mcpField
F
fp-ts-task-either

fp-ts-task-either

github.com/whatiskadudoingDevelopment1451 installsUpdated 39w ago
Repository

About fp-ts-task-either

A guide for TypeScript developers using fp-ts to compose lazy asynchronous operations with TaskEither and handle errors. It covers sequential and parallel execution, recovery, result matching, and API and database patterns.

Key features

  1. 01Convert Promises to TaskEither
  2. 02Compose asynchronous operations sequentially
  3. 03Run operations in parallel or with limits
  4. 04Map, recover from, and retry errors
  5. 05Match and handle async results
  6. 06Build API and database operation patterns

Use cases

  1. 01Model fallible API requests
  2. 02Orchestrate async steps that depend on prior results
  3. 03Fetch multiple data items in parallel
  4. 04Try a fallback when a primary request fails

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/whatiskadudoing/fp-ts-skills --skill fp-ts-task-either

After installing, type/fp-ts-task-eitherin Claude Code to confirm it loaded.

Details

From the project README

fp-ts TaskEither Async Patterns

TaskEither combines the laziness of Task with the error handling of Either, providing a powerful abstraction for async operations that can fail.

Core Concepts

import * as TE from 'fp-ts/TaskEither'
import * as T from 'fp-ts/Task'
import * as E from 'fp-ts/Either'
import { pipe, flow } from 'fp-ts/function'

TaskEither<E, A> is equivalent to () => Promise<Either<E, A>>

  • E = Error type (left)
  • A = Success type (right)
  • Lazy: nothing executes until you call the function
  • Composable: chain operations without try/catch

1. Converting Promises to TaskEither

Using tryCatch

The primary way to lift Promises into TaskEither:

import * as TE from 'fp-ts/TaskEither'

// Basic tryCatch pattern
const fetchUser = (id: string): TE.TaskEither<Error, User> =>
  TE.tryCatch(
    () => fetch(`/api/users/${id}`).then(res => res.json()),
    (reason) => new Error(String(reason))
  )

// With typed errors
interface ApiError {
  code: string
  message: string
  status: number
}

const fetchUserTyped = (id: string): TE.TaskEither<ApiError, User> =>
  TE.tryCatch(
    () => fetch(`/api/users/${id}`).then(res => res.json()),
    (reason): ApiError => ({
      code: 'FETCH_ERROR',
      message: reason instanceof Error ? reason.message : 'Unknown error',
      status: 500
    })
  )
From existing Either
// Lift an Either into TaskEither
const fromEither: TE.TaskEither<Error, number> = TE.fromEither(E.right(42))

// From a nullable value
const fromNullable = TE.fromNullable(new Error('Value was null'))
const result = fromNullable(maybeValue) // TaskEither<Error, NonNullable<T>>

// From an Option
import * as O from 'fp-ts/Option'
const fromOption = TE.fromOption(() => new Error('None value'))
const optionResult = fromOption(O.some(42)) // TaskEither<Error, number>
Creating TaskEither values directly
// Success value
const success = TE.right<Error, number>(42)

// Error value
const failure = TE.left<Error, number>(new Error('Something failed'))

// From a predicate
const validatePositive = TE.fromPredicate(
  (n: number) => n > 0,
  (n) => new Error(`Expected positive, got ${n}`)
)

2. Handling Async Errors Functionally

Mapping over errors
// Transform the error type
const withMappedError = pipe(
  fetchUser('123'),
  TE.mapLeft((error) => ({
    type: 'USER_FETCH_ERROR' as const,
    originalError: error,
    timestamp: Date.now()
  }))
)

// Bifunctor: map both sides
const mapped = pipe(
  fetchUser('123'),
  TE.bimap(
    (error) => new DetailedError(error),  // map error
    (user) => user.profile                 // map success
  )
)
Error filtering
// Filter with error on false
const validateAge = pipe(
  fetchUser('123'),
  TE.filterOrElse(
    (user) => user.age >= 18,
    (user) => new Error(`User ${user.name} is underage`)
  )
)

