mirror of
https://github.com/bytedance/deer-flow.git
synced 2026-07-29 01:15:59 +00:00
* feat(subagents): persist and display subagent step history (#3779) Capture both assistant turns and tool outputs during subagent execution, stream them in task_running events, and persist them as subagent.* run events so the subtask card's step timeline survives a reload. Backend: - step_events.py: pure layer (capture_step_message, build_subagent_step, subagent_run_event) shared by streaming and persistence - executor.py: capture ToolMessage outputs, not just AIMessage turns - worker.py: persist task_* custom events to RunEventStore (category "subagent" keeps them out of the thread feed; list_events backfills) Frontend: - core/tasks/steps.ts + api.ts: SubtaskStep model, messageToStep, eventsToSteps, mergeSteps, fetchSubtaskSteps - subtask card accumulates live steps and backfills on expand - carry run_id onto history content messages for the events endpoint * fix(subagents): show AI turns in subtask card + paginate step backfill (#3779) Two follow-ups to the subagent step-history feature: Problem 1 — reload backfill could silently truncate the step timeline because list_events capped at 500 events (seq-ASC) across the whole run. Add task_id filtering + an after_seq forward cursor to list_events (all three stores + abstract base + the /events route), and make fetchSubtaskSteps page through one task's subagent.step events until a short page. No schema migration: the DB filter rides the existing run-scoped index via event_metadata["task_id"]. Problem 2 — the card only rendered tool steps, so persisted AI turns were never shown. Replace toolStepsForDisplay with stepsForDisplay: interleave AI reasoning turns (with text) and tool steps by message_index, drop blank-text AI turns, and drop the trailing final-answer AI turn when completed (already shown as result). Card renders AI steps as muted clamped markdown with a sparkles icon. Tests: store task_id/after_seq filtering + pagination across memory/db/jsonl, the /events route forwarding, stepsForDisplay rules, and fetchSubtaskSteps pagination. Docs updated in both AGENTS.md. * make format * fix(subagents): capture full multi-tool step tail, batch step persistence, cap tool-call args (#3779) Address PR review findings on the subagent step-history feature: 1. executor.py streamed on stream_mode="values" and captured only messages[-1] per chunk, so a multi-tool-call turn (ToolNode appends one ToolMessage per call in a single super-step) lost all but the last tool output in both the live task_running stream and the persisted history. Replace with capture_new_step_messages, which walks the newly-appended tail (and still re-checks the trailing message on no-growth chunks so id-less in-place replacements survive). 2. worker.py persisted each step with the store's low-frequency put() (a per-thread advisory lock per call); a deep subagent (max_turns=150) emits hundreds of steps on the hot stream loop. Replace with _SubagentEventBuffer, which batches via put_batch (flush on terminal subagent.end, at FLUSH_THRESHOLD, and in the worker finally). 3. build_subagent_step capped only text; tool_calls[].args were copied verbatim, so a large write_file/bash payload produced an unbounded subagent.step row. Cap each call's serialized args at SUBAGENT_STEP_MAX_CHARS, flagged args_truncated. Tests updated/added for all three; AGENTS.md refreshed. * fix(subagents): merge backfill into latest subtask state; reuse message_content_to_text (#3779) Address the remaining two PR review findings: 4. subtask-card's fetchSubtaskSteps().then(updateSubtask) closed over a stale tasks snapshot: a late-resolving backfill wrote setTasks({...stale}), clobbering SSE steps/status and sibling subtasks that arrived during the fetch. useUpdateSubtask now reads/writes through a tasksRef mirroring the latest state (ref-to-latest), and the pure per-subtask transition is extracted to core/tasks/subtask-update.ts::computeNextSubtask (unit-tested). 5. step_events._content_to_text duplicated deerflow.utils.messages. message_content_to_text; call the shared helper instead (guarding None content with 'or ""' so a tool-call-only turn still renders as ""). Tests added for computeNextSubtask and the None-content case; AGENTS.md docs updated.
