> **Can't find what you're looking for?** Use `search_docs` on the docs MCP server at `https://docs.ampdb.io/api/mcp` to find what you need. # Domain Types The types the SDK's modules share: block ranges, dataset names and references, manifests, jobs, and tokens. Import them from `@edgeandnode/amp`'s `Domain` namespace, or from `@edgeandnode/amp/core`. Each export is both a schema and a TypeScript type, of the same name: `Domain.BlockRange` is the type of a block range, and the schema that checks one. A schema can: * Check and convert a value: `Schema.decodeUnknownSync(Domain.DatasetManifest)(json)` reads a manifest file's JSON, and throws when it isn't valid. `Schema.encodeSync` does the reverse. * Create a branded value, for the types that are branded: `Domain.DatasetName.make("eth_mainnet")` throws when the value isn't valid (see [Effect Basics](/sdks/typescript/effect#schemas-and-branded-types)). Many have different property names in JSON, where Amp uses snake_case: `BlockRange`'s `prevHash` is `prev_hash`. Decoding and encoding convert them. ## Blocks and batches ### `BlockRange` A range of blocks on one chain, which a batch of rows comes from. | Property | Type | Description | | ---------- | ------------------------------------------ | -------------------------------------------------------------------------------------- | | `network` | `Network` | The chain, like `"mainnet"`. | | `numbers` | `{ start: BlockNumber; end: BlockNumber }` | The first and last blocks, both included. | | `hash` | `BlockHash` | The last block's hash. | | `prevHash` | `BlockHash`, optional | The hash of the block before the first, its parent. Missing for a chain's first block. | ```ts { network: "mainnet", numbers: { start: 24000000, end: 24000099 }, hash: "0x6ed4b3511195052eea47925fe5035f5e3fc528440eed62649ba82058a1b70e40", prevHash: "0xe8f20171e085c248ac4d60b65ddeb9c1bd679149cc1be9925ad58799dc380b25", } ``` ### `RecordBatchMetadata` What Amp says about a batch of results: `ranges`, its `BlockRange`s, one per chain; and `rangesComplete`, `true` once all the rows for those ranges have arrived. `RecordBatchMetadataFromUint8Array` decodes it from the bytes Amp sends with each batch. ### Block types | Type | What it is | Check | | ------------- | ----------------------------------------------------------------------- | ------------------------------------------------- | | `BlockNumber` | A block's number | A non-negative integer. Branded. | | `BlockHash` | A block's hash | `0x` and 64 lowercase letters or digits. Branded. | | `Network` | A chain's name, as its datasets and providers call it, like `"mainnet"` | Lowercase. Branded. | ## Datasets ### `DatasetReference` A dataset and one of its revisions: `{ namespace, name, revision }`. As a string, `namespace/name@revision`, like `acme/eth_mainnet@1.0.0`. * `DatasetReferenceFromString` reads one from that string: `Schema.decodeUnknownSync(Domain.DatasetReferenceFromString)("acme/eth_mainnet@1.0.0")`. Without a namespace, it's `_`, and without a revision, `dev`. * `DatasetReferenceString` is the string, checked against that format. ### `DatasetRevision` Which version of a dataset to use: a `DatasetVersion` (`"1.0.0"`), a `DatasetHash`, or a `DatasetTag` (`"latest"` or `"dev"`). `latest` is the highest version, and `dev` the last manifest registered, with or without a version. ### Name and version types | Type | What it is | Check | | ----------------------- | -------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------- | | `DatasetNamespace` | Who a dataset belongs to, like `"acme"` or `"edgeandnode"`; `"_"` when none is given | Lowercase letters, digits, and underscores. Branded. | | `DatasetName` | A dataset's name, like `"eth_mainnet"` | Lowercase letters, digits, and underscores, not starting with a digit. Branded. | | `DatasetVersion` | A version, in [semantic versioning](https://semver.org), like `"1.0.0"` or `"1.0.0-dev123"` | A semantic version. Branded. | | `DatasetHash` | A manifest's SHA-256 hash, which identifies its exact contents | 64 hexadecimal characters. Branded. | | `DatasetTag` | `"latest"` or `"dev"` | Branded. | | `DatasetNameAndVersion` | A name and a version, like `"uniswap@1.0.0"` | A name, `@`, and a semantic version. | | `DatasetKind` | What kind of dataset: `"manifest"` (derived), `"static"`, or a raw kind | Branded. | | `RawDatasetKind` | The kinds of raw datasets: `"evm-rpc"`, `"solana"`, `"eth-beacon"`, `"bittensor"`, `"tempo"`, `"bitcoin-rpc"`, `"stellar"`, `"phaser"` | | | `DatasetVisibility` | `"public"` or `"private"` | | [Datasets](/get-started/concepts/datasets) explains namespaces, names, versions, and kinds. ### Metadata types The metadata a dataset can have, for people to find and understand it: | Type | What it is | | -------------------- | ------------------------------------------------------------------------------------------------------------------------------- | | `DatasetMetadata` | `namespace` and `name`, and optional `readme`, `repository`, `description`, `keywords`, `sources`, `license`, and `visibility`. | | `DatasetDescription` | What it's for: a string of up to 1,024 characters. | | `DatasetReadme` | Its documentation, a string. | | `DatasetRepository` | The `URL` of its source code. | | `DatasetKeyword` | A word to find it by, like `"DeFi"`. | | `DatasetSource` | Where its data comes