refactor(core): inline auxiliary model operations (#45974)

This commit is contained in:
Kit Langton
2026-08-28 12:34:01 -04:00
committed by GitHub
Unverified
parent e8fa7daed5
commit 1f77f4a4ed
2 changed files with 290 additions and 319 deletions
+186 -193
View File
@@ -1,7 +1,6 @@
export * as SessionCompaction from "./compaction.js"
import { LLMClient, AIError, LLMEvent, Message, type LLMRequest } from "@opencode-ai/ai"
import type { StreamOptions } from "@opencode-ai/ai/route"
import { LLMClient, LLMEvent, Message } from "@opencode-ai/ai"
import { SessionError } from "@opencode-ai/schema/session-error"
import { Context, Effect, Layer, Stream } from "effect"
import { Bus } from "../bus.js"
@@ -65,13 +64,6 @@ export type Draft = {
configure: (settings: Partial<Settings>) => void
}
type Dependencies = {
readonly bus: Bus.Interface
readonly llm: {
readonly stream: (request: LLMRequest, options?: StreamOptions) => Stream.Stream<LLMEvent, AIError>
}
}
export type AutoInput = {
readonly session: SessionSchema.Info
readonly messages: readonly SessionMessage.Info[]
@@ -240,195 +232,196 @@ const planContent = (messages: readonly SessionMessage.Info[], tokens: number) =
}
}
const make = (dependencies: Dependencies) => {
const state = State.create<Settings, Draft>({
name: "session-compaction",
initial: () => ({ auto: true, buffer: DEFAULT_BUFFER, tokens: DEFAULT_KEEP_TOKENS }),
draft: (draft) => ({
configure: (settings) => {
if (settings.auto !== undefined) draft.auto = settings.auto
if (settings.buffer !== undefined) draft.buffer = settings.buffer
if (settings.tokens !== undefined) draft.tokens = settings.tokens
},
}),
})
const failed = Effect.fnUntraced(function* (input: {
readonly sessionID: SessionSchema.ID
readonly reason: SessionMessage.Compaction["reason"]
readonly error: SessionError.Error
readonly inputID?: SessionMessage.ID
}) {
yield* dependencies.bus.publish(SessionEvent.Compaction.Failed, input)
return { status: "failed" as const, error: input.error }
})
const execute = Effect.fn("SessionCompaction.execute")(function* (plan: Plan) {
if (!plan.started)
yield* dependencies.bus.publish(SessionEvent.Compaction.Started, {
sessionID: plan.session.id,
reason: plan.reason,
recent: plan.recent,
inputID: plan.inputID,
})
const chunks: string[] = []
let failure: SessionError.Error | undefined
let usage: SessionUsage.Recorded | undefined
const recordUsage = Effect.suspend(() =>
usage
? dependencies.bus.publish(SessionEvent.UsageRecorded, {
sessionID: plan.session.id,
source: "compaction",
...usage,
})
: Effect.void,
)
const prepared = yield* plan.prepare({
scope: { session: plan.session, agentID: Agent.ID.make("compaction"), model: plan.resolved },
transcript: { system: [], messages: [Message.user(plan.prompt)] },
contextHooks: false,
})
yield* dependencies.llm.stream(prepared.request, prepared.options).pipe(
Stream.runForEach((event) => {
if (LLMEvent.is.providerError(event))
failure = {
type: event.classification === "context-overflow" ? "provider.invalid-request" : "provider.error",
message: event.message,
}
if (LLMEvent.is.textDelta(event)) {
chunks.push(event.text)
return dependencies.bus.publish(SessionEvent.Compaction.Delta, {
sessionID: plan.session.id,
text: event.text,
})
}
if (LLMEvent.is.stepFinish(event)) {
const step = SessionUsage.record(event.usage, plan.resolved.cost)
usage = usage ? SessionUsage.add(usage, step) : step
}
return Effect.void
}),
Effect.catchTag("AI.Error", (error) =>
Effect.sync(() => {
failure = toSessionError(error)
}),
),
Effect.onInterrupt(() =>
recordUsage.pipe(
Effect.andThen(
plan.reason === "auto"
? failed({
sessionID: plan.session.id,
reason: plan.reason,
error: { type: "compaction.interrupted", message: "Compaction was interrupted" },
inputID: plan.inputID,
}).pipe(Effect.asVoid)
: Effect.void,
),
),
),
)
yield* recordUsage
const summary = chunks.join("")
if (failure || !summary.trim()) {
const error = failure ?? { type: "compaction.failed" as const, message: "Compaction produced no summary" }
