compressed-context 0.1.0 → 0.2.0

package/src/compression/transcript-store.ts

@@ -1,258 +1,258 @@
-import type { Schema } from "effect"
-
-import { isSessionDeleted } from "./deleted-sessions.ts"
-import { forEachScannedEntry } from "../shared/scan.ts"
-import type { Block, StorageLike } from "./state.ts"
-import type { MediaSourceLike, MessageLike } from "./select.ts"
-
-/** 落盘的媒体 source：bytes 统一转 base64；url/ref 只存引用。 */
-export interface StoredMediaSource {
-  readonly type: "base64" | "url" | "ref" | "unknown"
-  readonly data?: string
-  readonly mediaType?: string
-  readonly url?: string
-  readonly provider?: string
-  readonly id?: string
-  readonly expiresAt?: number
-}
-
-/** 落盘的单条 part；reasoning 不进入这里。 */
-export interface StoredPart {
-  readonly type: "text" | "tool-call" | "tool-result" | "media"
-  readonly text?: string
-  readonly name?: string
-  readonly input?: unknown
-  readonly result?: unknown
-  readonly filename?: string
-  readonly media?: {
-    readonly source: StoredMediaSource
-    readonly info?: unknown
-  }
-}
-
-/** 落盘的单条消息。 */
-export interface StoredMessage {
-  readonly id?: string
-  readonly role: string
-  readonly content: readonly StoredPart[]
-}
-
-/** 一个 T1 块的 transcript manifest；消息本体按 msg:<index> 分开存。 */
-export interface TranscriptManifest {
-  readonly v: 1
-  readonly sessionID: string
-  readonly blockID: string
-  readonly fromId: string
-  readonly toId: string
-  readonly count: number
-  readonly createdAt: number
-}
-
-const PREFIX = "compression:transcript:"
-
-/** 输入：sessionID；输出：该会话全部 transcript key 的公共前缀。 */
-export function transcriptSessionPrefix(sessionID: string): string {
-  return `${PREFIX}${sessionID}:`
-}
-
-/** 输入：sessionID + blockID；输出：manifest key。 */
-export function transcriptManifestKey(sessionID: string, blockID: string): string {
-  return `${transcriptSessionPrefix(sessionID)}${blockID}:manifest`
-}
-
-/** 输入：sessionID + blockID + 序号；输出：单条消息 key。 */
-export function transcriptMessageKey(sessionID: string, blockID: string, index: number): string {
-  return `${transcriptSessionPrefix(sessionID)}${blockID}:msg:${index.toString()}`
-}
-
-/** 输入：任意数据；输出：是否 Uint8Array。 */
-function isBytes(value: unknown): value is Uint8Array {
-  return value instanceof Uint8Array
-}
-
-/** 输入：media source；输出：可 JSON 化的 source；未知形状保留 unknown。 */
-function storeSource(source: MediaSourceLike | undefined): StoredMediaSource {
-  if (source === undefined) return { type: "unknown" }
-  if (source.type === "bytes") {
-    const data = isBytes(source.data)
-      ? Buffer.from(source.data).toString("base64")
-      : typeof source.data === "string"
-        ? source.data
-        : ""
-    return {
-      type: "base64",
-      ...(data.length === 0 ? {} : { data }),
-      ...(source.mediaType === undefined ? {} : { mediaType: source.mediaType }),
-    }
-  }
-  if (source.type === "base64") {
-    return {
-      type: "base64",
-      ...(typeof source.data === "string" ? { data: source.data } : {}),
-      ...(source.mediaType === undefined ? {} : { mediaType: source.mediaType }),
-    }
-  }
-  if (source.type === "url") {
-    return {
-      type: "url",
-      ...(source.url === undefined ? {} : { url: source.url }),
-      ...(source.mediaType === undefined ? {} : { mediaType: source.mediaType }),
-      ...(source.expiresAt === undefined ? {} : { expiresAt: source.expiresAt }),
-    }
-  }
-  if (source.type === "ref") {
-    return {
-      type: "ref",
-      ...(source.provider === undefined ? {} : { provider: source.provider }),
-      ...(source.id === undefined ? {} : { id: source.id }),
