meshai/work/tests/test_persistence.py
Matt Johnson bc94cf237a fix(persistence): derive SCHEMA_VERSION from the migrations directory
db.py hardcoded SCHEMA_VERSION = 26 while migrations/ had already reached
v29 (v27 dispatcher floor-drop counter, v28 mesh_observations, v29 IPAWS).
The migration runner globs the directory and applies every vN.sql it finds
regardless of the constant, so a fresh DB actually landed at 29 while the
constant claimed 26 -- a three-version drift that three tests were
correctly catching.

Derive it from the highest vN.sql present instead of bumping the literal,
so it cannot drift again the next time someone adds a migration. Falls back
to 0 if the directory is missing so import never fails; the migrations dir
sits alongside db.py and ships with the package (Dockerfile COPYs meshai/).

Adds a regression guard asserting the constant matches the highest
migration file, and updates three tests that hardcoded 26 as a literal.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
2026-07-16 18:26:07 +00:00

356 lines
12 KiB
Python

"""Tests for the meshai/persistence foundation (v1 schema).
Foundation-only: no adapter handlers wired yet. Tests cover migration
runner, schema shape, idempotency, WAL mode, env-var path resolution,
in-memory mode, and basic insert/query against representative tables.
"""
import os
import sqlite3
import tempfile
import time
from pathlib import Path
import pytest
from meshai.persistence import (
MESHAI_DB_PATH,
SCHEMA_VERSION,
close_thread_connection,
get_db,
init_db,
)
from meshai.persistence import db as persistence_db
# ---------- helpers --------------------------------------------------------
def _force_reinit(monkeypatch, db_path):
"""Point every get_db() at a fresh DB and reset the per-process init
cache + thread-local connection so tests don't bleed into one another."""
monkeypatch.setenv("MESHAI_DB_PATH", db_path)
# Wipe the initialised-set so migrations re-run against the new file.
persistence_db._initialised.clear()
close_thread_connection()
@pytest.fixture
def tmp_db(monkeypatch, tmp_path):
db = str(tmp_path / "meshai-test.sqlite")
_force_reinit(monkeypatch, db)
yield db
close_thread_connection()
persistence_db._initialised.discard(db)
# ---------- migration runner ----------------------------------------------
# Every table the v1 schema is contracted to produce. The migration runner
# is tested by checking that ALL of these exist after init.
_V1_TABLES = {
"schema_meta",
"event_log",
"traffic_events",
"fires",
"firms_pixels",
"quake_events",
"nws_alerts",
"gauge_readings",
"swpc_events",
"mesh_nodes",
"mesh_telemetry",
"mesh_positions",
"mesh_messages_in",
"mesh_broadcasts_out",
"mesh_health_events",
}
def test_migration_v1_creates_all_tables(tmp_db):
conn = init_db()
rows = conn.execute(
"SELECT name FROM sqlite_master WHERE type='table' "
"AND name NOT LIKE 'sqlite_%' AND name NOT LIKE 'idx_%'"
).fetchall()
names = {r["name"] for r in rows}
missing = _V1_TABLES - names
extra = names - _V1_TABLES
assert not missing, f"missing tables after v1: {sorted(missing)}"
# extras are fine (autoindex tables etc.); just make sure we didn't
# leave the contract unfulfilled.
assert len(names) >= len(_V1_TABLES)
def test_schema_version_recorded(tmp_db):
conn = init_db()
row = conn.execute("SELECT value FROM schema_meta WHERE key='version'").fetchone()
assert row is not None
assert int(row["value"]) == SCHEMA_VERSION
def test_schema_version_matches_highest_migration_file():
"""SCHEMA_VERSION is derived from migrations/ at import time (see
db._derive_schema_version); this test independently re-derives the
expected value straight off the filenames so a future regression
(e.g. someone re-hardcoding the constant) is caught even if the
derivation logic itself is what breaks."""
import re
versions = []
for p in persistence_db.MIGRATIONS_DIR.iterdir():
m = re.match(r"^v(\d+)", p.stem)
if p.suffix.lower() == ".sql" and m:
versions.append(int(m.group(1)))
assert versions, "no vN.sql migration files found"
assert SCHEMA_VERSION == max(versions)
def test_migration_idempotent_rerun(tmp_db):
init_db()
# Force a "second startup" by closing the connection and clearing the
# per-process initialised cache so the migration runner re-evaluates.
close_thread_connection()
persistence_db._initialised.discard(tmp_db)
conn = init_db()