from, like `"mainnet_rpc.logs"` or a contract address. | | `DatasetLicense` | Its license, like `"MIT"`. | | `DatasetInfo` | `name`, `kind`, and `tables`, each a `TableInfo`: `name`, `network`, and `activeLocation`, where its files are. | | `DatasetConfig` | A derived dataset's configuration: its metadata, `network`, `startBlock`, `dependencies`, and its `tables` and `functions`. | ## Manifests ### `DatasetManifest` A dataset's [manifest](/datasets/manifests): one of three kinds, told apart by `kind`. Read a manifest file with `Schema.decodeUnknownSync(Domain.DatasetManifest)(json)`. | Type | `kind` | Properties | | ---------------- | ------------------ | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | `DatasetRaw` | A `RawDatasetKind` | `network`; `tables`, each a `RawDatasetTable`; and optional `startBlock`, `finalizedBlocksOnly`, and `stateSnapshots`. See [Raw Datasets](/datasets/raw-datasets). | | `DatasetDerived` | `"manifest"` | Optional `dependencies`, the datasets it reads, by alias, as `DatasetReference`s; `tables`, each a `Table`; and `functions`, each a `FunctionManifest`. See [Derived Datasets](/datasets/derived-datasets). | | `DatasetStatic` | `"static"` | `tables`, each a `StaticDatasetTable`. See [Static Datasets](/datasets/static-datasets). | ### Manifest parts | Type | What it is | | ------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | `RawDatasetTable` | A raw dataset's table: `schema`, `network`, and optional `sortedBy`, `bloomFilterColumns`, and `segmentSelectionPolicy`. | | `Table` | A derived dataset's table: `input`, a `TableInput` (`{ sql }`); `schema`; and optional `bloomFilterColumns`. | | `StaticDatasetTable` | A static dataset's table: `path`, the CSV file; `schema`; `format` (`"csv"`); `hasHeader`; and optional `sortedBy`. | | `BloomFilterColumnConfig` | A column to build a [bloom filter](/datasets/bloom-filters) for: `column`, and optional `ndv`, how many distinct values to expect. | | `SegmentSelectionPolicy` | How a raw table chooses its files: `"hash-linked"` or `"ordered"`. | | `StateSnapshots` | A raw dataset's state snapshots: optional `selectors`, each a `StateSnapshotSelector`, a selector string or `{ selector, blockInterval }`. | | `FunctionManifest` | A [JavaScript function](/querying/functions/javascript): `source`, a `FunctionSource` (`{ source, filename }`), `inputTypes`, and `outputType`, as `ArrowDataType`s. | | `FunctionDefinition` | The same, as `getOutputSchema` takes it. | | `TableDefinition` | `{ sql }`, a table in a `DatasetConfig`. | ### `TableSchema` A table's columns: `{ arrow: ArrowSchema }`, where `ArrowSchema` is `{ fields }`, and each `ArrowField` has `name`, `type`, an `ArrowDataType`, and `nullable`. `TableSchemaInfo` is a table's `name`, `network`, and `schema`. `ArrowDataType` is an Arrow type, as Amp writes it: a string for a type without parameters, like `"UInt64"`, or an object for one with them, like `{ "FixedSizeBinary": 32 }` or `{ "Timestamp": ["Nanosecond", "+00:00"] }`. ## Jobs ### `JobInfo` A [job](/managing-amp/jobs): | Property | Type | Description | | ---------------- | -------------- | -------------------------------------------------------------------------------------------------------------------------- | | `id` | `JobId` | Its ID, a branded number. | | `status` | `JobStatus` | `SCHEDULED`, `RUNNING`, `COMPLETED`, `STOPPED`, `STOP_REQUESTED`, `STOPPING`, `ERROR`, `CANCELLED`, `FATAL`, or `UNKNOWN`. | | `descriptor` | `any` | What it does, like `{ kind: "materialize-raw", dataset_name: "eth_mainnet", … }`. | | `nodeId` | `string` | The worker, or controller, it runs on. | | `idempotencyKey` | `string` | A hash that identifies the work the job does. | | `createdAt` | `DateTime.Utc` | When it was created. `DateTime.toDate` turns it into a `Date`. | | `updatedAt` | `DateTime.Utc` | When it last changed. | ## Authentication ### `AuthInfo` A sign-in, as the [`Auth`](/sdks/typescript/authentication#auth) service saves it: | Property | Type | Description | | -------------- | --------------------------------- | ---------------------------------------------------------- | | `accessToken` | `Redacted` | The token requests send. | | `refreshToken` | `Redacted` | The token that gets a new access token. | | `userId` | `UserId` | Who signed in. | | `accounts` | `ReadonlyArray`, optional | Their accounts, like wallet addresses. | | `expiry` | `number`, optional | When the access token expires, in milliseconds since 1970. | ### Token types | Type | What it is | | --------------- | --------------------------------------------------------------------------------------------------------------- | | `AccessToken` | An OAuth2 access token: a branded non-empty string. | | `RefreshToken` | An OAuth2 refresh token: a branded non-empty string. | | `TokenDuration` | How long a generated token lasts: a number and a unit, like `"7 days"`, `"1 hour"`, or `"30 minutes"`. Branded. | | `UserId` | A user's ID, like `cmfd6bf6u006vjx0b7xb2eybx`, optionally after `did:privy:`. Branded. | | `Address` | An Ethereum address, `0x` and 40 hexadecimal characters. Branded. | ## Basic types | Type | What it is | | ----------------------- | ------------------------------------------------------ | | `NonEmptyTrimmedString` | A string that isn't empty, without spaces at its ends. | | `NonNegativeInt` | An integer of 0 or more. |