dongarandClaude Opus 5.5
154a0c3fa8
What isn't the dev agent's moves to nisse-core, unchanged in behaviour, with its tests: canonical, render, render_anthropic, llm, catalog, probe, context, views, state, reconcile, worker, cancel, dump, transport, matrix, execution, testing, vault, memory, oplog, envfile and gitea (whole). - nisse_core.state.StateStore holds the tables every agent has; an agent's subclass adds its tables (version, migrations), roles and views. nisse_dev.state.StateStore adds the build VM blocks (v5). - views.project takes what each role is shown, rather than a fixed table. - Vault takes the agent's name (agent=), for provenance. - A test that nisse_core never imports an agent. - The loop, config, git, delegation, expert, tools and evals stay in nisse-dev until nisse-ops has a loop to share. Docs: the Development Plan marks M7.1 done and notes that Phase 6 waits on the agent host and Terraform; Architecture Part L names the package. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01J5Utyfzw7gF7UhjRDEtjTk
121 lines
3.9 KiB
Python
121 lines
3.9 KiB
Python
import sqlite3
|
|
from pathlib import Path
|
|
|
|
import pytest
|
|
|
|
from nisse_core.canonical import (
|
|
AssistantText,
|
|
Compaction,
|
|
DelegationCompleted,
|
|
DelegationStarted,
|
|
GateOpened,
|
|
PlanSubmitted,
|
|
ToolCall,
|
|
ToolResult,
|
|
UserMessage,
|
|
from_payload,
|
|
mint_call_id,
|
|
)
|
|
from nisse_core.state import StateSchemaError, StateStore
|
|
|
|
ALL_EVENTS = [
|
|
UserMessage(text="use demo", source="matrix:$abc"),
|
|
AssistantText(text="starting", model="local/qwen"),
|
|
ToolCall(call_id="call-1", tool="run_tests", arguments={"path": "tests/"}),
|
|
ToolResult(call_id="call-1", result="3 passed", success=True),
|
|
PlanSubmitted(plan="the plan", model="claude-opus-5"),
|
|
GateOpened(gate="arch", reason="plan submitted (rev-1)"),
|
|
DelegationStarted(
|
|
brief="fix the flaky test",
|
|
role="helper",
|
|
model="claude-opus-5",
|
|
caps={"turns": 20, "tokens": 100000},
|
|
),
|
|
DelegationCompleted(
|
|
handback="test fixed; root cause was a shared tmpdir",
|
|
artifacts=["nisse-dev/tests/test_x.py"],
|
|
tokens=51234,
|
|
duration_s=61.5,
|
|
exit_status="success",
|
|
),
|
|
Compaction(summary="earlier work summarized", model="local/qwen", replaces=[1, 4]),
|
|
]
|
|
|
|
|
|
def test_all_event_types_round_trip(tmp_path: Path) -> None:
|
|
with StateStore(tmp_path / "state.db") as store:
|
|
for event in ALL_EVENTS:
|
|
store.transcript("t", "driver").append(event)
|
|
with StateStore(tmp_path / "state.db") as store:
|
|
stored = store.transcript("t", "driver").read()
|
|
assert [s.event for s in stored] == ALL_EVENTS
|
|
assert [s.id for s in stored] == sorted(s.id for s in stored)
|
|
assert {(s.work_item, s.actor) for s in stored} == {("t", "driver")}
|
|
|
|
|
|
def test_read_events_after_id(tmp_path: Path) -> None:
|
|
with StateStore(tmp_path / "state.db") as store:
|
|
log = store.transcript("t", "operator")
|
|
first = log.append(UserMessage(text="one", source="m:1"))
|
|
log.append(UserMessage(text="two", source="m:2"))
|
|
tail = log.read(after=first)
|
|
assert [s.event.text for s in tail] == ["two"]
|
|
|
|
|
|
def test_mint_call_id_is_unique_and_prefixed() -> None:
|
|
ids = {mint_call_id() for _ in range(100)}
|
|
assert len(ids) == 100
|
|
assert all(i.startswith("call-") for i in ids)
|
|
|
|
|
|
def test_unknown_event_type_rejected() -> None:
|
|
with pytest.raises(ValueError, match="unknown canonical event type"):
|
|
from_payload("mystery", {})
|
|
|
|
|
|
def test_unknown_fields_rejected() -> None:
|
|
with pytest.raises(ValueError, match="unknown fields"):
|
|
from_payload("user_message", {"text": "hi", "source": "m", "extra": 1})
|
|
|
|
|
|
# The v1 schema as it shipped (PR #2), frozen here to prove a newer image
|
|
# opens and upgrades a database written by the older one (L8).
|
|
V1_SCHEMA = [
|
|
"CREATE TABLE meta (key TEXT PRIMARY KEY, value TEXT NOT NULL)",
|
|
"""
|
|
CREATE TABLE agent_state (
|
|
id INTEGER PRIMARY KEY CHECK (id = 1),
|
|
state TEXT NOT NULL,
|
|
project TEXT,
|
|
data TEXT NOT NULL DEFAULT '{}',
|
|
updated_at TEXT NOT NULL
|
|
)
|
|
""",
|
|
"""
|
|
CREATE TABLE intent_log (
|
|
id INTEGER PRIMARY KEY AUTOINCREMENT,
|
|
kind TEXT NOT NULL,
|
|
payload TEXT NOT NULL,
|
|
status TEXT NOT NULL DEFAULT 'pending',
|
|
created_at TEXT NOT NULL,
|
|
resolved_at TEXT
|
|
)
|
|
""",
|
|
]
|
|
|
|
|
|
def test_a_store_from_before_work_items_is_refused_clearly(tmp_path: Path) -> None:
|
|
# No production data predates work items, so there is no migration from
|
|
# these versions (Development Plan M1.1): the store must say so rather
|
|
# than open half-understood.
|
|
db = tmp_path / "state.db"
|
|
conn = sqlite3.connect(db)
|
|
for statement in V1_SCHEMA:
|
|
conn.execute(statement)
|
|
conn.execute("INSERT INTO meta (key, value) VALUES ('schema_version', '3')")
|
|
conn.commit()
|
|
conn.close()
|
|
|
|
with pytest.raises(StateSchemaError, match="predates work items"):
|
|
StateStore(db)
|