# No exception means we didn't try to CREATE TABLE twice; verify the
# row count of schema_meta hasn't drifted.
rows = conn.execute("SELECT COUNT(*) AS n FROM schema_meta").fetchone()
assert rows["n"] >= 1
# ---------- WAL mode ------------------------------------------------------
def test_wal_mode_enabled_on_file_db(tmp_db):
conn = init_db()
mode = conn.execute("PRAGMA journal_mode").fetchone()[0]
assert mode.lower() == "wal", f"expected WAL, got {mode!r}"
# ---------- MESHAI_DB_PATH resolution -------------------------------------
def test_meshai_db_path_env_var_respected(monkeypatch, tmp_path):
custom = str(tmp_path / "custom-path.sqlite")
monkeypatch.setenv("MESHAI_DB_PATH", custom)
assert MESHAI_DB_PATH() == custom
def test_meshai_db_path_falls_back_to_default(monkeypatch):
monkeypatch.delenv("MESHAI_DB_PATH", raising=False)
assert MESHAI_DB_PATH() == persistence_db.DEFAULT_DB_PATH
# ---------- in-memory mode for tests --------------------------------------
def test_in_memory_db_works(monkeypatch):
# Special-cased :memory: routes via uri=shared-cache so init runs cleanly.
monkeypatch.setenv("MESHAI_DB_PATH", ":memory:")
persistence_db._initialised.clear()
close_thread_connection()
conn = init_db()
# All v1 tables present even in memory.
rows = conn.execute(
"SELECT name FROM sqlite_master WHERE type='table' "
"AND name NOT LIKE 'sqlite_%'"
).fetchall()
names = {r["name"] for r in rows}
assert _V1_TABLES.issubset(names)
close_thread_connection()
# ---------- basic insert/query: fires -------------------------------------
def test_fires_insert_and_query(tmp_db):
conn = init_db()
now = int(time.time())
irwin = "{E7FCBC00-2D0A-49D6-889F-550D4EDCBFD6}"
conn.execute(
"INSERT INTO fires(irwin_id, incident_name, incident_type, "
"current_acres, current_contained_pct, status, lat, lon, "
"county, state, landclass, declared_at, last_event_at) "
"VALUES (?,?,?,?,?,?,?,?,?,?,?,?,?)",
(irwin, "IA 1", "wildfire", None, None, "active",
43.5213, -115.1665, "Elmore", "ID", "Sawtooth National Forest",
now - 3600, now),
)
row = conn.execute(
"SELECT incident_name, county, state, landclass, last_event_at "
"FROM fires WHERE irwin_id = ?", (irwin,)).fetchone()
assert row is not None
assert row["incident_name"] == "IA 1"
assert row["county"] == "Elmore"
assert row["state"] == "ID"
assert row["landclass"] == "Sawtooth National Forest"
assert row["last_event_at"] == now
def test_fires_last_broadcast_change_detection_columns(tmp_db):
"""The 8h-rate-limit change-detection logic uses last_broadcast_acres
and last_broadcast_contained -- verify they accept NULL and updates."""
conn = init_db()
now = int(time.time())
irwin = "{ABC}"
conn.execute(
"INSERT INTO fires(irwin_id, incident_name, last_event_at, "
"last_broadcast_at, last_broadcast_acres, last_broadcast_contained) "
"VALUES (?,?,?,?,?,?)",
(irwin, "X", now, now, 1847.0, 23),
)
row = conn.execute(
"SELECT last_broadcast_acres, last_broadcast_contained "
"FROM fires WHERE irwin_id = ?", (irwin,)).fetchone()
assert row["last_broadcast_acres"] == 1847.0
assert row["last_broadcast_contained"] == 23
# ---------- basic insert/query: mesh_nodes -------------------------------
def test_mesh_nodes_insert_and_query(tmp_db):
conn = init_db()
now = int(time.time())
conn.execute(
"INSERT INTO mesh_nodes(node_id, long_name, short_name, hw_model, "
"last_lat, last_lon, last_battery_pct, first_seen_at, last_seen_at) "
"VALUES (?,?,?,?,?,?,?,?,?)",
("!85098cea", "AIDA Northgate", "AIDA", "TBEAM",
43.6535, -116.2674, 88, now - 86400, now),
)
row = conn.execute(
"SELECT long_name, short_name, last_battery_pct, is_stale "
"FROM mesh_nodes WHERE node_id = ?", ("!85098cea",)).fetchone()
assert row is not None
assert row["long_name"] == "AIDA Northgate"
assert row["last_battery_pct"] == 88
assert row["is_stale"] == 0 # default
# ---------- traffic_events composite PK ----------------------------------
def test_traffic_events_composite_pk_uniqueness(tmp_db):
conn = init_db()
now = int(time.time())
conn.execute(
"INSERT INTO traffic_events(source, external_id, road, county, state, "
"first_seen_at, last_seen_at) VALUES (?,?,?,?,?,?,?)",
("state_511_atis", "ID:Construction:33868", "I-86", "Bannock", "ID", now, now),
)
