paperclip/server/src/__tests__/gemini-local-adapter.test.ts

215 lines
6.5 KiB
TypeScript
Raw Normal View History

import { describe, expect, it, vi } from "vitest";
[codex] Retry max-turn exhausted heartbeats (#5096) ## Thinking Path > - Paperclip orchestrates AI agents for autonomous companies, and heartbeat execution is the control-plane loop that keeps assigned work moving. > - Max-turn exhaustion is a recoverable local-adapter stop condition for Claude and Gemini agents when a run needs another heartbeat to continue safely. > - The previous behavior could leave max-turn continuation details hard to inspect, and duplicate/stale continuation wakes could keep running after issue state changed. > - The adapter layer also needed to avoid trusting arbitrary stdout/stderr text as scheduler control metadata. > - This pull request adds bounded max-turn continuation scheduling, visible retry state, structured stop metadata handling, and stale/duplicate continuation guards. > - The benefit is safer automatic continuation after max-turn stops, clearer operator visibility, and fewer duplicate or stale agent runs. ## What Changed - Replaces closed PR #4952, whose head repository was deleted. - Rebases the recovered max-turn continuation branch onto current `paperclipai/paperclip:master`. - Adds max-turn continuation scheduling and retry-state plumbing for heartbeat runs. - Adds stale/duplicate continuation suppression when issue status, ownership, or execution locks change. - Normalizes Claude/Gemini max-turn detection around structured stop metadata instead of unstructured stdout/stderr text. - Surfaces max-turn continuation settings and retry visibility in the board UI. - Adds focused server, adapter, and UI tests for max-turn stop metadata, retry scheduling, stale queued-run invalidation, adapter parsing/execution, run ledger display, and agent config patching. ## Verification - `pnpm install --no-frozen-lockfile` to refresh local dependencies after rebasing onto current `master`. - `pnpm run preflight:workspace-links && pnpm exec vitest run server/src/__tests__/claude-local-adapter.test.ts server/src/__tests__/claude-local-execute.test.ts server/src/__tests__/gemini-local-adapter.test.ts server/src/__tests__/gemini-local-execute.test.ts server/src/__tests__/heartbeat-retry-scheduling.test.ts server/src/__tests__/heartbeat-stale-queue-invalidation.test.ts server/src/services/heartbeat-stop-metadata.test.ts ui/src/components/IssueRunLedger.test.tsx ui/src/lib/agent-config-patch.test.ts ui/src/lib/runRetryState.test.ts --testTimeout=20000` - `pnpm --filter @paperclipai/adapter-claude-local typecheck && pnpm --filter @paperclipai/adapter-gemini-local typecheck && pnpm --filter @paperclipai/server typecheck && pnpm --filter @paperclipai/ui typecheck` - UI screenshot note: the UI changes are limited to config/ledger state rendering rather than layout changes; component/unit coverage above verifies the rendered behavior. ## Risks - Medium behavior risk: heartbeat retry gating now suppresses max-turn continuations when issue state or execution locks drift, so any callers that relied on stale continuations running will now see cancellation instead. - Low adapter risk: Claude/Gemini unstructured text no longer triggers max-turn scheduler metadata, so only structured stop signals and Gemini exit code 53 are trusted. - No database migrations. > For core feature work, check [`ROADMAP.md`](ROADMAP.md) first and discuss it in `#dev` before opening the PR. Feature PRs that overlap with planned core work may need to be redirected — check the roadmap first. See `CONTRIBUTING.md`. ## Model Used - OpenAI Codex coding agent, GPT-5-class model, tool-enabled local repository editing and command execution. ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have run tests locally and they pass - [x] I have added or updated tests where applicable - [x] If this change affects the UI, I have included before/after screenshots (not applicable: state/default rendering only; covered by component/unit tests) - [x] I have updated relevant documentation to reflect my changes (not applicable: no user-facing command or docs contract changed) - [x] I have considered and documented any risks above - [x] I will address all Greptile and reviewer comments before requesting merge --------- Co-authored-by: Paperclip <noreply@paperclip.ing>
2026-05-03 11:30:48 -05:00
import {
isGeminiTurnLimitResult,
isGeminiUnknownSessionError,
parseGeminiJsonl,
} from "@paperclipai/adapter-gemini-local/server";
import { parseGeminiStdoutLine } from "@paperclipai/adapter-gemini-local/ui";
import { printGeminiStreamEvent } from "@paperclipai/adapter-gemini-local/cli";
