Python rewrite of the agent harness on top of deepagents 0.6.1 + langchain 1.x, replacing the abandoned TS attempt in packages/. 388 unit/integration tests pass. Steps ----- 0. Scaffolding — uv workspace, ruff/mypy/pre-commit/alembic, src/tests/docs trees with docs/schemas/ seeded from my-deepagent-seed/. 1. Core — config (pydantic-settings with MYDEEPAGENT_ env prefix and TOML source), enums (Backend, Capability, RiskLevel, ApprovalDecisionAction, ApprovalState, RunState, RunPhaseState, SessionState, ErrorClass), errors (MyDeepAgentError + BudgetExhaustedError with PEP-3134 cause + context suppression), hash (canonical JSON + sha256). 2. Persona/Workflow/Binding — pydantic v2 schemas with tuple-based deep immutability (post-construction hash drift prevented), YAML loaders, deterministic auto-select (preferred_backends → version → name → hash), override resolution with ineligibility diagnostics, PersonaConsentStore with fcntl.flock + tmp+fsync+rename atomic write. 3. Artifact schema registry — Draft202012Validator, multi-root resolution, structured ValidationFinding output. 4. Persistence — 18 SQLAlchemy 2.0 async ORM models with FK CASCADE/RESTRICT, WAL + busy_timeout + foreign_keys PRAGMA, alembic baseline + ux_active_run_repo_base partial unique index, LangGraph SqliteSaver as context manager only (lifecycle safety). 5. DeepAgent session — build_agent wires Persona → create_deep_agent with LocalShellBackend / FilesystemBackend / StateBackend / CompositeBackend, ChatOpenAI(base_url=openrouter) for openrouter: model strings, and 4 middleware classes (cost / audit-tool / safety-shell / fallback-model). Critical workarounds -------------------- - deepagents 0.6.1 rejects FilesystemPermission together with backends that implement SandboxBackendProtocol (LocalShellBackend). SafetyShellMiddleware enforces destructive-command and secret-path policy at the tool layer instead, and build_agent strips the permissions kwarg when the persona's deepagents_backend is local_shell. - FilesystemOperation in deepagents is Literal['read', 'write'] only; _map_operations collapses our richer schema (read/write/edit/ls) safely. Real OpenRouter smoke --------------------- test_openrouter_deepagents_local_shell_smoke calls DeepSeek via deepagents + LocalShellBackend + SafetyShellMiddleware end-to-end. PASS, ~$0.000001 cost, input=9 / output=1 tokens with content "OK". Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
230 lines
7.1 KiB
Python
230 lines
7.1 KiB
Python
"""Unit tests for src/my_deepagent/monitoring/pricing.py."""
|
|
|
|
from __future__ import annotations
|
|
|
|
import httpx
|
|
import pytest
|
|
import respx
|
|
|
|
from my_deepagent.errors import MyDeepAgentError
|
|
from my_deepagent.monitoring.pricing import (
|
|
ModelPrice,
|
|
PricingCache,
|
|
_parse_pricing_payload,
|
|
fetch_openrouter_pricing,
|
|
)
|
|
|
|
# ---------------------------------------------------------------------------
|
|
# _parse_pricing_payload
|
|
# ---------------------------------------------------------------------------
|
|
|
|
|
|
def test_parse_valid_payload_returns_model_prices() -> None:
|
|
data = {
|
|
"data": [
|
|
{
|
|
"id": "deepseek/deepseek-chat",
|
|
"pricing": {"prompt": "0.000001", "completion": "0.000002"},
|
|
"context_length": 32768,
|
|
},
|
|
{
|
|
"id": "anthropic/claude-sonnet",
|
|
"pricing": {"prompt": "0.000003", "completion": "0.000015"},
|
|
"context_length": 200000,
|
|
},
|
|
]
|
|
}
|
|
result = _parse_pricing_payload(data)
|
|
assert len(result) == 2
|
|
assert result[0].model == "deepseek/deepseek-chat"
|
|
assert result[0].input_per_1k_usd == pytest.approx(0.001)
|
|
assert result[0].output_per_1k_usd == pytest.approx(0.002)
|
|
assert result[0].context_length == 32768
|
|
assert result[1].model == "anthropic/claude-sonnet"
|
|
|
|
|
|
def test_parse_empty_data_list_returns_empty() -> None:
|
|
result = _parse_pricing_payload({"data": []})
|
|
assert result == []
|
|
|
|
|
|
def test_parse_data_is_not_list_returns_empty() -> None:
|
|
# data is a dict instead of list — malformed response
|
|
result = _parse_pricing_payload({"data": {"id": "bad"}})
|
|
assert result == []
|
|
|
|
|
|
def test_parse_missing_data_key_returns_empty() -> None:
|
|
result = _parse_pricing_payload({})
|
|
assert result == []
|
|
|
|
|
|
def test_parse_skips_entries_without_id() -> None:
|
|
data = {
|
|
"data": [
|
|
{"pricing": {"prompt": "0.000001", "completion": "0.000002"}, "context_length": 1000},
|
|
]
|
|
}
|
|
result = _parse_pricing_payload(data)
|
|
assert result == []
|
|
|
|
|
|
def test_parse_skips_entries_with_invalid_pricing_values() -> None:
|
|
data = {
|
|
"data": [
|
|
{
|
|
"id": "model/x",
|
|
