cartethyia 2.7.9 → 2.8.1

package/src/providers/integrations/buddy/buddy-checkin.ts

@@ -1,521 +0,0 @@
-import { asCoercedNumber as sharedAsNumber, asRecord as sharedAsRecord } from "../../../shared/coerce";
-
-// Buddy-family daily check-in (daily login credit claim), shared by both
-// variants (`cb`, `cbcn`): same Tencent billing facade, same
-// status/claim contract — only the endpoint base and identity headers differ
-// per variant, and those already live in each provider's own modules.
-//
-// The desktop client claims a free daily credit grant once per calendar day.
-// The gateway mirrors that for every stored account so an operator can see,
-// per account, whether today's grant was claimed — without opening the app.
-//
-// Contract (from the reverse-engineered desktop bundle, vendor/research):
-//   Status : POST {base}/v2/billing/meter/checkin-activity-status  body {}
-//   Claim  : POST {base}/v2/billing/meter/daily-checkin            body {}
-//   Envelope: { code: 0, data: { today_checked_in, streak_days, ... } }
-//   "already claimed" business code is 1001 (NOT 10001 — that belongs to the
-//   artifact-release API).
-//
-// One request per account per day: the sweep asks the status endpoint first and
-// only POSTs the claim when the account has not claimed today. A successful
-// claim short-circuits the day, so a restart cannot double-post.
-
-import { providerBaseUrl } from "../../provider-metadata";
-import { sleep } from "../../../shared/sleep";
-import { BUDDY_PROVIDER_IDS } from "../../provider-metadata";
-import { QODER_CHECKIN_PROVIDER_IDS } from "../qoder/qoder-checkin";
-import { buildCodeBuddyUserAgent } from "../../operations/client-versions";
-import { codebuddyDomain, type CodeBuddyVariant } from "./codebuddy/shared";
-/** Providers whose billing facade exposes the daily check-in routes. */
-/**
- * The Buddy providers the daily check-in claims a credit for. Derived from the
- * shared family set so this list cannot drift from the health recorder's or the
- * request preparer's.
- */
-export const DAILY_CHECKIN_PROVIDER_IDS: readonly string[] = [
-  ...BUDDY_PROVIDER_IDS,
-  ...QODER_CHECKIN_PROVIDER_IDS,
-];
-
-/** Abort budget for one upstream check-in request. */
-const TIMEOUT_MS = 15_000;
-
-export const CHECKIN_STATUS_PATH = "/billing/meter/checkin-activity-status";
-export const CHECKIN_CLAIM_PATH = "/billing/meter/daily-checkin";
-
-export type DailyCheckinState =
-  | "claimed"
-  | "already_claimed"
-  | "not_eligible"
-  | "event_ended"
-  | "unavailable"
-  | "error";
-
-export interface DailyCheckinResult {
-  readonly providerId: string;
-  readonly accountId: string;
-  readonly state: DailyCheckinState;
-  /** Credits granted by this claim; `null` when nothing was granted. */
-  readonly credit: number | null;
-  readonly streakDays: number | null;
-  /** Provider-facing detail for the audit log; `null` on success. */
-  readonly error: string | null;
-}
-
-/**
- * One `chat_request_send` growth event for `POST {base}/v2/report`.
- *
- * The growth backend scores conversation activity, not credit grants: one
- * report lights the login streak and unlocks the `first_buddy` task. The full
- * client shape is sent (not a minimal subset) because the upstream may tighten
- * validation later. `userId` is mandatory — the server answers 200 but silently
- * drops events without it.