-      ...(source.mediaType === undefined ? {} : { mediaType: source.mediaType }),
-    }
-  }
-  return { type: "unknown", ...(source.mediaType === undefined ? {} : { mediaType: source.mediaType }) }
-}
-
-/** 输入：可压缩消息；输出：落盘消息；预期：跳过 reasoning 和其他未知 part。 */
-export function toStoredMessage(message: MessageLike): StoredMessage {
-  const content: StoredPart[] = []
-  for (const part of message.content) {
-    if (part.type === "text" && typeof part.text === "string") {
-      content.push({ type: "text", text: part.text })
-      continue
-    }
-    if (part.type === "tool-call") {
-      content.push({
-        type: "tool-call",
-        ...(part.name === undefined ? {} : { name: part.name }),
-        input: part.input,
-      })
-      continue
-    }
-    if (part.type === "tool-result") {
-      content.push({
-        type: "tool-result",
-        ...(part.name === undefined ? {} : { name: part.name }),
-        result: part.result,
-      })
-      continue
-    }
-    if (part.type === "media") {
-      content.push({
-        type: "media",
-        ...(part.filename === undefined ? {} : { filename: part.filename }),
-        media: {
-          source: storeSource(part.media?.source),
-          ...(part.media?.info === undefined ? {} : { info: part.media.info }),
-        },
-      })
-    }
-  }
-  return {
-    ...(message.id === undefined ? {} : { id: message.id }),
-    role: message.role,
-    content,
-  }
-}
-
-/**
- * 输入：storage + 会话 + T1 块 + 覆盖的消息。
- * 输出：无。
- * 预期：先写 manifest，再逐条写消息；删除墓碑存在时整体跳过。
- */
-export async function saveTranscript(
-  storage: StorageLike,
-  sessionID: string,
-  block: Pick<Block, "id" | "fromId" | "toId">,
-  messages: readonly MessageLike[],
-): Promise<void> {
-  if (isSessionDeleted(sessionID)) return
-  const manifest: TranscriptManifest = {
-    v: 1,
-    sessionID,
-    blockID: block.id,
-    fromId: block.fromId,
-    toId: block.toId,
-    count: messages.length,
-    createdAt: Date.now(),
-  }
-  await storage.set(transcriptManifestKey(sessionID, block.id), manifest as unknown as Schema.Json)
-  for (const [index, message] of messages.entries()) {
-    await storage.set(
-      transcriptMessageKey(sessionID, block.id, index),
-      toStoredMessage(message) as unknown as Schema.Json,
-    )
-  }
-}
-
-/** 输入：storage + 会话 + 块；输出：合法 manifest 或 undefined。 */
-export async function loadTranscriptManifest(
-  storage: StorageLike,
-  sessionID: string,
-  blockID: string,
-): Promise<TranscriptManifest | undefined> {
-  const raw = await storage.get(transcriptManifestKey(sessionID, blockID))
-  if (typeof raw !== "object" || raw === null) return undefined
-  const record = raw as Record<string, unknown>
-  if (
-    record["v"] !== 1 ||
-    typeof record["sessionID"] !== "string" ||
-    typeof record["blockID"] !== "string" ||
-    typeof record["fromId"] !== "string" ||
-    typeof record["toId"] !== "string" ||
-    typeof record["count"] !== "number" ||
-    !Number.isInteger(record["count"]) ||
-    record["count"] < 0
-  ) {
-    return undefined
-  }
-  return {
-    v: 1,
-    sessionID: record["sessionID"],
-    blockID: record["blockID"],
-    fromId: record["fromId"],
-    toId: record["toId"],
-    count: record["count"],
-    createdAt: typeof record["createdAt"] === "number" ? record["createdAt"] : 0,
-  }
-}
-
-/** 输入：storage + 会话 + 块 + 序号；输出：落盘消息或 undefined。 */
-export async function loadTranscriptMessage(
-  storage: StorageLike,
-  sessionID: string,
-  blockID: string,
-  index: number,
-): Promise<StoredMessage | undefined> {