return yield* failed({
sessionID: plan.session.id,
reason: plan.reason,
error,
inputID: plan.inputID,
})
}
yield* dependencies.bus.publish(SessionEvent.Compaction.Ended, {
sessionID: plan.session.id,
reason: plan.reason,
text: summary,
recent: plan.recent,
})
return { status: "completed" as const }
})
const compact = Effect.fn("SessionCompaction.compact")(function* (input: AutoInput) {
const content = planContent(input.messages, state.get().tokens)
if (content)
return yield* execute({
session: input.session,
resolved: input.resolved,
prepare: input.prepare,
reason: "auto",
...content,
})
return yield* failed({
sessionID: input.session.id,
reason: "auto",
error: { type: "compaction.unavailable", message: "Nothing to compact yet" },
})
})
const required = (input: RequiredInput) => {
const config = state.get()
if (!config.auto) return false
const limit = input.resolved.limit
const context = limit.context
if (context <= 0) return false
const last = input.messages.findLast(
(message): message is SessionMessage.Assistant & { tokens: NonNullable<SessionMessage.Assistant["tokens"]> } =>
message.type === "assistant" && message.tokens !== undefined,
)
if (!last) return false
const output = Math.min(limit.output, OUTPUT_TOKEN_MAX)
const promptCeiling = Math.min(
limit.input === undefined ? Number.POSITIVE_INFINITY : limit.input - config.buffer,
context - Math.max(output, config.buffer),
)
const used =
last.tokens.input + last.tokens.output + last.tokens.reasoning + last.tokens.cache.read + last.tokens.cache.write
if (used <= 0) return false
return used >= promptCeiling
}
const compactManual = Effect.fn("SessionCompaction.compactManual")(function* (input: ManualInput) {
const content = planContent(input.messages, state.get().tokens)
if (!content)
return yield* failed({
sessionID: input.session.id,
reason: "manual",
error: { type: "compaction.unavailable", message: "Nothing to compact yet" },
inputID: input.inputID,
})
const resolved = yield* input.resolveModel(input.session).pipe(
Effect.catch((cause) =>
failed({
sessionID: input.session.id,
reason: "manual",
error: toSessionError(cause),
inputID: input.inputID,
}),
),
)
if ("status" in resolved) return resolved
return yield* execute({
session: input.session,
resolved,
prepare: input.prepare,
reason: "manual",
inputID: input.inputID,
started: input.started,
...content,
})
})
return Service.of({
transform: state.transform,
reload: state.reload,
enabled: () => state.get().auto,
required,
compact,
compactManual,
})
}
export const layer = Layer.effect(
Service,
Effect.gen(function* () {
const bus = yield* Bus.Service
const llm = yield* LLMClient.Service
return make({ bus, llm })
const state = State.create<Settings, Draft>({
name: "session-compaction",
initial: () => ({ auto: true, buffer: DEFAULT_BUFFER, tokens: DEFAULT_KEEP_TOKENS }),
draft: (draft) => ({
configure: (settings) => {
if (settings.auto !== undefined) draft.auto = settings.auto
if (settings.buffer !== undefined) draft.buffer = settings.buffer
if (settings.tokens !== undefined) draft.tokens = settings.tokens
},
}),
})
const failed = Effect.fnUntraced(function* (input: {
readonly sessionID: SessionSchema.ID
readonly reason: SessionMessage.Compaction["reason"]
readonly error: SessionError.Error
readonly inputID?: SessionMessage.ID
}) {
yield* bus.publish(SessionEvent.Compaction.Failed, input)
return { status: "failed" as const, error: input.error }
})
const execute = Effect.fn("SessionCompaction.execute")(function* (plan: Plan) {
if (!plan.started)
yield* bus.publish(SessionEvent.Compaction.Started, {
sessionID: plan.session.id,
reason: plan.reason,
recent: plan.recent,
inputID: plan.inputID,
})
const chunks: string[] = []
let failure: SessionError.Error | undefined
let usage: SessionUsage.Recorded | undefined
const recordUsage = Effect.suspend(() =>
usage
? bus.publish(SessionEvent.UsageRecorded, {
sessionID: plan.session.id,
source: "compaction",
...usage,
})
: Effect.void,
)