3. Chaining Async Operations

Sequential chaining with chain/flatMap
interface User { id: string; name: string; teamId: string }
interface Team { id: string; name: string; orgId: string }
interface Org { id: string; name: string }

const fetchUser = (id: string): TE.TaskEither<Error, User> =>
  TE.tryCatch(() => api.getUser(id), toError)

const fetchTeam = (teamId: string): TE.TaskEither<Error, Team> =>
  TE.tryCatch(() => api.getTeam(teamId), toError)

const fetchOrg = (orgId: string): TE.TaskEither<Error, Org> =>
  TE.tryCatch(() => api.getOrg(orgId), toError)

// Chain operations sequentially
const getUserOrg = (userId: string): TE.TaskEither<Error, Org> =>
  pipe(
    fetchUser(userId),
    TE.chain((user) => fetchTeam(user.teamId)),
    TE.chain((team) => fetchOrg(team.orgId))
  )

// flatMap is an alias for chain
const getUserOrgAlt = (userId: string): TE.TaskEither<Error, Org> =>
  pipe(
    fetchUser(userId),
    TE.flatMap((user) => fetchTeam(user.teamId)),
    TE.flatMap((team) => fetchOrg(team.orgId))
  )
Chaining with intermediate values
// Use bind to accumulate values
const getFullContext = (userId: string) =>
  pipe(
    TE.Do,
    TE.bind('user', () => fetchUser(userId)),
    TE.bind('team', ({ user }) => fetchTeam(user.teamId)),
    TE.bind('org', ({ team }) => fetchOrg(team.orgId)),
    TE.map(({ user, team, org }) => ({
      userName: user.name,
      teamName: team.name,
      orgName: org.name
    }))
  )

4. Parallel vs Sequential Execution

Parallel execution with sequenceArray
import * as A from 'fp-ts/Array'

const userIds = ['1', '2', '3', '4', '5']

// Parallel: all requests start immediately
// Fails fast: returns first error encountered
const fetchAllUsersParallel = pipe(
  userIds.map(fetchUser),
  TE.sequenceArray  // TaskEither<Error, readonly User[]>
)

// Sequential: one at a time (use when order matters or rate limiting)
const fetchAllUsersSequential = pipe(
  userIds,
  A.traverse(TE.ApplicativeSeq)(fetchUser)
)
Parallel with traverseArray
// More idiomatic: traverse combines map + sequence
const fetchAllUsers = (ids: string[]): TE.TaskEither<Error, readonly User[]> =>
  pipe(ids, TE.traverseArray(fetchUser))

// With index
const fetchWithIndex = pipe(
  userIds,
  TE.traverseArrayWithIndex((index, id) =>
    pipe(
      fetchUser(id),
      TE.map(user => ({ ...user, index }))
    )
  )
)
Collecting all errors vs fail fast
import * as These from 'fp-ts/These'
import * as TH from 'fp-ts/TaskThese'

// For collecting all errors, consider TaskThese
// Or use validation with sequenceT

import { sequenceT } from 'fp-ts/Apply'

// Parallel execution, collects results
const parallel = sequenceT(TE.ApplyPar)(
  fetchUser('1'),
  fetchTeam('team-1'),
  fetchOrg('org-1')
) // TaskEither<Error, [User, Team, Org]>

// Sequential execution
const sequential = sequenceT(TE.ApplySeq)(
  fetchUser('1'),
  fetchTeam('team-1'),
  fetchOrg('org-1')
)
Concurrent with limit
// For controlled concurrency, batch your operations
const batchSize = 3

const fetchInBatches = (ids: string[]): TE.TaskEither<Error, User[]> => {
  const batches = chunk(ids, batchSize)

  return pipe(
    batches,
    A.traverse(TE.ApplicativeSeq)((batch) =>
      pipe(batch, TE.traverseArray(fetchUser))
    ),
    TE.map(A.flatten)
  )
}