- */
-export interface BuddyActivityReportEvent {
-  readonly eventCode: "chat_request_send";
-  readonly timestamp: number;
-  readonly reportDelay: number;
-  readonly mode: string;
-  readonly conversationId: string;
-  readonly requestId: string;
-  readonly inputLength: number;
-  readonly requestModelId: string;
-  readonly requestModelName: string;
-  readonly isPlan: boolean;
-  readonly isAutoExecuteTerminal: boolean;
-  readonly isAutoModify: boolean;
-  readonly codebaseEnable: boolean;
-  readonly maxToken: number;
-  readonly maxSteps: number;
-  readonly temperature: number;
-  readonly maxRetries: number;
-  readonly mentionContexts: readonly unknown[];
-  readonly knowledgeId: readonly unknown[];
-  readonly knowledgeName: readonly unknown[];
-  readonly codebaseId: string;
-  readonly mentionContextCount: number;
-  readonly command: string;
-  readonly recommendId: string;
-  readonly skillId: string;
-  readonly skillCount: number;
-  readonly totalCount: number;
-  readonly presentAt: number;
-  readonly rootRequestId: string;
-  readonly parentConversationId: string;
-  readonly agentName: string;
-  readonly agentType: string;
-  readonly product: string;
-  readonly userId: string;
-}
-
-export const ACTIVITY_REPORT_PATH = "/report";
-export type BuddyActivityReportState = "reported" | "unavailable" | "error";
-
-export interface BuddyActivityReportResult {
-  readonly providerId: string;
-  readonly accountId: string;
-  readonly state: BuddyActivityReportState;
-  /** Provider-facing detail for the audit log; `null` on success. */
-  readonly error: string | null;
-}
-
-/**
- * Builds one growth-activity event. `conversationId` is caller-generated (a
- * synthetic `activity-<millis>` id is fine — the server does not validate it
- * against a real conversation); `requestId` defaults to it. The model fields
- * mirror the desktop client's own report so the event reads as a normal chat
- * send, and `userId` carries the account uid the streak is scored against.
- */
-export function buildActivityReportEvent(args: {
-  readonly conversationId: string;
-  readonly userId: string;
-  readonly requestId?: string;
-  readonly now?: number;
-}): BuddyActivityReportEvent {
-  const now = args.now ?? Date.now();
-  const requestId = args.requestId ?? args.conversationId;
-  return {
-    eventCode: "chat_request_send",
-    timestamp: now,
-    reportDelay: 0,
-    mode: "unknown",
-    conversationId: args.conversationId,
-    requestId,
-    inputLength: 0,
-    requestModelId: "auto",
-    requestModelName: "Auto",
-    isPlan: false,
-    isAutoExecuteTerminal: false,
-    isAutoModify: false,
-    codebaseEnable: false,
-    maxToken: 0,
-    maxSteps: 0,
-    temperature: 0,
-    maxRetries: 0,
-    mentionContexts: [],
-    knowledgeId: [],
-    knowledgeName: [],
-    codebaseId: "",
-    mentionContextCount: 0,
-    command: "",
-    recommendId: "",
-    skillId: "",
-    skillCount: 0,
-    totalCount: 0,
-    presentAt: now,
-    rootRequestId: args.conversationId,
-    parentConversationId: args.conversationId,
-    agentName: "default",
-    agentType: "main",
-    product: "SaaS",
-    userId: args.userId,
-  };
-}
-
-/**
- * How the growth report is sent.
- *
- * The billing facade scores conversation activity, and one event reads as one
- * chat send. A single event did not move the growth tasks; four events per
- * batch, sent twice with a pause between the rounds, is what does. The pause
- * is what keeps the two rounds from landing as one burst. Both numbers apply
- * to every buddy variant — the request is built the same way for `cb` and
- * `cbcn`, so neither can drift back to a single shot.
- */
-export const ACTIVITY_REPORT_EVENTS_PER_BATCH = 4;
-export const ACTIVITY_REPORT_ROUNDS = 2;
-export const ACTIVITY_REPORT_ROUND_DELAY_MS = 1_500;
-
-/**
- * Sends the growth-activity report (`POST {base}/v2/report`, the versioned
- * billing-meter path joined the same way as the status/claim routes:
- * `VERSION_SEGMENT` supplies nothing for `cb`/`cbcn`, which already end in
- * `/v2`).
- *
- * Two rounds of four events, paused between rounds. Each event carries its
- * own conversation and request id: the facade drops a repeated id as a
- * duplicate, so a batch of identical events would score as one. The first
- * round that fails stops the sequence — a credential or transport failure
- * will not be fixed by the second round, and sending it only adds load.