const prepared = yield* plan.prepare({
scope: { session: plan.session, agentID: Agent.ID.make("compaction"), model: plan.resolved },
transcript: { system: [], messages: [Message.user(plan.prompt)] },
contextHooks: false,
})
yield* llm.stream(prepared.request, prepared.options).pipe(
Stream.runForEach((event) => {
if (LLMEvent.is.providerError(event))
failure = {
type: event.classification === "context-overflow" ? "provider.invalid-request" : "provider.error",
message: event.message,
}
if (LLMEvent.is.textDelta(event)) {
chunks.push(event.text)
return bus.publish(SessionEvent.Compaction.Delta, {
sessionID: plan.session.id,
text: event.text,
})
}
if (LLMEvent.is.stepFinish(event)) {
const step = SessionUsage.record(event.usage, plan.resolved.cost)
usage = usage ? SessionUsage.add(usage, step) : step
}
return Effect.void
}),
Effect.catchTag("AI.Error", (error) =>
Effect.sync(() => {
failure = toSessionError(error)
}),
),
Effect.onInterrupt(() =>
recordUsage.pipe(
Effect.andThen(
plan.reason === "auto"
? failed({
sessionID: plan.session.id,
reason: plan.reason,
error: { type: "compaction.interrupted", message: "Compaction was interrupted" },
inputID: plan.inputID,
}).pipe(Effect.asVoid)
: Effect.void,
),
),
),
)
yield* recordUsage
const summary = chunks.join("")
if (failure || !summary.trim()) {
const error = failure ?? { type: "compaction.failed" as const, message: "Compaction produced no summary" }
return yield* failed({
sessionID: plan.session.id,
reason: plan.reason,
error,
inputID: plan.inputID,
})
}
yield* bus.publish(SessionEvent.Compaction.Ended, {
sessionID: plan.session.id,
reason: plan.reason,
text: summary,
recent: plan.recent,
})
return { status: "completed" as const }
})
const compact = Effect.fn("SessionCompaction.compact")(function* (input: AutoInput) {
const content = planContent(input.messages, state.get().tokens)
if (content)
return yield* execute({
session: input.session,
resolved: input.resolved,
prepare: input.prepare,
reason: "auto",
...content,
})
return yield* failed({
sessionID: input.session.id,
reason: "auto",
error: { type: "compaction.unavailable", message: "Nothing to compact yet" },
})
})
const required = (input: RequiredInput) => {
const config = state.get()
if (!config.auto) return false
const limit = input.resolved.limit
const context = limit.context
if (context <= 0) return false
const last = input.messages.findLast(
(message): message is SessionMessage.Assistant & { tokens: NonNullable<SessionMessage.Assistant["tokens"]> } =>
message.type === "assistant" && message.tokens !== undefined,
)
if (!last) return false
const output = Math.min(limit.output, OUTPUT_TOKEN_MAX)
const promptCeiling = Math.min(
limit.input === undefined ? Number.POSITIVE_INFINITY : limit.input - config.buffer,
context - Math.max(output, config.buffer),
)
const used =
last.tokens.input +
last.tokens.output +
last.tokens.reasoning +
last.tokens.cache.read +
last.tokens.cache.write
if (used <= 0) return false
return used >= promptCeiling
}
const compactManual = Effect.fn("SessionCompaction.compactManual")(function* (input: ManualInput) {
const content = planContent(input.messages, state.get().tokens)
if (!content)
return yield* failed({
sessionID: input.session.id,
reason: "manual",
error: { type: "compaction.unavailable", message: "Nothing to compact yet" },
inputID: input.inputID,
})
const resolved = yield* input.resolveModel(input.session).pipe(
Effect.catch((cause) =>
failed({
sessionID: input.session.id,
reason: "manual",
error: toSessionError(cause),
inputID: input.inputID,
}),
),
)
if ("status" in resolved) return resolved
return yield* execute({
session: input.session,
resolved,
prepare: input.prepare,
reason: "manual",
inputID: input.inputID,
started: input.started,
...content,
})
})
return Service.of({
transform: state.transform,
reload: state.reload,
enabled: () => state.get().auto,
required,
compact,
compactManual,
})
}),
)
+104 -126
View File
@@ -1,8 +1,7 @@
export * as SessionTitle from "./title.js"
