bfeafa2d68
Ports #119027::finalize_continuation_partial: when a stream drop entered the continuation path and the next request exhausted retries before a token, the _length_continuation_fragment/_nudge rows were persisted as-is, so resume replayed a dangling synthetic user nudge. Collapse them into one assistant row before persistence and feed the collapsed text to the #119081 partial-retention path (the fragment rows are gone by then). Co-authored-by: fangliquan <fangliquan@qq.com>
141 lines
5.4 KiB
Python
141 lines
5.4 KiB
Python
"""#119001: a 429 / network error that exhausts retries AFTER partial output was
|
|
delivered must retain that output instead of ending the turn error-only.
|
|
|
|
Repro: attempt 1 streams text then dies mid-stream (partial-stub path appends a
|
|
``_length_continuation_fragment`` + nudge to ``messages``); the continuation
|
|
attempt 429s before the stream starts and every retry re-hits the limit.
|
|
``build_api_request`` resets ``_current_streamed_assistant_text`` per attempt,
|
|
so the only record of the delivered text is the fragment rows — the terminal
|
|
builders below must recover it, flag ``partial``, and keep ``final_response``
|
|
distinct from ``error`` (the gateway/desktop retention contract).
|
|
"""
|
|
from __future__ import annotations
|
|
|
|
from agent.error_classifier import classify_api_error
|
|
from agent.turn_recovery import (
|
|
max_retries_exhausted_result,
|
|
nonretryable_client_error_result,
|
|
)
|
|
|
|
|
|
class _Agent:
|
|
log_prefix = ""
|
|
verbose = False
|
|
verbose_logging = False
|
|
provider = "openrouter"
|
|
model = "m"
|
|
base_url = "https://openrouter.ai/api/v1"
|
|
_current_streamed_assistant_text = ""
|
|
|
|
def _summarize_api_error(self, error):
|
|
return str(error)
|
|
|
|
def _has_pending_fallback(self):
|
|
return False
|
|
|
|
def _try_activate_fallback(self, **_kw):
|
|
return False
|
|
|
|
def __getattr__(self, name):
|
|
return lambda *args, **kwargs: None
|
|
|
|
|
|
class _Http(Exception):
|
|
def __init__(self, status_code, message):
|
|
super().__init__(message)
|
|
self.status_code = status_code
|
|
|
|
|
|
PARTIAL = "Here is the first half of the report, already shown to the user."
|
|
|
|
|
|
def _messages_with_fragment():
|
|
return [
|
|
{"role": "user", "content": "write a long report"},
|
|
{
|
|
"role": "assistant",
|
|
"content": PARTIAL,
|
|
"_length_continuation_fragment": True,
|
|
},
|
|
{
|
|
"role": "user",
|
|
"content": "continue where you left off",
|
|
"_length_continuation_nudge": True,
|
|
},
|
|
]
|
|
|
|
|
|
def _exhausted(messages, agent=None):
|
|
error = _Http(429, "HTTP 429: RequestBurstTooFast — slow down traffic growth")
|
|
classified = classify_api_error(error, provider="openrouter", model="m")
|
|
assert classified.reason.value == "rate_limit"
|
|
return max_retries_exhausted_result(
|
|
agent or _Agent(), error, classified, max_retries=3, is_rate_limited=True,
|
|
error_msg=str(error).lower(), api_kwargs=None, api_messages=[], messages=messages,
|
|
conversation_history=None, api_call_count=3, approx_tokens=10, provider="openrouter",
|
|
base_url="https://openrouter.ai/api/v1", model="m",
|
|
)
|
|
|
|
|
|
def test_exhausted_429_keeps_delivered_partial():
|
|
result = _exhausted(_messages_with_fragment())
|
|
assert result.get("partial") is True
|
|
assert PARTIAL in result["final_response"]
|
|
# Gateway retention contract: final must differ from the error string.
|
|
assert result["final_response"].strip() != str(result["error"]).strip()
|
|
assert result["failure_reason"] == "rate_limit"
|
|
assert result["failed"] is True
|
|
|
|
|
|
def test_exhausted_429_without_partial_is_unchanged():
|
|
result = _exhausted([])
|
|
assert "partial" not in result
|
|
assert "/retry" in result["final_response"]
|
|
assert result["failure_reason"] == "rate_limit"
|
|
|
|
|
|
def test_exhausted_429_live_accumulator_counts_as_delivered():
|
|
agent = _Agent()
|
|
agent._current_streamed_assistant_text = "fresh streamed text"
|
|
result = _exhausted([], agent=agent)
|
|
assert result.get("partial") is True
|
|
assert "fresh streamed text" in result["final_response"]
|
|
|
|
|
|
def test_whitespace_only_fragment_is_not_partial():
|
|
messages = _messages_with_fragment()
|
|
messages[1] = dict(messages[1], content=" ")
|
|
result = _exhausted(messages)
|
|
assert "partial" not in result
|
|
|
|
|
|
def test_nonretryable_terminal_keeps_delivered_partial():
|
|
error = _Http(400, "HTTP 400: Bad request")
|
|
classified = classify_api_error(error, provider="openrouter", model="m")
|
|
result = nonretryable_client_error_result(
|
|
_Agent(), error, classified, status_code=400, api_kwargs=None, api_messages=[],
|
|
messages=_messages_with_fragment(), conversation_history=None, api_call_count=1,
|
|
approx_tokens=10, provider="openrouter",
|
|
base_url="https://openrouter.ai/api/v1", model="m",
|
|
)
|
|
assert result.get("partial") is True
|
|
assert PARTIAL in result["final_response"]
|
|
assert result["final_response"].strip() != str(result["error"]).strip()
|
|
|
|
|
|
def test_exhausted_429_collapses_continuation_trail_into_one_assistant_row():
|
|
messages = _messages_with_fragment()
|
|
error = _Http(429, "HTTP 429: RequestBurstTooFast — slow down traffic growth")
|
|
classified = classify_api_error(error, provider="openrouter", model="m")
|
|
result = max_retries_exhausted_result(
|
|
_Agent(), error, classified, max_retries=3, is_rate_limited=True,
|
|
error_msg=str(error).lower(), api_kwargs=None, api_messages=[], messages=messages,
|
|
conversation_history=None, api_call_count=3, approx_tokens=10, provider="openrouter",
|
|
base_url="https://openrouter.ai/api/v1", model="m", current_turn_user_idx=0,
|
|
)
|
|
# No dangling synthetic nudge: the turn persists as user -> one assistant row.
|
|
assert [m["role"] for m in messages] == ["user", "assistant"]
|
|
assert messages[1]["content"] == PARTIAL
|
|
assert not any(m.get("_length_continuation_nudge") for m in messages)
|
|
assert result.get("partial") is True and PARTIAL in result["final_response"]
|