-  const raw = await storage.get(transcriptMessageKey(sessionID, blockID, index))
-  if (typeof raw !== "object" || raw === null) return undefined
-  const record = raw as Record<string, unknown>
-  if (typeof record["role"] !== "string" || !Array.isArray(record["content"])) return undefined
-  return {
-    ...(typeof record["id"] === "string" ? { id: record["id"] } : {}),
-    role: record["role"],
-    content: record["content"] as StoredPart[],
-  }
-}
-
-/** 输入：storage + 会话；输出：无；预期：删掉该会话所有 manifest/message key。 */
-export async function deleteTranscriptsForSession(
-  storage: StorageLike & {
-    scan?: (options: { prefix: string; limit: number; after?: string }) => Promise<{
-      entries: readonly { key: string; value: unknown }[]
-      next?: string
-    }>
-  },
-  sessionID: string,
-): Promise<void> {
-  await forEachScannedEntry(storage, transcriptSessionPrefix(sessionID), async (entry) => {
-    await storage.remove(entry.key)
-  })
-}
-
-/** 输入：storage + 会话 + 块；输出：manifest 是否存在。 */
-export async function hasTranscript(
-  storage: StorageLike,
-  sessionID: string,
-  blockID: string,
-): Promise<boolean> {
-  return (await loadTranscriptManifest(storage, sessionID, blockID)) !== undefined
-}
+import type { Schema } from "effect"
+
+import { isSessionDeleted } from "./deleted-sessions.ts"
+import { forEachScannedEntry } from "../shared/scan.ts"
+import type { Block, StorageLike } from "./state.ts"
+import type { MediaSourceLike, MessageLike } from "./select.ts"
+
+/** 落盘的媒体 source：bytes 统一转 base64；url/ref 只存引用。 */
+export interface StoredMediaSource {
+  readonly type: "base64" | "url" | "ref" | "unknown"
+  readonly data?: string
+  readonly mediaType?: string
+  readonly url?: string
+  readonly provider?: string
+  readonly id?: string
+  readonly expiresAt?: number
+}
+
+/** 落盘的单条 part；reasoning 不进入这里。 */
+export interface StoredPart {
+  readonly type: "text" | "tool-call" | "tool-result" | "media"
+  readonly text?: string
+  readonly name?: string
+  readonly input?: unknown
+  readonly result?: unknown
+  readonly filename?: string
+  readonly media?: {
+    readonly source: StoredMediaSource
+    readonly info?: unknown
+  }
+}
+
+/** 落盘的单条消息。 */
+export interface StoredMessage {
+  readonly id?: string
+  readonly role: string
+  readonly content: readonly StoredPart[]
+}
+
+/** 一个 T1 块的 transcript manifest；消息本体按 msg:<index> 分开存。 */
+export interface TranscriptManifest {
+  readonly v: 1
+  readonly sessionID: string
+  readonly blockID: string
+  readonly fromId: string
+  readonly toId: string
+  readonly count: number
+  readonly createdAt: number
+}
+
+const PREFIX = "compression:transcript:"
+
+/** 输入：sessionID；输出：该会话全部 transcript key 的公共前缀。 */
+export function transcriptSessionPrefix(sessionID: string): string {
+  return `${PREFIX}${sessionID}:`
+}
+
+/** 输入：sessionID + blockID；输出：manifest key。 */
+export function transcriptManifestKey(sessionID: string, blockID: string): string {
+  return `${transcriptSessionPrefix(sessionID)}${blockID}:manifest`
+}
+
+/** 输入：sessionID + blockID + 序号；输出：单条消息 key。 */
+export function transcriptMessageKey(sessionID: string, blockID: string, index: number): string {
+  return `${transcriptSessionPrefix(sessionID)}${blockID}:msg:${index.toString()}`
+}
+
+/** 输入：任意数据；输出：是否 Uint8Array。 */
+function isBytes(value: unknown): value is Uint8Array {
+  return value instanceof Uint8Array
+}
+
+/** 输入：media source；输出：可 JSON 化的 source；未知形状保留 unknown。 */
+function storeSource(source: MediaSourceLike | undefined): StoredMediaSource {
+  if (source === undefined) return { type: "unknown" }