5. Error Recovery with orElse

Basic error recovery
// Try primary, fall back to secondary
const fetchWithFallback = pipe(
  fetchFromPrimaryApi(id),
  TE.orElse((primaryError) =>
    pipe(
      fetchFromBackupApi(id),
      TE.mapLeft((backupError) => ({
        primary: primaryError,
        backup: backupError
      }))
    )
  )
)

// Recover to a default value
const fetchWithDefault = pipe(
  fetchUser(id),
  TE.orElse(() => TE.right(defaultUser))
)

// orElseW when recovery has different error type
const fetchWithTypedFallback = pipe(
  fetchFromApi(id),           // TaskEither<ApiError, User>
  TE.orElseW((apiError) =>    // orElseW allows different error type
    fetchFromCache(id)         // TaskEither<CacheError, User>
  )
) // TaskEither<CacheError, User>
Retry patterns
const retry = <E, A>(
  te: TE.TaskEither<E, A>,
  retries: number,
  delay: number
): TE.TaskEither<E, A> =>
  pipe(
    te,
    TE.orElse((error) =>
      retries > 0
        ? pipe(
            T.delay(delay)(T.of(undefined)),
            T.chain(() => retry(te, retries - 1, delay * 2))
          )
        : TE.left(error)
    )
  )

// Usage
const fetchWithRetry = retry(fetchUser('123'), 3, 1000)
Conditional recovery
// Only recover from specific errors
const recoverFromNotFound = pipe(
  fetchUser(id),
  TE.orElse((error) =>
    error.code === 'NOT_FOUND'
      ? TE.right(createDefaultUser(id))
      : TE.left(error)  // re-throw other errors
  )
)

// Alt: try alternatives in order
import { alt } from 'fp-ts/TaskEither'

const fetchFromAnywhere = pipe(
  fetchFromCache(id),
  TE.alt(() => fetchFromApi(id)),
  TE.alt(() => fetchFromBackup(id))
)

6. Pattern Matching Async Results

Using fold/match
// fold executes the TaskEither and handles both cases
const handleResult = pipe(
  fetchUser('123'),
  TE.fold(
    (error) => T.of(`Error: ${error.message}`),
    (user) => T.of(`Welcome, ${user.name}`)
  )
) // Task<string> - no longer has error channel

// match is an alias for fold
const handleWithMatch = pipe(
  fetchUser('123'),
  TE.match(
    (error) => ({ success: false, error }),
    (user) => ({ success: true, data: user })
  )
)

// matchW when handlers return different types
const handleWithMatchW = pipe(
  fetchUser('123'),
  TE.matchW(
    (error) => ({ type: 'error' as const, error }),
    (user) => ({ type: 'success' as const, user })
  )
)
Getting the underlying Either
// Execute and get the Either
const getEither = async () => {
  const either = await fetchUser('123')()

  if (E.isLeft(either)) {
    console.error('Failed:', either.left)
  } else {
    console.log('User:', either.right)
  }
}

// Using getOrElse for default
const getWithDefault = pipe(
  fetchUser('123'),
  TE.getOrElse((error) => T.of(defaultUser))
) // Task<User>

// getOrElseW when default has different type
const getOrNull = pipe(
  fetchUser('123'),
  TE.getOrElseW(() => T.of(null))
) // Task<User | null>

7. Do Notation for Complex Workflows

Building complex operations
interface OrderContext {
  user: User
  cart: Cart
  payment: PaymentMethod
  shipping: ShippingAddress
}

const processOrder = (
  userId: string,
  cartId: string
): TE.TaskEither<OrderError, OrderConfirmation> =>
  pipe(
    TE.Do,
    // Bind values sequentially
    TE.bind('user', () => fetchUser(userId)),
    TE.bind('cart', () => fetchCart(cartId)),

    // Validate intermediate results
    TE.filterOrElse(
      ({ cart }) => cart.items.length > 0,
      () => ({ code: 'EMPTY_CART', message: 'Cart is empty' })
    ),

    // Continue building context
    TE.bind('payment', ({ user }) => getDefaultPayment(user.id)),
    TE.bind('shipping', ({ user }) => getDefaultShipping(user.id)),

