Evenfire is an intelligence layer platform for running LLM agents that take real actions, on your infrastructure
TypeScript
import type { SecretEditState } from '@clerum/frontend-components'
import { CHANNEL_CREDENTIAL_FIELDS } from '@components/ChannelCredentialsPanel/constants'
import type { CredentialEditStates, CredentialKey } from '@components/ChannelCredentialsPanel/types'
import type { ChannelType } from './channelTypes'
export type CredentialOperation =
| { kind: 'replace'; key: CredentialKey; label: string; value: string }
| { kind: 'clear'; key: CredentialKey; label: string }
export function credentialStateIsDirty(state: SecretEditState | undefined): boolean {
return state?.status === 'replaced' || state?.status === 'cleared'
}
export function buildCredentialOperationPlan(
states: CredentialEditStates,
validChannelTypes: readonly ChannelType[]
): { operations: CredentialOperation[]; invalidLabels: string[] } {
const valid = new Set(validChannelTypes)
const operations: CredentialOperation[] = []
const invalidLabels: string[] = []
for (const field of CHANNEL_CREDENTIAL_FIELDS) {
const state = states[field.key]
if (!credentialStateIsDirty(state)) continue
if (!valid.has(field.channelType)) {
invalidLabels.push(field.label)
continue
}
if (state?.status === 'replaced') {
const value = state.value.trim()
if (value) operations.push({ kind: 'replace', key: field.key, label: field.label, value })
else invalidLabels.push(field.label)
} else if (state?.status === 'cleared') {
operations.push({ kind: 'clear', key: field.key, label: field.label })
}
}
return { operations, invalidLabels }
}
export function cleanCredentialStates(
states: CredentialEditStates,
successfulKeys: readonly CredentialKey[]
): CredentialEditStates {
const successful = new Set(successfulKeys)
return Object.fromEntries(
Object.entries(states).map(([key, state]) => [
key,
successful.has(key as CredentialKey) ? { status: 'untouched' } : state,
])
) as CredentialEditStates
}