+  if (source.type === "bytes") {
+    const data = isBytes(source.data)
+      ? Buffer.from(source.data).toString("base64")
+      : typeof source.data === "string"
+        ? source.data
+        : ""
+    return {
+      type: "base64",
+      ...(data.length === 0 ? {} : { data }),
+      ...(source.mediaType === undefined ? {} : { mediaType: source.mediaType }),
+    }
+  }
+  if (source.type === "base64") {
+    return {
+      type: "base64",
+      ...(typeof source.data === "string" ? { data: source.data } : {}),
+      ...(source.mediaType === undefined ? {} : { mediaType: source.mediaType }),
+    }
+  }
+  if (source.type === "url") {
+    return {
+      type: "url",
+      ...(source.url === undefined ? {} : { url: source.url }),
+      ...(source.mediaType === undefined ? {} : { mediaType: source.mediaType }),
+      ...(source.expiresAt === undefined ? {} : { expiresAt: source.expiresAt }),
+    }
+  }
+  if (source.type === "ref") {
+    return {
+      type: "ref",
+      ...(source.provider === undefined ? {} : { provider: source.provider }),
+      ...(source.id === undefined ? {} : { id: source.id }),
+      ...(source.mediaType === undefined ? {} : { mediaType: source.mediaType }),
+    }
+  }
+  return { type: "unknown", ...(source.mediaType === undefined ? {} : { mediaType: source.mediaType }) }
+}
+
+/** 输入：可压缩消息；输出：落盘消息；预期：跳过 reasoning 和其他未知 part。 */
+export function toStoredMessage(message: MessageLike): StoredMessage {
+  const content: StoredPart[] = []
+  for (const part of message.content) {
+    if (part.type === "text" && typeof part.text === "string") {
+      content.push({ type: "text", text: part.text })
+      continue
+    }
+    if (part.type === "tool-call") {
+      content.push({
+        type: "tool-call",
+        ...(part.name === undefined ? {} : { name: part.name }),
+        input: part.input,
+      })
+      continue
+    }
+    if (part.type === "tool-result") {
+      content.push({
+        type: "tool-result",
+        ...(part.name === undefined ? {} : { name: part.name }),
+        result: part.result,
+      })
+      continue
+    }
+    if (part.type === "media") {
+      content.push({
+        type: "media",
+        ...(part.filename === undefined ? {} : { filename: part.filename }),
+        media: {
+          source: storeSource(part.media?.source),
+          ...(part.media?.info === undefined ? {} : { info: part.media.info }),
+        },
+      })
+    }
+  }
+  return {
+    ...(message.id === undefined ? {} : { id: message.id }),
+    role: message.role,
+    content,
+  }
+}
+
+/**
+ * 输入：storage + 会话 + T1 块 + 覆盖的消息。
+ * 输出：无。
+ * 预期：先写 manifest，再逐条写消息；删除墓碑存在时整体跳过。
+ */
+export async function saveTranscript(
+  storage: StorageLike,
+  sessionID: string,
+  block: Pick<Block, "id" | "fromId" | "toId">,
+  messages: readonly MessageLike[],
+): Promise<void> {
+  if (isSessionDeleted(sessionID)) return
+  const manifest: TranscriptManifest = {
+    v: 1,
+    sessionID,
+    blockID: block.id,
+    fromId: block.fromId,
+    toId: block.toId,
+    count: messages.length,
+    createdAt: Date.now(),
+  }
+  await storage.set(transcriptManifestKey(sessionID, block.id), manifest as unknown as Schema.Json)
+  for (const [index, message] of messages.entries()) {
+    await storage.set(
+      transcriptMessageKey(sessionID, block.id, index),
+      toStoredMessage(message) as unknown as Schema.Json,
+    )
+  }
+}
+
+/** 输入：storage + 会话 + 块；输出：合法 manifest 或 undefined。 */
+export async function loadTranscriptManifest(
+  storage: StorageLike,
+  sessionID: string,
+  blockID: string,