    // Calculate derived values
    TE.bind('total', ({ cart }) => TE.right(calculateTotal(cart))),

    // Validate before final operation
    TE.filterOrElse(
      ({ payment, total }) => payment.limit >= total,
      ({ total }) => ({ code: 'LIMIT_EXCEEDED', message: `Order total ${total} exceeds limit` })
    ),

    // Final operation
    TE.chain(({ user, cart, payment, shipping, total }) =>
      createOrder({ user, cart, payment, shipping, total })
    )
  )
Parallel fetching within Do
const getOrderDetails = (orderId: string) =>
  pipe(
    TE.Do,
    TE.bind('order', () => fetchOrder(orderId)),

    // Parallel fetch based on order data
    TE.bind('details', ({ order }) =>
      sequenceT(TE.ApplyPar)(
        fetchUser(order.userId),
        fetchProducts(order.productIds),
        fetchShipping(order.shippingId)
      )
    ),

    TE.map(({ order, details: [user, products, shipping] }) => ({
      order,
      user,
      products,
      shipping
    }))
  )
Using apS for simpler additions
// apS is like bind but doesn't depend on previous values
const enrichUser = (userId: string) =>
  pipe(
    fetchUser(userId),
    TE.bindTo('user'),  // Wrap in { user: ... }
    TE.apS('config', fetchAppConfig()),  // Add independent value
    TE.apS('features', fetchFeatureFlags()),
    TE.bind('preferences', ({ user }) => fetchPreferences(user.id))  // Dependent
  )

8. Real-World API Call Patterns

Typed API client
interface ApiConfig {
  baseUrl: string
  timeout: number
}

interface ApiError {
  code: string
  message: string
  status: number
  details?: unknown
}

const createApiClient = (config: ApiConfig) => {
  const request = <T>(
    method: string,
    path: string,
    body?: unknown
  ): TE.TaskEither<ApiError, T> =>
    TE.tryCatch(
      async () => {
        const response = await fetch(`${config.baseUrl}${path}`, {
          method,
          headers: { 'Content-Type': 'application/json' },
          body: body ? JSON.stringify(body) : undefined,
          signal: AbortSignal.timeout(config.timeout)
        })

        if (!response.ok) {
          const error = await response.json().catch(() => ({}))
          throw { status: response.status, ...error }
        }

        return response.json()
      },
      (error): ApiError => ({
        code: 'API_ERROR',
        message: error instanceof Error ? error.message : 'Request failed',
        status: (error as any)?.status ?? 500,
        details: error
      })
    )

  return {
    get: <T>(path: string) => request<T>('GET', path),
    post: <T>(path: string, body: unknown) => request<T>('POST', path, body),
    put: <T>(path: string, body: unknown) => request<T>('PUT', path, body),
    delete: <T>(path: string) => request<T>('DELETE', path)
  }
}

// Usage
const api = createApiClient({ baseUrl: '/api', timeout: 5000 })

const getUser = (id: string) => api.get<User>(`/users/${id}`)
const createUser = (data: CreateUserDto) => api.post<User>('/users', data)
Request with validation
import * as t from 'io-ts'
import { PathReporter } from 'io-ts/PathReporter'

const UserCodec = t.type({
  id: t.string,
  name: t.string,
  email: t.string
})

type User = t.TypeOf<typeof UserCodec>

const fetchAndValidate = <A>(
  codec: t.Type<A>,
  url: string
): TE.TaskEither<Error, A> =>
  pipe(
    TE.tryCatch(
      () => fetch(url).then(r => r.json()),
      (e) => new Error(`Fetch failed: ${e}`)
    ),
    TE.chainEitherK((data) =>
      pipe(
        codec.decode(data),
        E.mapLeft((errors) =>
          new Error(`Validation failed: ${PathReporter.report(E.left(errors)).join(', ')}`)
        )
      )
    )
  )

const getValidatedUser = (id: string) =>
  fetchAndValidate(UserCodec, `/api/users/${id}`)