# Same source + external_id should violate the PK constraint.
with pytest.raises(sqlite3.IntegrityError):
conn.execute(
"INSERT INTO traffic_events(source, external_id, road, county, state, "
"first_seen_at, last_seen_at) VALUES (?,?,?,?,?,?,?)",
("state_511_atis", "ID:Construction:33868", "I-86", "Bannock", "ID", now, now),
)
# Different source with same external_id is fine.
conn.execute(
"INSERT INTO traffic_events(source, external_id, road, county, state, "
"first_seen_at, last_seen_at) VALUES (?,?,?,?,?,?,?)",
("wzdx", "ID:Construction:33868", "I-86", "Bannock", "ID", now, now),
)
rows = conn.execute("SELECT COUNT(*) AS n FROM traffic_events").fetchone()
assert rows["n"] == 2
# ---------- event_log basic ------------------------------------------------
def test_event_log_insert_and_query(tmp_db):
conn = init_db()
now = int(time.time())
conn.execute(
"INSERT INTO event_log(received_at, source, category, severity_word, "
"event_id_external, nats_subject, handled, table_name, table_pk) "
"VALUES (?,?,?,?,?,?,?,?,?)",
(now, "wfigs_incidents", "fire.incident.wildfire", "routine",
"{E7FCBC00}", "central.fire.incident.id.elmore", 1, "fires", "{E7FCBC00}"),
)
row = conn.execute(
"SELECT source, table_name, handled FROM event_log "
"WHERE event_id_external = ?", ("{E7FCBC00}",)).fetchone()
assert row["source"] == "wfigs_incidents"
assert row["table_name"] == "fires"
assert row["handled"] == 1
# ---------- firms_pixels nullable FK to fires ----------------------------
def test_firms_pixels_fk_to_fires_nullable(tmp_db):
conn = init_db()
now = int(time.time())
# Unattached pixel (no irwin_id yet -- v0.6 fire-tracker will attach later).
conn.execute(
"INSERT INTO firms_pixels(lat, lon, acq_time, frp, confidence, satellite) "
"VALUES (?,?,?,?,?,?)",
(42.197, -113.710, now, 135.93, "high", "N"),
)
row = conn.execute(
"SELECT irwin_id, frp, confidence FROM firms_pixels "
"WHERE acq_time = ?", (now,)).fetchone()
assert row is not None
assert row["irwin_id"] is None # nullable
assert row["frp"] == 135.93
# Insert a fire, then attach a new pixel via the FK.
irwin = "{LINK-IRWIN}"
conn.execute(
"INSERT INTO fires(irwin_id, incident_name, last_event_at) "
"VALUES (?,?,?)", (irwin, "Linked Fire", now),
)
conn.execute(
"INSERT INTO firms_pixels(irwin_id, lat, lon, acq_time, frp) "
"VALUES (?,?,?,?,?)", (irwin, 42.197, -113.710, now + 1, 50.0),
)
rows = conn.execute(
"SELECT COUNT(*) AS n FROM firms_pixels WHERE irwin_id = ?",
(irwin,)).fetchone()
assert rows["n"] == 1
# ---------- indexes present (spot-check) ---------------------------------
def test_v1_indexes_created(tmp_db):
conn = init_db()
rows = conn.execute(
"SELECT name FROM sqlite_master WHERE type='index' "
"AND name LIKE 'idx_%'"
).fetchall()
names = {r["name"] for r in rows}
must_have = {
"idx_event_log_received",
"idx_traffic_last_seen",
"idx_fires_last_event",
"idx_firms_pixels_acq_time",
"idx_mesh_tel_node_time",
"idx_mesh_pos_node_time",
"idx_mesh_nodes_last_seen",
"idx_gauge_site_time",
}
missing = must_have - names
assert not missing, f"missing indexes: {sorted(missing)}"
# ---------- gauge_readings autoincrement -------------------------------
def test_gauge_readings_autoincrement_pk(tmp_db):
conn = init_db()
now = int(time.time())
for value in (3490.0, 3520.0, 3505.0):
conn.execute(
"INSERT INTO gauge_readings(site_id, reading_value, reading_unit, "
"reading_time) VALUES (?,?,?,?)",
("USGS-13038000", value, "ft^3/s", now),
)
rows = conn.execute(
"SELECT COUNT(*) AS n FROM gauge_readings WHERE site_id = ?",
("USGS-13038000",)).fetchone()
assert rows["n"] == 3