"pricing": {"prompt": "not-a-number", "completion": "also-bad"},
|
|
"context_length": 1000,
|
|
}
|
|
]
|
|
}
|
|
result = _parse_pricing_payload(data)
|
|
assert result == []
|
|
|
|
|
|
def test_parse_handles_null_pricing_gracefully() -> None:
|
|
data = {
|
|
"data": [
|
|
{"id": "model/y", "pricing": None, "context_length": 0},
|
|
]
|
|
}
|
|
result = _parse_pricing_payload(data)
|
|
# pricing=None -> {} -> prompt/completion default to "0"
|
|
assert len(result) == 1
|
|
assert result[0].input_per_1k_usd == 0.0
|
|
assert result[0].output_per_1k_usd == 0.0
|
|
|
|
|
|
def test_parse_handles_missing_context_length() -> None:
|
|
data = {
|
|
"data": [
|
|
{"id": "model/z", "pricing": {"prompt": "0.000001", "completion": "0.000002"}},
|
|
]
|
|
}
|
|
result = _parse_pricing_payload(data)
|
|
assert len(result) == 1
|
|
assert result[0].context_length == 0
|
|
|
|
|
|
def test_parse_non_dict_entry_is_skipped() -> None:
|
|
data = {"data": ["not-a-dict", None]}
|
|
result = _parse_pricing_payload(data)
|
|
assert result == []
|
|
|
|
|
|
# ---------------------------------------------------------------------------
|
|
# PricingCache.compute_cost
|
|
# ---------------------------------------------------------------------------
|
|
|
|
|
|
def test_compute_cost_known_model() -> None:
|
|
cache = PricingCache()
|
|
cache.set(
|
|
[
|
|
ModelPrice(
|
|
model="deepseek/deepseek-chat",
|
|
input_per_1k_usd=0.001,
|
|
output_per_1k_usd=0.002,
|
|
context_length=32768,
|
|
)
|
|
]
|
|
)
|
|
cost = cache.compute_cost("deepseek/deepseek-chat", input_tokens=1000, output_tokens=500)
|
|
assert cost == pytest.approx(0.001 * 1.0 + 0.002 * 0.5)
|
|
|
|
|
|
def test_compute_cost_openrouter_prefix_stripped() -> None:
|
|
cache = PricingCache()
|
|
cache.set(
|
|
[
|
|
ModelPrice(
|
|
model="deepseek/deepseek-chat",
|
|
input_per_1k_usd=0.001,
|
|
output_per_1k_usd=0.002,
|
|
context_length=32768,
|
|
)
|
|
]
|
|
)
|
|
# Should strip "openrouter:" prefix when looking up
|
|
cost = cache.compute_cost(
|
|
"openrouter:deepseek/deepseek-chat", input_tokens=1000, output_tokens=0
|
|
)
|
|
assert cost == pytest.approx(0.001)
|
|
|
|
|
|
def test_compute_cost_unknown_model_returns_zero() -> None:
|
|
cache = PricingCache()
|
|
cost = cache.compute_cost("unknown/model", input_tokens=1000, output_tokens=1000)
|
|
assert cost == 0.0
|
|
|
|
|
|
def test_compute_cost_zero_tokens_returns_zero() -> None:
|
|
cache = PricingCache()
|
|
cache.set(
|
|
[ModelPrice(model="m/x", input_per_1k_usd=1.0, output_per_1k_usd=2.0, context_length=1000)]
|
|
)
|
|
assert cache.compute_cost("m/x", input_tokens=0, output_tokens=0) == 0.0
|
|
|
|
|
|
def test_pricing_cache_get_strips_openrouter_prefix() -> None:
|
|
cache = PricingCache()
|
|
cache.set(
|
|
[ModelPrice(model="a/b", input_per_1k_usd=0.5, output_per_1k_usd=1.0, context_length=0)]
|
|
)
|
|
assert cache.get("openrouter:a/b") is not None
|
|
assert cache.get("a/b") is not None
|
|
|
|
|
|
# ---------------------------------------------------------------------------
|
|
# fetch_openrouter_pricing (respx mock)
|
|
# ---------------------------------------------------------------------------
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_fetch_openrouter_pricing_success() -> None:
|
|
payload = {
|
|
"data": [
|
|
{
|
|
"id": "deepseek/deepseek-chat",
|
|
"pricing": {"prompt": "0.000001", "completion": "0.000002"},
|
|
"context_length": 64000,
|
|
}
|
|
]
|
|
}
|
|
with respx.mock:
|
|
respx.get("https://openrouter.ai/api/v1/models").mock(
|
|
return_value=httpx.Response(200, json=payload)
|
|
)
|
|
result = await fetch_openrouter_pricing(
|
|
api_key="sk-or-test", base_url="https://openrouter.ai/api/v1"
|
|
)
|
|
assert len(result) == 1
|
|
assert result[0].model == "deepseek/deepseek-chat"
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_fetch_openrouter_pricing_http_error_raises_recoverable() -> None:
|
|
with respx.mock:
|
|
respx.get("https://openrouter.ai/api/v1/models").mock(
|
|
return_value=httpx.Response(401, json={"error": "unauthorized"})
|
|
)
|
|
with pytest.raises(MyDeepAgentError) as exc_info:
|
|
await fetch_openrouter_pricing(
|
|
api_key="bad-key", base_url="https://openrouter.ai/api/v1"
|
|
)
|
|
assert exc_info.value.code == "network_blip"
|
|
|
|
|
|
@pytest.mark.asyncio
|
|
async def test_fetch_openrouter_pricing_connect_error_raises_recoverable() -> None:
|
|
with respx.mock:
|
|
respx.get("https://openrouter.ai/api/v1/models").mock(
|
|
side_effect=httpx.ConnectError("connection refused")
|
|
)
|
|
with pytest.raises(MyDeepAgentError) as exc_info:
|
|
await fetch_openrouter_pricing(
|
|
api_key="sk-or-test", base_url="https://openrouter.ai/api/v1"
|
|
)
|
|
assert exc_info.value.code == "network_blip"
|