+): Promise<TranscriptManifest | undefined> {
+  const raw = await storage.get(transcriptManifestKey(sessionID, blockID))
+  if (typeof raw !== "object" || raw === null) return undefined
+  const record = raw as Record<string, unknown>
+  if (
+    record["v"] !== 1 ||
+    typeof record["sessionID"] !== "string" ||
+    typeof record["blockID"] !== "string" ||
+    typeof record["fromId"] !== "string" ||
+    typeof record["toId"] !== "string" ||
+    typeof record["count"] !== "number" ||
+    !Number.isInteger(record["count"]) ||
+    record["count"] < 0
+  ) {
+    return undefined
+  }
+  return {
+    v: 1,
+    sessionID: record["sessionID"],
+    blockID: record["blockID"],
+    fromId: record["fromId"],
+    toId: record["toId"],
+    count: record["count"],
+    createdAt: typeof record["createdAt"] === "number" ? record["createdAt"] : 0,
+  }
+}
+
+/** 输入：storage + 会话 + 块 + 序号；输出：落盘消息或 undefined。 */
+export async function loadTranscriptMessage(
+  storage: StorageLike,
+  sessionID: string,
+  blockID: string,
+  index: number,
+): Promise<StoredMessage | undefined> {
+  const raw = await storage.get(transcriptMessageKey(sessionID, blockID, index))
+  if (typeof raw !== "object" || raw === null) return undefined
+  const record = raw as Record<string, unknown>
+  if (typeof record["role"] !== "string" || !Array.isArray(record["content"])) return undefined
+  return {
+    ...(typeof record["id"] === "string" ? { id: record["id"] } : {}),
+    role: record["role"],
+    content: record["content"] as StoredPart[],
+  }
+}
+
+/** 输入：storage + 会话；输出：无；预期：删掉该会话所有 manifest/message key。 */
+export async function deleteTranscriptsForSession(
+  storage: StorageLike & {
+    scan?: (options: { prefix: string; limit: number; after?: string }) => Promise<{
+      entries: readonly { key: string; value: unknown }[]
+      next?: string
+    }>
+  },
+  sessionID: string,
+): Promise<void> {
+  await forEachScannedEntry(storage, transcriptSessionPrefix(sessionID), async (entry) => {
+    await storage.remove(entry.key)
+  })
+}
+
+/** 输入：storage + 会话 + 块；输出：manifest 是否存在。 */
+export async function hasTranscript(
+  storage: StorageLike,
+  sessionID: string,
+  blockID: string,
+): Promise<boolean> {
+  return (await loadTranscriptManifest(storage, sessionID, blockID)) !== undefined
+}

package/src/compression/transcript.ts

@@ -1,63 +1,63 @@
-import { stringify, toolInput, toolResult, toolResultText } from "../shared/parts.ts"
-import type { ContentPartLike, MessageLike } from "./select.ts"
-
-/** 单条工具输出最多保留的字符数。 */
-const TOOL_RESULT_MAX = 2_000
-/** 工具调用参数最多保留的字符数。 */
-const TOOL_INPUT_MAX = 500
-/** 整份 transcript 的字符上限，防止侧调用自身溢出。 */
-export const TRANSCRIPT_MAX = 200_000
-
-function truncate(value: string, max: number): string {
-  return value.length <= max ? value : `${value.slice(0, max)}…[truncated]`
-}
-
-function renderMessage(message: MessageLike): string {
-  const lines: string[] = []
-  if (typeof message.text === "string" && message.text.length > 0) {
-    lines.push(message.role === "user" ? `[User]\n${message.text}` : message.text)
-  }
-  const content: readonly ContentPartLike[] = Array.isArray(message.content)
-    ? message.content
-    : []
-  for (const part of content) {
-    if (part.type === "text") {
-      const text = part.text ?? ""
-      if (text.length > 0) lines.push(message.role === "user" ? `[User]\n${text}` : text)
-      continue
-    }
-    if (part.type === "tool") {
-      lines.push(`[Tool call ${part.name ?? "?"}] ${truncate(stringify(toolInput(part)), TOOL_INPUT_MAX)}`)
