feat(#1): shared SSE resume orchestration; wire cli --send

v1 coverage-audit slice b1. The audit found reconnect_turn had no
caller — every presenter dropped the stream on disconnect instead of
resuming, leaving the "reference SSE-resume implementation" (design-
brief §3/§8d) unreachable. Add stream_turn_resilient as the single
shared resume surface (design-brief §8b "share the consumer, branch
the presenter") and route cli --send through it.

- stream_turn_resilient wraps stream_turn + reconnect_turn: on
  SseConnectionDropped (mid-stream drop or clean EOF before terminal),
  resume from the last-seen sse_id via reconnect_turn (Last-Event-ID),
  up to max_reconnects (default 5). last_seen persists across attempts.
- Non-drop reconnect failures (412/410/400/TurnIdFlip/SseConnectFailed)
  propagate unchanged, per contract #1's "surface, not recover".
- cli.py: --send consumer now drives stream_turn_resilient (transparent
  reconnect). tui/web still consume bare stream_turn (follow-up).
- Contract #1 amended (FN stream_turn_resilient) + validated; 8 TDD
  cases (happy, resume-after-1/2-drops, clean-EOF resume, unresumable
  zero-event, max-reconnects-exhausted, zero-budget, buffer-expired-
  propagates). Suite 518 green; ruff + mypy clean on touched code.
This commit is contained in:
vh
2026-06-30 15:42:33 -07:00
parent 1f289098ba
commit 0c7660791f
7 changed files with 255 additions and 4 deletions
+153
View File
@@ -19,6 +19,7 @@ from ratatoskr.sse_client import (
ResumeBufferExpired,
ResumeTurnFinished,
SseConnectFailed,
SseConnectionDropped,
SseId,
Text,
TurnIdFlip,
@@ -27,6 +28,7 @@ from ratatoskr.sse_client import (
cancel_turn,
reconnect_turn,
stream_turn,
stream_turn_resilient,
)
_DONE_42_6 = {
@@ -51,6 +53,36 @@ def _sse_chunk(sse_id: str, body: dict[str, object]) -> bytes:
return f"id: {sse_id}\ndata: {json.dumps(body)}\n\n".encode()
_EVENT_STREAM = {"content-type": "text/event-stream"}
class _DropStream(httpx.AsyncByteStream):
"""Yield the given chunks, then raise a mid-stream drop (RemoteProtocolError).
Mirrors the inline `_DropAfter` used by TestStreamTurn.test_connection_drop;
hoisted to module scope because the resilient-wrapper tests reuse it.
"""
def __init__(self, chunks: list[bytes]) -> None:
self._chunks = chunks
async def __aiter__(self): # type: ignore[no-untyped-def]
for c in self._chunks:
yield c
raise httpx.RemoteProtocolError("simulated mid-stream drop")
async def aclose(self) -> None:
return None
def _drop_response(chunks: list[bytes]) -> httpx.Response:
return httpx.Response(200, headers=_EVENT_STREAM, stream=_DropStream(chunks))
def _stream_response(content: bytes) -> httpx.Response:
return httpx.Response(200, headers=_EVENT_STREAM, content=content)
class TestParseSseId:
def test_happy_simple(self) -> None:
"""happy_simple [happy,tracer]: '42:3' -> SseId(turn_id=42, seq=3)."""
@@ -1108,3 +1140,124 @@ class TestAwaitingLlmFirstToken:
elapsed = [b.elapsed_ms_since_building_prompt for b in beats]
assert elapsed == sorted(elapsed) # monotonically increasing
assert all(b.turn_id == 42 for b in beats)
_URL = "https://w.example/sessions/s1/messages"
class TestStreamTurnResilient:
"""docs/contracts/issues/1.contract.md FN stream_turn_resilient (amendment 2026-06-30)."""
@respx.mock
async def test_happy_no_drop(self) -> None:
"""happy_no_drop [happy]: clean stream passes through; no reconnect issued."""
stream = _sse_chunk("42:1", {"type": "text", "content": "a"}) + _sse_chunk(
"42:2", _DONE_42_6
)
route = respx.post(_URL).mock(return_value=_stream_response(stream))
async with httpx.AsyncClient(base_url="https://w.example") as client:
events = [e async for e in stream_turn_resilient(client, "s1", "hi")]
assert [e.sse_id for e in events] == [SseId(42, 1), SseId(42, 2)]
assert isinstance(events[-1], Done)
assert route.call_count == 1 # POST-001: no reconnect on a clean stream
@respx.mock
async def test_resume_after_one_drop(self) -> None:
"""resume_after_one_drop [tracer]: a mid-stream drop resumes via reconnect; one stream."""