describe("gemini_local parser", () => {
it("extracts session, summary, usage, cost, and terminal error message", () => {
const stdout = [
JSON.stringify({ type: "system", subtype: "init", session_id: "gemini-session-1", model: "gemini-2.5-pro" }),
JSON.stringify({
type: "assistant",
message: {
content: [{ type: "output_text", text: "hello" }],
},
}),
JSON.stringify({
type: "result",
subtype: "success",
session_id: "gemini-session-1",
usage: {
promptTokenCount: 12,
cachedContentTokenCount: 3,
candidatesTokenCount: 7,
},
total_cost_usd: 0.00123,
result: "done",
}),
JSON.stringify({ type: "error", message: "model access denied" }),
].join("\n");
const parsed = parseGeminiJsonl(stdout);
expect(parsed.sessionId).toBe("gemini-session-1");
expect(parsed.summary).toBe("hello");
expect(parsed.usage).toEqual({
inputTokens: 12,
cachedInputTokens: 3,
outputTokens: 7,
});
expect(parsed.costUsd).toBeCloseTo(0.00123, 6);
expect(parsed.errorMessage).toBe("model access denied");
});
it("extracts structured questions", () => {
const stdout = [
JSON.stringify({
type: "assistant",
message: {
content: [
{ type: "output_text", text: "I have a question." },
{
type: "question",
prompt: "Which model?",
choices: [
{ key: "pro", label: "Gemini Pro", description: "Better" },
{ key: "flash", label: "Gemini Flash" },
],
},
],
},
}),
].join("\n");
const parsed = parseGeminiJsonl(stdout);
expect(parsed.summary).toBe("I have a question.");
expect(parsed.question).toEqual({
prompt: "Which model?",
choices: [
{ key: "pro", label: "Gemini Pro", description: "Better" },
{ key: "flash", label: "Gemini Flash", description: undefined },
],
});
});
});
describe("gemini_local stale session detection", () => {
it("treats missing session messages as an unknown session error", () => {
expect(isGeminiUnknownSessionError("", "unknown session id abc")).toBe(true);
expect(isGeminiUnknownSessionError("", "checkpoint latest not found")).toBe(true);
});
});
[codex] Retry max-turn exhausted heartbeats (#5096) ## Thinking Path > - Paperclip orchestrates AI agents for autonomous companies, and heartbeat execution is the control-plane loop that keeps assigned work moving. > - Max-turn exhaustion is a recoverable local-adapter stop condition for Claude and Gemini agents when a run needs another heartbeat to continue safely. > - The previous behavior could leave max-turn continuation details hard to inspect, and duplicate/stale continuation wakes could keep running after issue state changed. > - The adapter layer also needed to avoid trusting arbitrary stdout/stderr text as scheduler control metadata. > - This pull request adds bounded max-turn continuation scheduling, visible retry state, structured stop metadata handling, and stale/duplicate continuation guards. > - The benefit is safer automatic continuation after max-turn stops, clearer operator visibility, and fewer duplicate or stale agent runs. ## What Changed - Replaces closed PR #4952, whose head repository was deleted. - Rebases the recovered max-turn continuation branch onto current `paperclipai/paperclip:master`. - Adds max-turn continuation scheduling and retry-state plumbing for heartbeat runs. - Adds stale/duplicate continuation suppression when issue status, ownership, or execution locks change. - Normalizes Claude/Gemini max-turn detection around structured stop metadata instead of unstructured stdout/stderr text. - Surfaces max-turn continuation settings and retry visibility in the board UI. - Adds focused server, adapter, and UI tests for max-turn stop metadata, retry scheduling, stale queued-run invalidation, adapter parsing/execution, run ledger display, and agent config patching. ## Verification - `pnpm install --no-frozen-lockfile` to refresh local dependencies after rebasing onto current `master`. - `pnpm run preflight:workspace-links && pnpm exec vitest run server/src/__tests__/claude-local-adapter.test.ts server/src/__tests__/claude-local-execute.test.ts server/src/__tests__/gemini-local-adapter.test.ts server/src/__tests__/gemini-local-execute.test.ts server/src/__tests__/heartbeat-retry-scheduling.test.ts server/src/__tests__/heartbeat-stale-queue-invalidation.test.ts server/src/services/heartbeat-stop-metadata.test.ts ui/src/components/IssueRunLedger.test.tsx ui/src/lib/agent-config-patch.test.ts ui/src/lib/runRetryState.test.ts --testTimeout=20000` - `pnpm --filter @paperclipai/adapter-claude-local typecheck && pnpm --filter @paperclipai/adapter-gemini-local typecheck && pnpm --filter @paperclipai/server typecheck && pnpm --filter @paperclipai/ui typecheck` - UI screenshot note: the UI changes are limited to