- */
-export async function fetchBuddyActivityReport(args: {
-  readonly providerId: string;
-  readonly accountId: string;
-  readonly credential: string;
-  readonly userId: string;
-  readonly fetcher: typeof fetch;
-  readonly conversationId?: string;
-  /** Pause between rounds. Injectable so a test does not wait on the real delay. */
-  readonly roundDelayMs?: number;
-  readonly sleep?: (ms: number) => Promise<void>;
-}): Promise<BuddyActivityReportResult> {
-  const { providerId, accountId, credential, userId, fetcher } = args;
-  const base = { providerId, accountId, error: null };
-  if (!credential) {
-    return { ...base, state: "error", error: "Account has no stored credential." };
-  }
-  if (!userId) {
-    return { ...base, state: "error", error: "Account uid is unavailable; cannot report activity." };
-  }
-  const headers = {
-    ...(await checkinHeaders(providerId)),
-    Authorization: `Bearer ${credential}`,
-    "X-User-Id": userId,
-  };
-  const url = `${providerBaseUrl(providerId)}${VERSION_SEGMENT[providerId] ?? ""}${ACTIVITY_REPORT_PATH}`;
-  const delayMs = args.roundDelayMs ?? ACTIVITY_REPORT_ROUND_DELAY_MS;
-  const sleep = args.sleep ?? ((ms: number) => new Promise((resolve) => setTimeout(resolve, ms)));
-  const conversationId = args.conversationId ?? `activity-${Date.now()}`;
-
-  for (let round = 0; round < ACTIVITY_REPORT_ROUNDS; round += 1) {
-    if (round > 0 && delayMs > 0) await sleep(delayMs);
-    const events = Array.from({ length: ACTIVITY_REPORT_EVENTS_PER_BATCH }, (_, index) => {
-      const sequence = round * ACTIVITY_REPORT_EVENTS_PER_BATCH + index;
-      return buildActivityReportEvent({
-        conversationId: `${conversationId}-${sequence}`,
-        requestId: `${conversationId}-req-${sequence}`,
-        userId,
-      });
-    });
-    let payload: unknown;
-    try {
-      const response = await fetcher(url, {
-        method: "POST",
-        headers,
-        body: JSON.stringify(events),
-        signal: AbortSignal.timeout(TIMEOUT_MS),
-      });
-      if (response.status === 401 || response.status === 403) {
-        return { ...base, state: "error", error: "Credential invalid or expired." };
-      }
-      if (!response.ok) {
-        return { ...base, state: "error", error: `Activity report API error (${response.status}).` };
-      }
-      payload = await response.text().then((text) => (text ? JSON.parse(text) : null));
-    } catch (error) {
-      if (error instanceof SyntaxError) {
-        return { ...base, state: "error", error: "Activity report response returned invalid JSON." };
-      }
-      return {
-        ...base,
-        state: "error",
-        error: error instanceof Error ? error.message : "Activity report request failed.",
-      };
-    }
-    const code = envelopeCode(payload);
-    if (code !== null && code !== 0) {
-      return { ...base, state: "error", error: `Activity report error (code ${code}).` };
-    }
-  }
-  return { ...base, state: "reported" };
-}
-
-/**
- * Check-in business codes, per the desktop bundle's `mapCheckinStatus()`.
- * `1001` is "already claimed" — NOT `10001`, which belongs to the
- * artifact-release API and never appears in this path.
- */
-const ALREADY_CLAIMED_CODE = 1001;
-const NOT_ELIGIBLE_CODE = 1002;
-const EVENT_ENDED_CODE = 1003;
-/** Delays between 5xx/network retries on the claim POST, mirroring the buddy gateway's own backoff. */
-const CLAIM_RETRY_DELAYS_MS = [1_000, 2_000] as const;
-/**
- * Check-in paths are joined onto the provider's manifest base URL, which does
- * NOT carry a consistent version segment: `cb`/`cbcn` already end in `/v2`,
- * so no segment is prepended. This map keeps the join aligned with each
- * provider's own quota collector, so the gateways cannot drift apart.
- */
-const VERSION_SEGMENT: Readonly<Record<string, string>> = {
-  cb: "",
-  cbcn: "",
-};
-
-function statusUrl(providerId: string): string {
-  return `${providerBaseUrl(providerId)}${VERSION_SEGMENT[providerId] ?? ""}${CHECKIN_STATUS_PATH}`;
-}
-
-function claimUrl(providerId: string): string {
-  return `${providerBaseUrl(providerId)}${VERSION_SEGMENT[providerId] ?? ""}${CHECKIN_CLAIM_PATH}`;
-}
-
-/**
- * Per-provider billing headers. The Buddy gateways share the Tencent billing
- * facade but each risk-gates on its own client identity: CodeBuddy intl
- * wants the `IDE` identity, and CN wants `CLI` (plus its own `X-Domain`).