import { isDeepStrictEqual } from "node:util"
import { LLMClient, AIError, LLMEvent, Message, SystemPart, type LLMRequest } from "@opencode-ai/ai"
import type { StreamOptions } from "@opencode-ai/ai/route"
import { LLMClient, LLMEvent, Message, SystemPart } from "@opencode-ai/ai"
import { Context, DateTime, Effect, Layer, Stream } from "effect"
import type { Agent } from "../agent.js"
import { Database } from "../database/database.js"
@@ -23,15 +22,6 @@ const MAX_CONTEXT_LENGTH = 8_000
const MAX_FIRST_MESSAGE_LENGTH = 2_000
const titleChanged = Symbol("Session title changed")
type Dependencies = {
readonly bus: Bus.Interface
readonly llm: {
readonly stream: (request: LLMRequest, options?: StreamOptions) => Stream.Stream<LLMEvent, AIError>
}
readonly context: SessionContext.Interface
readonly store: SessionStore.Interface
}
export interface Interface {
/** Generates an initial title or regenerates one from bounded conversation history. */
readonly generate: (sessionID: SessionSchema.ID) => Effect.Effect<void>
@@ -46,117 +36,6 @@ export const isUntitled = (session: SessionSchema.Info) =>
time: { created: DateTime.toEpochMillis(session.time.created) },
})
const attempt = Effect.fn("SessionTitle.attempt")(function* (
dependencies: Dependencies,
input: {
readonly session: SessionSchema.Info
readonly agent: Agent.Info
readonly text: string
readonly model: SessionRunnerModel.Resolved
},
) {
const chunks: string[] = []
let failed = false
let usage: SessionUsage.Recorded | undefined
const recordUsage = Effect.suspend(() =>
usage
? dependencies.bus.publish(SessionEvent.UsageRecorded, {
sessionID: input.session.id,
source: "title",
...usage,
})
: Effect.void,
)
const prepared = yield* dependencies.context.prepare({
scope: { session: input.session, agentID: input.agent.id, model: input.model },
transcript: {
system: input.agent.system ? [SystemPart.make(input.agent.system)] : [],
messages: [Message.user(input.text)],
},
contextHooks: false,
})
yield* dependencies.llm.stream(prepared.request, prepared.options).pipe(
Stream.runForEach((event) => {
if (LLMEvent.is.providerError(event)) failed = true
if (LLMEvent.is.textDelta(event)) chunks.push(event.text)
if (LLMEvent.is.stepFinish(event)) {
const step = SessionUsage.record(event.usage, input.model.cost)
usage = usage ? SessionUsage.add(usage, step) : step
}
return Effect.void
}),
Effect.catchTag("AI.Error", () =>
Effect.sync(() => {
failed = true
}),
),
Effect.onInterrupt(() => recordUsage.pipe(Effect.asVoid)),
)
yield* recordUsage
if (failed) return
return chunks
.join("")
.split("\n")
.map((line) => line.trim())
.find((line) => line.length > 0)
})
const make = (dependencies: Dependencies) => {
const generate = Effect.fn("SessionTitle.generate")(function* (
db: Database.Interface["db"],
sessionID: SessionSchema.ID,
) {
const session = yield* dependencies.store.get(sessionID)
if (!session) return
const firstUser = yield* SessionHistory.firstUserMessage(db, session.id)
if (!firstUser) return
const text = !isUntitled(session)
? yield* dependencies.store.context(session.id).pipe(
Effect.map((messages) => {
const original = `Original request:\n${firstUser.text.slice(0, MAX_FIRST_MESSAGE_LENGTH)}`
const recent = messages
.flatMap((message) => {
if (message.type === "user" && message.id !== firstUser.id) return [`User: ${message.text.trim()}`]
if (message.type !== "assistant") return []
const text = message.content
.flatMap((part) => (part.type === "text" ? [part.text.trim()] : []))
.filter(Boolean)
.join("\n")
return text ? [`Assistant: ${text}`] : []
})
.join("\n\n")
if (!recent) return original
const prefix = `${original}\n\nRecent conversation:\n`
return `${prefix}${recent.slice(-(MAX_CONTEXT_LENGTH - prefix.length))}`
}),
Effect.orElseSucceed(() => firstUser.text),
)
: firstUser.text
const selection = yield* dependencies.context.selectTitle(session)
if (!selection) return
const title =
(yield* attempt(dependencies, { session, agent: selection.agent, text, model: selection.selected })) ??