config/ledger state rendering rather than layout changes; component/unit coverage above verifies the rendered behavior. ## Risks - Medium behavior risk: heartbeat retry gating now suppresses max-turn continuations when issue state or execution locks drift, so any callers that relied on stale continuations running will now see cancellation instead. - Low adapter risk: Claude/Gemini unstructured text no longer triggers max-turn scheduler metadata, so only structured stop signals and Gemini exit code 53 are trusted. - No database migrations. > For core feature work, check [`ROADMAP.md`](ROADMAP.md) first and discuss it in `#dev` before opening the PR. Feature PRs that overlap with planned core work may need to be redirected — check the roadmap first. See `CONTRIBUTING.md`. ## Model Used - OpenAI Codex coding agent, GPT-5-class model, tool-enabled local repository editing and command execution. ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have run tests locally and they pass - [x] I have added or updated tests where applicable - [x] If this change affects the UI, I have included before/after screenshots (not applicable: state/default rendering only; covered by component/unit tests) - [x] I have updated relevant documentation to reflect my changes (not applicable: no user-facing command or docs contract changed) - [x] I have considered and documented any risks above - [x] I will address all Greptile and reviewer comments before requesting merge --------- Co-authored-by: Paperclip <noreply@paperclip.ing>
2026-05-03 11:30:48 -05:00
describe("gemini_local turn-limit detection", () => {
it("detects structured turn-limit signals and exit code 53", () => {
expect(isGeminiTurnLimitResult({ status: "turn_limit" })).toBe(true);
expect(isGeminiTurnLimitResult({ stopReason: "max_turns_exhausted" })).toBe(true);
expect(isGeminiTurnLimitResult(null, 53)).toBe(true);
});
it("checks every structured stop field for turn-limit exhaustion", () => {
expect(
isGeminiTurnLimitResult({
status: "success",
stopReason: "turn_limit_exhausted",
}),
).toBe(true);
});
it("does not detect turn-limit exhaustion from unstructured error text", () => {
expect(isGeminiTurnLimitResult({ error: "max_turns reached" })).toBe(false);
});
});
describe("gemini_local ui stdout parser", () => {
it("parses assistant, thinking, and result events", () => {
const ts = "2026-03-08T00:00:00.000Z";
expect(
parseGeminiStdoutLine(
JSON.stringify({
type: "assistant",
message: {
content: [
{ type: "output_text", text: "I checked the repo." },
{ type: "thinking", text: "Reviewing adapter registry" },
{ type: "tool_call", name: "shell", input: { command: "ls -1" } },
{ type: "tool_result", tool_use_id: "tool_1", output: "AGENTS.md\n", status: "ok" },
],
},
}),
ts,
),
).toEqual([
{ kind: "assistant", ts, text: "I checked the repo." },
{ kind: "thinking", ts, text: "Reviewing adapter registry" },
{ kind: "tool_call", ts, name: "shell", input: { command: "ls -1" } },
{ kind: "tool_result", ts, toolUseId: "tool_1", content: "AGENTS.md\n", isError: false },
]);
expect(
parseGeminiStdoutLine(
JSON.stringify({
type: "result",
subtype: "success",
result: "Done",
usage: {
promptTokenCount: 10,
candidatesTokenCount: 5,
cachedContentTokenCount: 2,
},
total_cost_usd: 0.00042,
is_error: false,
}),
ts,
),
).toEqual([
{
kind: "result",
ts,
text: "Done",
inputTokens: 10,
outputTokens: 5,
cachedTokens: 2,
costUsd: 0.00042,
subtype: "success",
isError: false,
errors: [],
},
]);
});
});
function stripAnsi(value: string): string {
return value.replace(/\x1b\[[0-9;]*m/g, "");
}
describe("gemini_local cli formatter", () => {
it("prints init, assistant, result, and error events", () => {
const spy = vi.spyOn(console, "log").mockImplementation(() => {});
let joined = "";
try {
printGeminiStreamEvent(
JSON.stringify({ type: "system", subtype: "init", session_id: "gemini-session-1", model: "gemini-2.5-pro" }),
false,
);
printGeminiStreamEvent(
JSON.stringify({
type: "assistant",
message: { content: [{ type: "output_text", text: "hello" }] },
}),
false,
);
printGeminiStreamEvent(
JSON.stringify({
type: "result",
subtype: "success",
usage: {
promptTokenCount: 10,
candidatesTokenCount: 5,
cachedContentTokenCount: 2,
},
total_cost_usd: 0.00042,
}),
false,
);
printGeminiStreamEvent(
JSON.stringify({ type: "error", message: "boom" }),
false,
);
joined = spy.mock.calls.map((call) => stripAnsi(call.join(" "))).join("\n");
} finally {
spy.mockRestore();
}
expect(joined).toContain("Gemini init");
expect(joined).toContain("assistant: hello");
expect(joined).toContain("tokens: in=10 out=5 cached=2 cost=$0.000420");
expect(joined).toContain("error: boom");
});
});