- * Sending the wrong one trips a 403 the same way the chat surface does.
- */
-async function checkinHeaders(providerId: string): Promise<Record<string, string>> {
-  const common = {
-    "X-Product": "SaaS",
-    "X-Requested-With": "XMLHttpRequest",
-    "X-CodeBuddy-Request": "1",
-    "Content-Type": "application/json",
-    Accept: "application/json",
-  };
-  const variant: CodeBuddyVariant = providerId === "cb" ? "IDE" : "CLI";
-  return {
-    ...common,
-    "User-Agent": buildCodeBuddyUserAgent(variant),
-    "X-IDE-Type": variant,
-    "X-IDE-Name": variant,
-    "X-Domain": codebuddyDomain(variant),
-  };
-}
-
-/**
- * Textual "already checked in" fingerprints. The claim POST answers `1001`
- * when the envelope is well-formed, but a locale or gateway variant can answer
- * the same state as prose (the reference gateway matches `已签到` and
- * `already…check…` case-insensitively across the message or raw body), so both
- * shapes resolve to `already_claimed` instead of a spurious error.
- */
-function alreadyClaimedMessage(payload: unknown, rawBody: string): string | null {
-  const message = asTextField(payload) || rawBody.trim();
-  if (!message) return null;
-  if (message.includes("已签到")) return message;
-  const lower = message.toLowerCase();
-  if (lower.includes("already") && lower.includes("check")) return message;
-  return null;
-}
-
-function asTextField(payload: unknown): string | null {
-  const data = asRecord(payload);
-  if (!data) return null;
-  const direct = data["msg"] ?? data["message"];
-  if (typeof direct === "string" && direct.trim()) return direct.trim();
-  return null;
-}
-
-/** True when the claim attempt is worth retrying: 5xx or a network throw, never a business code. */
-function isRetryableClaim(error: unknown, status: number | null): boolean {
-  if (status !== null) return status >= 500;
-  return !(error instanceof DOMException && error.name === "AbortError");
-}
-
-function asRecord(value: unknown): Record<string, unknown> | null {
-  return sharedAsRecord(value) ?? null;
-}
-
-function asNumber(value: unknown): number | null {
-  return sharedAsNumber(value) ?? null;
-}
-
-
-/** Reads the envelope's business code; `0` means success. */
-function envelopeCode(payload: unknown): number | null {
-  const envelope = asRecord(payload);
-  if (!envelope || !("code" in envelope)) return null;
-  return asNumber(envelope["code"]);
-}
-
-/**
- * Unwraps the check-in payload. The gateway returns a flat `{ code, data }`,
- * while the CodeBuddy billing facade nests Tencent's `data.Response.Data`
- * (the same double wrap its quota endpoint uses), so both shapes resolve.
- */
-function envelopeData(payload: unknown): Record<string, unknown> | null {
-  const data = asRecord(asRecord(payload)?.["data"]);
-  if (!data) return null;
-  if ("today_checked_in" in data || "active" in data || "streak_days" in data) return data;
-  return asRecord(asRecord(data["Response"])?.["Data"]);
-}
-
-/**
- * Performs one account's daily check-in.
- *
- * Status first, claim only when `today_checked_in` is false — so the daily
- * budget is one request for an already-claimed account and two on claim day.
- */
-export async function fetchDailyCheckin(args: {
-  readonly providerId: string;
-  readonly accountId: string;
-  readonly credential: string;
-  readonly fetcher: typeof fetch;
-  /** Overrides the status probe when the caller already knows the state. */
-  readonly forceClaim?: boolean;
-}): Promise<DailyCheckinResult> {
-  const { providerId, accountId, credential, fetcher } = args;
-  const base = { providerId, accountId, credit: null, streakDays: null, error: null };
-
-  if (!credential) {
-    return { ...base, state: "error", error: "Account has no stored credential." };
-  }
-
-  // `checkinHeaders()` already carries Content-Type/Accept; only auth is
-  // layered on here so the per-provider identity stays in one place.