first = _drop_response([_sse_chunk("42:1", {"type": "text", "content": "a"})])
resume = _stream_response(
_sse_chunk("42:2", {"type": "text", "content": "b"})
+ _sse_chunk("42:3", _DONE_42_6)
)
route = respx.post(_URL).mock(side_effect=[first, resume])
async with httpx.AsyncClient(base_url="https://w.example") as client:
events = [e async for e in stream_turn_resilient(client, "s1", "hi")]
assert [e.sse_id for e in events] == [SseId(42, 1), SseId(42, 2), SseId(42, 3)]
assert isinstance(events[-1], Done)
assert route.call_count == 2
# POST-003: reconnect carries the last yielded pre-drop event's id.
assert route.calls[1].request.headers.get("Last-Event-ID") == "42:1"
# PRE/wire: first attempt does NOT carry a Last-Event-ID.
assert route.calls[0].request.headers.get("Last-Event-ID") is None
@respx.mock
async def test_resume_after_clean_eof(self) -> None:
"""resume_after_clean_eof: a clean EOF before terminal also triggers resume (INV-001)."""
first = _stream_response(_sse_chunk("42:1", {"type": "text", "content": "a"}))
resume = _stream_response(_sse_chunk("42:2", _DONE_42_6))
route = respx.post(_URL).mock(side_effect=[first, resume])
async with httpx.AsyncClient(base_url="https://w.example") as client:
events = [e async for e in stream_turn_resilient(client, "s1", "hi")]
assert [e.sse_id for e in events] == [SseId(42, 1), SseId(42, 2)]
assert isinstance(events[-1], Done)
assert route.calls[1].request.headers.get("Last-Event-ID") == "42:1"
@respx.mock
async def test_two_drops_then_done(self) -> None:
"""two_drops_then_done: two transient drops, third attempt completes; ids thread through."""
a1 = _drop_response([_sse_chunk("42:1", {"type": "text", "content": "a"})])
a2 = _drop_response([_sse_chunk("42:2", {"type": "text", "content": "b"})])
a3 = _stream_response(_sse_chunk("42:3", _DONE_42_6))
route = respx.post(_URL).mock(side_effect=[a1, a2, a3])
async with httpx.AsyncClient(base_url="https://w.example") as client:
events = [e async for e in stream_turn_resilient(client, "s1", "hi")]
assert [e.sse_id for e in events] == [SseId(42, 1), SseId(42, 2), SseId(42, 3)]
assert route.call_count == 3
assert route.calls[1].request.headers.get("Last-Event-ID") == "42:1"
assert route.calls[2].request.headers.get("Last-Event-ID") == "42:2"
@respx.mock
async def test_unresumable_zero_event_drop(self) -> None:
"""unresumable_zero_event_drop [adversarial]: drop before any event → propagate."""
route = respx.post(_URL).mock(side_effect=[_drop_response([])])
async with httpx.AsyncClient(base_url="https://w.example") as client:
with pytest.raises(SseConnectionDropped):
_ = [e async for e in stream_turn_resilient(client, "s1", "hi")]
assert route.call_count == 1 # no id to resume from → no reconnect
@respx.mock
async def test_max_reconnects_exhausted(self) -> None:
"""max_reconnects_exhausted [adversarial]: every attempt drops; budget caps reconnects."""
side = [
_drop_response([_sse_chunk(f"42:{n}", {"type": "text", "content": "x"})])
for n in (1, 2, 3)
]
route = respx.post(_URL).mock(side_effect=side)
async with httpx.AsyncClient(base_url="https://w.example") as client:
with pytest.raises(SseConnectionDropped):
_ = [
e
async for e in stream_turn_resilient(
client, "s1", "hi", max_reconnects=2
)
]
assert route.call_count == 3 # initial + 2 reconnects, then give up
@respx.mock
async def test_zero_budget_no_resume(self) -> None:
"""zero_budget_no_resume [adversarial]: max_reconnects=0 → first drop propagates."""
first = _drop_response([_sse_chunk("42:1", {"type": "text", "content": "a"})])
route = respx.post(_URL).mock(side_effect=[first])
async with httpx.AsyncClient(base_url="https://w.example") as client:
with pytest.raises(SseConnectionDropped):
_ = [
e
async for e in stream_turn_resilient(
client, "s1", "hi", max_reconnects=0
)
]
assert route.call_count == 1
@respx.mock
async def test_buffer_expired_propagates(self) -> None:
"""buffer_expired_propagates [error]: a 412 on reconnect surfaces, not retried."""
first = _drop_response([_sse_chunk("42:1", {"type": "text", "content": "a"})])
resume = httpx.Response(412, json={"turn_id": 42, "buffered_from_seq": 7})
route = respx.post(_URL).mock(side_effect=[first, resume])
async with httpx.AsyncClient(base_url="https://w.example") as client:
collected: list[object] = []
with pytest.raises(ResumeBufferExpired):
async for e in stream_turn_resilient(client, "s1", "hi"):
collected.append(e)
assert [e.sse_id for e in collected] == [SseId(42, 1)] # type: ignore[attr-defined]
assert route.call_count == 2