9. Database Operation Patterns

Repository pattern
interface Repository<E, T, ID> {
  findById: (id: ID) => TE.TaskEither<E, T>
  findAll: () => TE.TaskEither<E, readonly T[]>
  save: (entity: T) => TE.TaskEither<E, T>
  delete: (id: ID) => TE.TaskEither<E, void>
}

interface DbError {
  code: 'NOT_FOUND' | 'DUPLICATE' | 'CONSTRAINT' | 'CONNECTION'
  message: string
  cause?: unknown
}

const createUserRepository = (db: Database): Repository<DbError, User, string> => ({
  findById: (id) =>
    pipe(
      TE.tryCatch(
        () => db.query('SELECT * FROM users WHERE id = ?', [id]),
        (e): DbError => ({ code: 'CONNECTION', message: String(e), cause: e })
      ),
      TE.chain((rows) =>
        rows.length === 0
          ? TE.left({ code: 'NOT_FOUND', message: `User ${id} not found` })
          : TE.right(rows[0] as User)
      )
    ),

  findAll: () =>
    TE.tryCatch(
      () => db.query('SELECT * FROM users'),
      (e): DbError => ({ code: 'CONNECTION', message: String(e), cause: e })
    ),

  save: (user) =>
    TE.tryCatch(
      () => db.query(
        'INSERT INTO users (id, name, email) VALUES (?, ?, ?) ON CONFLICT (id) DO UPDATE SET name = ?, email = ?',
        [user.id, user.name, user.email, user.name, user.email]
      ).then(() => user),
      (e): DbError => {
        if (String(e).includes('UNIQUE constraint')) {
          return { code: 'DUPLICATE', message: 'Email already exists', cause: e }
        }
        return { code: 'CONNECTION', message: String(e), cause: e }
      }
    ),

  delete: (id) =>
    TE.tryCatch(
      () => db.query('DELETE FROM users WHERE id = ?', [id]).then(() => undefined),
      (e): DbError => ({ code: 'CONNECTION', message: String(e), cause: e })
    )
})
Transaction handling
interface Transaction {
  query: (sql: string, params?: unknown[]) => Promise<unknown>
  commit: () => Promise<void>
  rollback: () => Promise<void>
}

const withTransaction = <E, A>(
  db: Database,
  operation: (tx: Transaction) => TE.TaskEither<E, A>
): TE.TaskEither<E | DbError, A> =>
  pipe(
    TE.tryCatch(
      () => db.beginTransaction(),
      (e): DbError => ({ code: 'CONNECTION', message: 'Failed to start transaction', cause: e })
    ),
    TE.chain((tx) =>
      pipe(
        operation(tx),
        TE.chainFirst(() =>
          TE.tryCatch(
            () => tx.commit(),
            (e): DbError => ({ code: 'CONNECTION', message: 'Commit failed', cause: e })
          )
        ),
        TE.orElse((error) =>
          pipe(
            TE.tryCatch(() => tx.rollback(), () => error),
            TE.chain(() => TE.left(error))
          )
        )
      )
    )
  )

// Usage
const transferFunds = (fromId: string, toId: string, amount: number) =>
  withTransaction(db, (tx) =>
    pipe(
      TE.Do,
      TE.bind('from', () => getAccount(tx, fromId)),
      TE.bind('to', () => getAccount(tx, toId)),
      TE.filterOrElse(
        ({ from }) => from.balance >= amount,
        () => ({ code: 'INSUFFICIENT_FUNDS', message: 'Not enough balance' })
      ),
      TE.chain(({ from, to }) =>
        sequenceT(TE.ApplySeq)(
          updateBalance(tx, fromId, from.balance - amount),
          updateBalance(tx, toId, to.balance + amount)
        )
      ),
      TE.map(() => ({ success: true, amount }))
    )
  )

10. Task vs TaskEither: When to Use Which

Use Task when:
import * as T from 'fp-ts/Task'

// 1. Operation cannot fail
const delay = (ms: number): T.Task<void> =>
  () => new Promise(resolve => setTimeout(resolve, ms))

