers with the data in a merkle proof.
import { Trie } from '@ethereumjs/trie'
import { bytesToUtf8, utf8ToBytes } from '@ethereumjs/util'
const trie = new Trie()
async function test() {
await trie.put(utf8ToBytes('test'), utf8ToBytes('one'))
await trie.put(utf8ToBytes('test2'), utf8ToBytes('two'))
const proof = await trie.createProof(utf8ToBytes('test2'))
proof[1].reverse()
try {
const value = await trie.verifyProof(trie.root(), utf8ToBytes('test2'), proof)
console.log(value ? bytesToUtf8(value) : 'not found')
} catch (err) {
console.log(err)
}
}
test()
Range Proofs
You may use the Trie.verifyRangeProof()
function to confirm if the given leaf nodes and edge proof possess the capacity to prove that the given trie leaves' range matches the specific root (which is useful for snap sync, for instance).
Examples
You can find additional examples complete with detailed explanations here.
Browser
With the breaking release round in Summer 2023 we have added hybrid ESM/CJS builds for all our libraries (see section below) and have eliminated many of the caveats which had previously prevented a frictionless browser usage.
It is now easily possible to run a browser build of one of the EthereumJS libraries within a modern browser using the provided ESM build. For a setup example see ./examples/browser.html.
API
Docs
Generated TypeDoc API Documentation
Hybrid CJS/ESM Builds
With the breaking releases from Summer 2023 we have started to ship our libraries with both CommonJS (cjs
folder) and ESM builds (esm
folder), see package.json
for the detailed setup.
If you use an ES6-style import
in your code files from the ESM build will be used:
import { EthereumJSClass } from '@ethereumjs/[PACKAGE_NAME]'
If you use Node.js specific require
, the CJS build will be used:
const { EthereumJSClass } = require('@ethereumjs/[PACKAGE_NAME]')
Using ESM will give you additional advantages over CJS beyond browser usage like static code analysis / Tree Shaking which CJS can not provide.
Buffer -> Uint8Array
With the breaking releases from Summer 2023 we have removed all Node.js specific Buffer
usages from our libraries and replace these with Uint8Array representations, which are available both in Node.js and the browser (Buffer
is a subclass of Uint8Array
).
We have converted existing Buffer conversion methods to Uint8Array conversion methods in the @ethereumjs/util bytes
module, see the respective README section for guidance.
BigInt Support
With the 5.0.0 release, BigInt takes the place of BN.js.
BigInt is a primitive that is used to represent and manipulate primitive bigint
values that the number primitive is incapable of representing as a result of their magnitude. ES2020
saw the introduction of this particular feature. Note that this version update resulted in the altering of number-related API signatures and that the minimal build target is now set to ES2020
.
Benchmarking
You will find two simple benchmarks in the benchmarks
folder:
random.ts
runs random PUT
operations on the tree, andcheckpointing.ts
runs checkpoints and commits between PUT
operations
A third benchmark using mainnet data to simulate real load is also being considered.
You may run benchmarks using:
npm run benchmarks
To run a profiler on the random.ts
benchmark and generate a flamegraph with 0x, you may use:
npm run profiling
0x processes the stacks and generates a profile folder (<pid>.0x
) containing flamegraph.html
.
Debugging
The Trie
class features optional debug logging.. Individual debug selections can be activated on the CL with DEBUG=ethjs,[Logger Selection]
.
ethjs
must be included in the DEBUG
environment variables to enable any logs.
Additional log selections can be added with a comma separated list (no spaces). Logs with extensions can be enabled with a colon :
, and *
can be used to include all extensions.
DEBUG=ethjs,thislog,thatlog,otherlog,otherlog:sublog,anotherLog:* node myscript.js
The following options are available:
Logger | Description |
---|
trie | minimal info logging for all trie methods |
trie:<METHOD> | debug logging for specific trie method |
trie:<METHOD>:* | verbose debug logging for specific trie method |
trie:* | verbose debug logging for all trie methods |
To observe the logging in action at different levels:
Run with minimal logging:
DEBUG=ethjs,trie npx vitest test/util/log.spec.ts
Run with put method logging:
DEBUG=ethjs,trie:PUT npx vitest test/util/log.spec.ts
Run with trie + put/get/del logging:
DEBUG=ethjs,trie,trie:PUT,trie:GET,trie:DEL npx vitest test/util/log.spec.ts
Run with findPath debug logging:
DEBUG=ethjs,trie:FIND_PATH npx vitest test/util/log.spec.ts
Run with findPath verbose logging:
DEBUG=ethjs,trie:FIND_PATH:* npx vitest test/util/log.spec.ts
Run with max logging:
DEBUG=ethjs,trie:* npx vitest test/util/log.spec.ts
References
EthereumJS
See our organizational documentation for an introduction to EthereumJS
as well as information on current standards and best practices. If you want to join for work or carry out improvements on the libraries, please review our contribution guidelines first.
License
MPL-2.0