(selection.primary && !isDeepStrictEqual(selection.selected.ref, selection.primary.ref)
? yield* attempt(dependencies, { session, agent: selection.agent, text, model: selection.primary })
: undefined)
if (!title) return
const expectedSequence = (yield* Bus.latestSequence(db, sessionID)) + 1
const current = yield* dependencies.store.get(sessionID)
if (!current || current.title !== session.title || current.title === truncate(title)) return
yield* dependencies.bus
.publish(
SessionEvent.Renamed,
{
sessionID: session.id,
title: truncate(title),
},
{ commit: (sequence) => (sequence === expectedSequence ? Effect.void : Effect.die(titleChanged)) },
)
.pipe(Effect.catchDefect((defect) => (defect === titleChanged ? Effect.void : Effect.die(defect))))
})
return { generate }
}
export const layer = Layer.effect(
Service,
Effect.gen(function* () {
@@ -164,11 +43,110 @@ export const layer = Layer.effect(
const llm = yield* LLMClient.Service
const context = yield* SessionContext.Service
const store = yield* SessionStore.Service
const database = yield* Database.Service
const title = make({ bus, llm, context, store })
return Service.of({
generate: (sessionID) => title.generate(database.db, sessionID),
const db = (yield* Database.Service).db
const attempt = Effect.fn("SessionTitle.attempt")(function* (input: {
readonly session: SessionSchema.Info
readonly agent: Agent.Info
readonly text: string
readonly model: SessionRunnerModel.Resolved
}) {
const chunks: string[] = []
let failed = false
let usage: SessionUsage.Recorded | undefined
const recordUsage = Effect.suspend(() =>
usage
? bus.publish(SessionEvent.UsageRecorded, {
sessionID: input.session.id,
source: "title",
...usage,
})
: Effect.void,
)
const prepared = yield* context.prepare({
scope: { session: input.session, agentID: input.agent.id, model: input.model },
transcript: {
system: input.agent.system ? [SystemPart.make(input.agent.system)] : [],
messages: [Message.user(input.text)],
},
contextHooks: false,
})
yield* llm.stream(prepared.request, prepared.options).pipe(
Stream.runForEach((event) => {
if (LLMEvent.is.providerError(event)) failed = true
if (LLMEvent.is.textDelta(event)) chunks.push(event.text)
if (LLMEvent.is.stepFinish(event)) {
const step = SessionUsage.record(event.usage, input.model.cost)
usage = usage ? SessionUsage.add(usage, step) : step
}
return Effect.void
}),
Effect.catchTag("AI.Error", () =>
Effect.sync(() => {
failed = true
}),
),
Effect.onInterrupt(() => recordUsage.pipe(Effect.asVoid)),
)
yield* recordUsage
if (failed) return
return chunks
.join("")
.split("\n")
.map((line) => line.trim())
.find((line) => line.length > 0)
})
const generate = Effect.fn("SessionTitle.generate")(function* (sessionID: SessionSchema.ID) {
const session = yield* store.get(sessionID)
if (!session) return
const firstUser = yield* SessionHistory.firstUserMessage(db, session.id)
if (!firstUser) return
const text = !isUntitled(session)
? yield* store.context(session.id).pipe(
Effect.map((messages) => {
const original = `Original request:\n${firstUser.text.slice(0, MAX_FIRST_MESSAGE_LENGTH)}`
const recent = messages
.flatMap((message) => {
if (message.type === "user" && message.id !== firstUser.id) return [`User: ${message.text.trim()}`]
if (message.type !== "assistant") return []
const text = message.content
.flatMap((part) => (part.type === "text" ? [part.text.trim()] : []))
.filter(Boolean)
.join("\n")
return text ? [`Assistant: ${text}`] : []
})
.join("\n\n")
if (!recent) return original
const prefix = `${original}\n\nRecent conversation:\n`
return `${prefix}${recent.slice(-(MAX_CONTEXT_LENGTH - prefix.length))}`
}),
Effect.orElseSucceed(() => firstUser.text),
)
: firstUser.text
const selection = yield* context.selectTitle(session)
if (!selection) return
const title =
(yield* attempt({ session, agent: selection.agent, text, model: selection.selected })) ??
(selection.primary && !isDeepStrictEqual(selection.selected.ref, selection.primary.ref)
? yield* attempt({ session, agent: selection.agent, text, model: selection.primary })
: undefined)
if (!title) return
const expectedSequence = (yield* Bus.latestSequence(db, sessionID)) + 1
const current = yield* store.get(sessionID)
if (!current || current.title !== session.title || current.title === truncate(title)) return
yield* bus
.publish(
SessionEvent.Renamed,
{
sessionID: session.id,
title: truncate(title),
},
{ commit: (sequence) => (sequence === expectedSequence ? Effect.void : Effect.die(titleChanged)) },
)
.pipe(Effect.catchDefect((defect) => (defect === titleChanged ? Effect.void : Effect.die(defect))))
})
return Service.of({ generate })
}),
)