// 2. Errors are handled elsewhere
const logMessage = (msg: string): T.Task<void> =>
  () => console.log(msg) as unknown as Promise<void>

// 3. You want to ignore errors
const fetchOrDefault = (url: string, defaultValue: Data): T.Task<Data> =>
  pipe(
    TE.tryCatch(() => fetch(url).then(r => r.json()), E.toError),
    TE.getOrElse(() => T.of(defaultValue))
  )

// 4. Fire and forget
const trackAnalytics = (event: Event): T.Task<void> =>
  () => analytics.track(event).catch(() => {}) // Errors swallowed
Use TaskEither when:
// 1. Operation can fail and you need to handle the error
const fetchUser = (id: string): TE.TaskEither<Error, User> =>
  TE.tryCatch(() => api.getUser(id), E.toError)

// 2. You need typed errors for different failure modes
type AuthError =
  | { type: 'INVALID_CREDENTIALS' }
  | { type: 'EXPIRED_TOKEN' }
  | { type: 'NETWORK_ERROR'; cause: Error }

const authenticate = (token: string): TE.TaskEither<AuthError, User> => { /* ... */ }

// 3. Error recovery is part of business logic
const getConfig = (): TE.TaskEither<ConfigError, Config> =>
  pipe(
    fetchRemoteConfig(),
    TE.orElse(() => loadLocalConfig()),
    TE.orElse(() => TE.right(defaultConfig))
  )

// 4. Composing multiple fallible operations
const processOrder = (orderId: string): TE.TaskEither<OrderError, Receipt> =>
  pipe(
    validateOrder(orderId),
    TE.chain(chargePayment),
    TE.chain(fulfillOrder),
    TE.chain(sendConfirmation)
  )
Converting between them
// Task to TaskEither (infallible to fallible)
const taskToTE = <A>(task: T.Task<A>): TE.TaskEither<never, A> =>
  pipe(task, T.map(E.right))

// TaskEither to Task (handle/ignore error)
const teToTask = <E, A>(te: TE.TaskEither<E, A>, defaultValue: A): T.Task<A> =>
  TE.getOrElse(() => T.of(defaultValue))(te)

// TaskEither to Task (throw on error - escape hatch)
const teToTaskThrow = <E, A>(te: TE.TaskEither<E, A>): T.Task<A> =>
  pipe(
    te,
    TE.getOrElse((e) => () => Promise.reject(e))
  )

Quick Reference

OperationFunctionDescription
Create successTE.right(value)Wrap value in Right
Create failureTE.left(error)Wrap error in Left
From PromiseTE.tryCatch(promise, onError)Convert Promise to TE
Transform valueTE.map(f)Apply f to success value
Transform errorTE.mapLeft(f)Apply f to error value
Chain operationsTE.chain(f) / TE.flatMap(f)Sequence dependent operations
Recover from errorTE.orElse(f)Try alternative on error
Handle both casesTE.fold(onError, onSuccess)Pattern match result
Parallel arrayTE.traverseArray(f)Map + sequence in parallel
Sequential arrayA.traverse(TE.ApplicativeSeq)(f)Map + sequence in order
Filter with errorTE.filterOrElse(pred, onFalse)Validate with error
Get or defaultTE.getOrElse(onError)Extract value with fallback

Common Patterns Summary

// 1. Fetch with error handling
const fetch = TE.tryCatch(() => api.get(url), toError)

// 2. Chain dependent calls
pipe(getA(), TE.chain(a => getB(a.id)), TE.chain(b => getC(b.id)))

// 3. Parallel independent calls
sequenceT(TE.ApplyPar)(getA(), getB(), getC())

// 4. Build context with Do
pipe(TE.Do, TE.bind('a', () => getA()), TE.bind('b', ({a}) => getB(a)))

// 5. Recover from errors
pipe(primary(), TE.orElse(() => fallback()))

// 6. Execute and handle result
pipe(operation(), TE.fold(handleError, handleSuccess))()

Comments

Markdown supported