-  const headers = {
-    ...(await checkinHeaders(providerId)),
-    Authorization: `Bearer ${credential}`,
-  };
-
-  let status: unknown;
-  try {
-    const response = await fetcher(statusUrl(providerId), {
-      method: "POST",
-      headers,
-      body: "{}",
-      signal: AbortSignal.timeout(TIMEOUT_MS),
-    });
-    if (response.status === 401 || response.status === 403) {
-      return { ...base, state: "error", error: "Credential invalid or expired." };
-    }
-    if (!response.ok) {
-      return { ...base, state: "error", error: `Check-in status API error (${response.status}).` };
-    }
-    status = await response.json().catch(() => null);
-  } catch (error) {
-    return {
-      ...base,
-      state: "error",
-      error: error instanceof Error ? error.message : "Check-in status request failed.",
-    };
-  }
-
-  const statusCode = envelopeCode(status);
-  if (statusCode !== null && statusCode !== 0) {
-    return { ...base, state: "error", error: `Check-in status error (code ${statusCode}).` };
-  }
-
-  const data = envelopeData(status);
-  if (!data) return { ...base, state: "unavailable", error: "Check-in activity unavailable." };
-  if (data["active"] === false) {
-    return { ...base, state: "unavailable", error: "No check-in activity running." };
-  }
-  const streak = asNumber(data["streak_days"]);
-  if (!args.forceClaim && data["today_checked_in"] === true) {
-    return { ...base, state: "already_claimed", streakDays: streak };
-  }
-
-  let claim: unknown;
-  let claimRaw = "";
-  let attempt = 0;
-  for (;;) {
-    try {
-      const response = await fetcher(claimUrl(providerId), {
-        method: "POST",
-        headers,
-        body: "{}",
-        signal: AbortSignal.timeout(TIMEOUT_MS),
-      });
-      if (response.status === 401 || response.status === 403) {
-        return { ...base, state: "error", error: "Credential invalid or expired." };
-      }
-      if (response.status >= 500 && attempt < CLAIM_RETRY_DELAYS_MS.length) {
-        await sleep(CLAIM_RETRY_DELAYS_MS[attempt] as number);
-        attempt += 1;
-        continue;
-      }
-      if (!response.ok) {
-        return { ...base, state: "error", error: `Daily check-in API error (${response.status}).` };
-      }
-      claimRaw = await response.text().catch(() => "");
-      claim = claimRaw ? JSON.parse(claimRaw) : null;
-      break;
-    } catch (error) {
-      if (error instanceof SyntaxError) {
-        return { ...base, state: "error", error: "Daily check-in response returned invalid JSON." };
-      }
-      if (isRetryableClaim(error, null) && attempt < CLAIM_RETRY_DELAYS_MS.length) {
-        await sleep(CLAIM_RETRY_DELAYS_MS[attempt] as number);
-        attempt += 1;
-        continue;
-      }
-      return {
-        ...base,
-        state: "error",
-        error: error instanceof Error ? error.message : "Daily check-in request failed.",
-      };
-    }
-  }
-
-  const claimCode = envelopeCode(claim);
-  if (claimCode !== null && claimCode !== 0) {
-    const alreadyText = alreadyClaimedMessage(claim, claimRaw);
-    if (claimCode === ALREADY_CLAIMED_CODE || alreadyText) {
-      return { ...base, state: "already_claimed", streakDays: streak };
-    }
-    if (claimCode === NOT_ELIGIBLE_CODE) return { ...base, state: "not_eligible" };
-    if (claimCode === EVENT_ENDED_CODE) return { ...base, state: "event_ended" };
-    return { ...base, state: "error", error: `Daily check-in error (code ${claimCode}).` };
-  }
-  // Some gateway variants answer the duplicate-claim state as prose with a
-  // success envelope; catch that before reporting a fresh claim.