109 lines
3.3 KiB
TypeScript
109 lines
3.3 KiB
TypeScript
import { describe, expect, it } from "@rstest/core";
|
|
|
|
import type { SubtaskStep } from "@/core/tasks/steps";
|
|
import {
|
|
computeNextSubtask,
|
|
isTerminalSubtaskStatus,
|
|
} from "@/core/tasks/subtask-update";
|
|
import type { Subtask } from "@/core/tasks/types";
|
|
|
|
function baseTask(overrides: Partial<Subtask> = {}): Subtask {
|
|
return {
|
|
id: "t1",
|
|
status: "in_progress",
|
|
subagent_type: "general-purpose",
|
|
description: "research",
|
|
prompt: "do it",
|
|
...overrides,
|
|
};
|
|
}
|
|
|
|
function step(message_index: number): SubtaskStep {
|
|
return {
|
|
kind: "tool",
|
|
message_index,
|
|
text: `step ${message_index}`,
|
|
truncated: false,
|
|
};
|
|
}
|
|
|
|
describe("computeNextSubtask", () => {
|
|
it("merges step deltas into the provided previous, preserving both", () => {
|
|
const previous = baseTask({ steps: [step(1), step(2)] });
|
|
|
|
const { next } = computeNextSubtask(previous, {
|
|
id: "t1",
|
|
steps: [step(3)],
|
|
});
|
|
|
|
// Regression for #3779 stale-closure race: next is derived from whatever
|
|
// `previous` is passed (the functional-update latest state), so concurrently
|
|
// arrived steps are kept, not clobbered by a backfill resolving late.
|
|
expect(next.steps?.map((s) => s.message_index)).toEqual([1, 2, 3]);
|
|
});
|
|
|
|
it("does not drop steps present only on the latest previous", () => {
|
|
// Simulates a backfill that fired when previous had 0 steps, but by the time
|
|
// it resolves the latest previous already has SSE steps 1..2. The backfill
|
|
// brings historical steps 1..3; the merge must retain all, not reset to [].
|
|
const latestPrevious = baseTask({ steps: [step(1), step(2)] });
|
|
|
|
const { next } = computeNextSubtask(latestPrevious, {
|
|
id: "t1",
|
|
steps: [step(1), step(2), step(3)],
|
|
});
|
|
|
|
expect(next.steps?.map((s) => s.message_index)).toEqual([1, 2, 3]);
|
|
});
|
|
|
|
it("keeps a terminal status stable against a late in_progress write", () => {
|
|
const previous = baseTask({ status: "completed" });
|
|
|
|
const { next, becameTerminal } = computeNextSubtask(previous, {
|
|
id: "t1",
|
|
status: "in_progress",
|
|
});
|
|
|
|
expect(next.status).toBe("completed");
|
|
expect(becameTerminal).toBe(false);
|
|
});
|
|
|
|
it("flags becameTerminal on the first transition to a terminal status", () => {
|
|
const previous = baseTask({ status: "in_progress" });
|
|
|
|
const { next, becameTerminal } = computeNextSubtask(previous, {
|
|
id: "t1",
|
|
status: "completed",
|
|
result: "done",
|
|
});
|
|
|
|
expect(next.status).toBe("completed");
|
|
expect(next.result).toBe("done");
|
|
expect(becameTerminal).toBe(true);
|
|
});
|
|
|
|
it("handles an undefined previous (first write for a task)", () => {
|
|
const { next, becameTerminal } = computeNextSubtask(undefined, {
|
|
id: "t1",
|
|
status: "in_progress",
|
|
subagent_type: "bash",
|
|
description: "run",
|
|
prompt: "p",
|
|
steps: [step(1)],
|
|
});
|
|
|
|
expect(next.id).toBe("t1");
|
|
expect(next.steps?.map((s) => s.message_index)).toEqual([1]);
|
|
expect(becameTerminal).toBe(false);
|
|
});
|
|
});
|
|
|
|
describe("isTerminalSubtaskStatus", () => {
|
|
it("recognizes terminal statuses only", () => {
|
|
expect(isTerminalSubtaskStatus("completed")).toBe(true);
|
|
expect(isTerminalSubtaskStatus("failed")).toBe(true);
|
|
expect(isTerminalSubtaskStatus("in_progress")).toBe(false);
|
|
expect(isTerminalSubtaskStatus(undefined)).toBe(false);
|
|
});
|
|
});
|