-      lines.push(`[Tool result ${part.name ?? "?"}] ${truncate(toolResultText(toolResult(part)), TOOL_RESULT_MAX)}`)
-      continue
-    }
-    if (part.type === "tool-call") {
-      lines.push(`[Tool call ${part.name ?? "?"}] ${truncate(stringify(toolInput(part)), TOOL_INPUT_MAX)}`)
-      continue
-    }
-    if (part.type === "tool-result") {
-      lines.push(`[Tool result ${part.name ?? "?"}] ${truncate(toolResultText(toolResult(part)), TOOL_RESULT_MAX)}`)
-    }
-  }
-  return lines.join("\n")
-}
-
-/**
- * 输入：可压缩消息。
- * 输出：给侧调用使用的纯文本 transcript。
- * 预期：跳过 reasoning/media；工具输出截断；整体超限时丢弃最老的消息。
- */
-export function buildTranscript(messages: readonly MessageLike[]): string {
-  const rendered = messages.map(renderMessage).filter((text) => text.length > 0)
-  let total = rendered.reduce((sum, text) => sum + text.length + 2, 0)
-  let start = 0
-  while (total > TRANSCRIPT_MAX && start < rendered.length - 1) {
-    const item = rendered[start]
-    if (item === undefined) break
-    total -= item.length + 2
-    start += 1
-  }
-  const kept = rendered.slice(start)
-  const prefix = start > 0 ? `[older messages omitted: ${start.toString()}]\n\n` : ""
-  return prefix + kept.join("\n\n")
-}
+import { stringify, toolInput, toolResult, toolResultText } from "../shared/parts.ts"
+import type { ContentPartLike, MessageLike } from "./select.ts"
+
+/** 单条工具输出最多保留的字符数。 */
+const TOOL_RESULT_MAX = 2_000
+/** 工具调用参数最多保留的字符数。 */
+const TOOL_INPUT_MAX = 500
+/** 整份 transcript 的字符上限，防止侧调用自身溢出。 */
+export const TRANSCRIPT_MAX = 200_000
+
+function truncate(value: string, max: number): string {
+  return value.length <= max ? value : `${value.slice(0, max)}…[truncated]`
+}
+
+function renderMessage(message: MessageLike): string {
+  const lines: string[] = []
+  if (typeof message.text === "string" && message.text.length > 0) {
+    lines.push(message.role === "user" ? `[User]\n${message.text}` : message.text)
+  }
+  const content: readonly ContentPartLike[] = Array.isArray(message.content)
+    ? message.content
+    : []
+  for (const part of content) {
+    if (part.type === "text") {
+      const text = part.text ?? ""
+      if (text.length > 0) lines.push(message.role === "user" ? `[User]\n${text}` : text)
+      continue
+    }
+    if (part.type === "tool") {
+      lines.push(`[Tool call ${part.name ?? "?"}] ${truncate(stringify(toolInput(part)), TOOL_INPUT_MAX)}`)
+      lines.push(`[Tool result ${part.name ?? "?"}] ${truncate(toolResultText(toolResult(part)), TOOL_RESULT_MAX)}`)
+      continue
+    }
+    if (part.type === "tool-call") {
+      lines.push(`[Tool call ${part.name ?? "?"}] ${truncate(stringify(toolInput(part)), TOOL_INPUT_MAX)}`)
+      continue
+    }
+    if (part.type === "tool-result") {
+      lines.push(`[Tool result ${part.name ?? "?"}] ${truncate(toolResultText(toolResult(part)), TOOL_RESULT_MAX)}`)
+    }
+  }
+  return lines.join("\n")
+}
+
+/**
+ * 输入：可压缩消息。
+ * 输出：给侧调用使用的纯文本 transcript。
+ * 预期：跳过 reasoning/media；工具输出截断；整体超限时丢弃最老的消息。
+ */
+export function buildTranscript(messages: readonly MessageLike[]): string {
+  const rendered = messages.map(renderMessage).filter((text) => text.length > 0)
+  let total = rendered.reduce((sum, text) => sum + text.length + 2, 0)
+  let start = 0
+  while (total > TRANSCRIPT_MAX && start < rendered.length - 1) {
+    const item = rendered[start]
+    if (item === undefined) break
+    total -= item.length + 2
+    start += 1
+  }
+  const kept = rendered.slice(start)
+  const prefix = start > 0 ? `[older messages omitted: ${start.toString()}]\n\n` : ""
+  return prefix + kept.join("\n\n")
+}