-  if (claimCode === 0) {
-    const alreadyText = alreadyClaimedMessage(claim, claimRaw);
-    if (alreadyText) return { ...base, state: "already_claimed", streakDays: streak };
-  }
-
-  const claimData = envelopeData(claim);
-  return {
-    ...base,
-    state: "claimed",
-    credit: asNumber(claimData?.["credit"]) ?? asNumber(claimData?.["today_credit"]) ?? null,
-    streakDays: asNumber(claimData?.["streak_days"]) ?? streak,
-  };
-}
-
-/** Calendar-day key in UTC, matching the provider's own day boundary. */
-export function checkinDayKey(now: Date = new Date()): string {
-  return now.toISOString().slice(0, 10);
-}

package/src/providers/integrations/buddy/buddy-cli-headers.ts

@@ -1,76 +0,0 @@
-/**
- * The header family only the CLI identity sends.
- *
- * Everything here comes from the two reference gateways, which agree: the
- * desktop product sends conversation/tracing headers and a device token, while
- * the CLI sends the agent envelope. The CLI still drops the browser *transport*
- * headers (`origin`, `referer`, `accept-language`) and the device token, but
- * keeps the request-shape markers (`x-requested-with`, `x-codebuddy-request`)
- * the reference sends on every call regardless of transport — see the CLI
- * branches in each provider's `*-shared.ts`.
- *
- * The CLI's cloud chat transport is axios (`RestOperations`), not the OpenAI
- * Node SDK, so it sends **no** `x-stainless-*` fingerprint. The reference's
- * chat request helper sets only `Content-Type`, `Accept`, and `Authorization`
- * before the conversation/agent/identity headers are injected; the stainless
- * set belongs to the SDK-backed secondary clients, not to CodeBuddy chat.
- */
-
-/**
- * The trace family the CLI carries as a child span.
- *
- * The desktop product omits the parent (`with_parent=False` in the reference);
- * the CLI conversation request is a child span, so it also sends `b3`'s fourth
- * field and `X-B3-ParentSpanId`. `buddyConversationHeaders` already produces
- * `traceparent`, `b3`, and the `X-B3-*` set, so only the parent is added here.
- */
-export function buddyCliTraceHeaders(
-  conversation: Readonly<Record<string, string>>,
-): Record<string, string> {
-  const traceId = conversation["x-trace-id"] ?? conversation["x-b3-traceid"];
-  const spanId = conversation["x-b3-spanid"];
-  if (traceId === undefined) return {};
-  const b3 = `${traceId}-${spanId ?? traceId.slice(0, 16)}-1-${traceId.slice(0, 16)}`;
-  return { ...(spanId === undefined ? {} : { "X-B3-ParentSpanId": spanId }), b3 };
-}
-
-/**
- * The full CLI identity envelope, minus what each caller already supplies.
- *
- * The caller keeps ownership of `User-Agent`, `X-IDE-*`, `X-Domain`, the
- * conversation family, and the OAuth identity headers; only the headers that
- * are the same for every CLI dispatch are built here.
- *
- * No `x-stainless-*` fingerprint: the CLI's chat transport is axios, not the
- * OpenAI Node SDK. `X-Private-Data` is added by the caller only when the
- * dispatch is actually private — the reference never sends it as `false`.
- */
-export function buddyCliHeaders(args: {
-  readonly userId: string;
-  readonly enterpriseId?: string | undefined;
-  readonly privateData?: boolean;
-}): Record<string, string> {
-  const enterpriseId =
-    args.enterpriseId !== undefined && args.enterpriseId.trim().length > 0
-      ? args.enterpriseId.trim()
-      : undefined;
-  return {
-    "X-Agent-Intent": "craft",
-    "X-Agent-Type": "main",
-    // The reference enum is local-agent / cloud-agent / agent-agent, and its
-    // parser returns "other" for anything unknown or absent. A gateway is none
-    // of the three, so "other" is the honest value rather than a guess.
-    "X-Agent-Runtime": "other",
-    "X-User-Id": args.userId,
-    ...(args.privateData === true ? { "X-Private-Data": "true" } : {}),
-    ...(enterpriseId === undefined
-      ? {}
-      : {
-          "X-Enterprise-Id": enterpriseId,
-          // The tenant falls back to the enterprise: the two are the same value
-          // for an account that belongs to one enterprise, and the upstream
-          // reads both.
-          "X-Tenant-Id": enterpriseId,
-        }),
-  };
-}