refactor: simplify mqtt pipeline state and remove event table
All checks were successful
Deploy to Production / deploy (push) Successful in 4m55s

This commit is contained in:
魏风
2026-03-19 22:08:16 +08:00
parent 8717b9c07b
commit e8846f4831
14 changed files with 269 additions and 118 deletions

View File

@@ -0,0 +1,108 @@
"""simplify mqtt parse tracking
Revision ID: b4c5d6e7f809
Revises: a1b2c3d4e5f6
Create Date: 2026-03-19 17:30:00.000000
"""
from alembic import op
import sqlalchemy as sa
# revision identifiers, used by Alembic.
revision = "b4c5d6e7f809"
down_revision = "a1b2c3d4e5f6"
branch_labels = None
depends_on = None
def upgrade() -> None:
op.add_column(
"terminal_message_raw",
sa.Column("parsed_at", sa.DateTime(timezone=True), nullable=True),
)
op.create_index(
op.f("ix_terminal_message_raw_parsed_at"),
"terminal_message_raw",
["parsed_at"],
unique=False,
)
# Preserve historical parse outcomes before removing state-machine columns.
op.execute(
sa.text(
"""
UPDATE terminal_message_raw
SET parsed_at = COALESCE(received_at, created_at)
WHERE parse_status <> 'pending'
"""
)
)
op.drop_index(
op.f("ix_terminal_message_raw_next_retry_at"),
table_name="terminal_message_raw",
)
op.drop_index(
op.f("ix_terminal_message_raw_parse_status"),
table_name="terminal_message_raw",
)
op.drop_column("terminal_message_raw", "next_retry_at")
op.drop_column("terminal_message_raw", "parse_attempts")
op.drop_column("terminal_message_raw", "parse_status")
def downgrade() -> None:
op.add_column(
"terminal_message_raw",
sa.Column(
"parse_status",
sa.String(length=32),
nullable=False,
server_default="pending",
),
)
op.add_column(
"terminal_message_raw",
sa.Column("parse_attempts", sa.Integer(), nullable=False, server_default="0"),
)
op.add_column(
"terminal_message_raw",
sa.Column("next_retry_at", sa.DateTime(timezone=True), nullable=True),
)
op.create_index(
op.f("ix_terminal_message_raw_parse_status"),
"terminal_message_raw",
["parse_status"],
unique=False,
)
op.create_index(
op.f("ix_terminal_message_raw_next_retry_at"),
"terminal_message_raw",
["next_retry_at"],
unique=False,
)
op.execute(
sa.text(
"""
UPDATE terminal_message_raw
SET parse_status = CASE
WHEN parsed_at IS NULL THEN 'pending'
WHEN parse_error IS NULL THEN 'succeeded'
ELSE 'failed'
END,
parse_attempts = CASE
WHEN parsed_at IS NULL THEN 0
ELSE 1
END
"""
)
)
op.drop_index(
op.f("ix_terminal_message_raw_parsed_at"),
table_name="terminal_message_raw",
)
op.drop_column("terminal_message_raw", "parsed_at")

View File

@@ -0,0 +1,60 @@
"""drop terminal observation event table
Revision ID: c6d7e8f90123
Revises: b4c5d6e7f809
Create Date: 2026-03-19 18:20:00.000000
"""
from alembic import op
import sqlalchemy as sa
# revision identifiers, used by Alembic.
revision = "c6d7e8f90123"
down_revision = "b4c5d6e7f809"
branch_labels = None
depends_on = None
def upgrade() -> None:
op.drop_table("terminal_observation_event")
def downgrade() -> None:
op.create_table(
"terminal_observation_event",
sa.Column("id", sa.Uuid(), nullable=False),
sa.Column("raw_id", sa.Uuid(), nullable=False),
sa.Column("device_id", sa.String(length=255), nullable=False),
sa.Column("event_time", sa.DateTime(timezone=True), nullable=False),
sa.Column("event_type", sa.String(length=64), nullable=False),
sa.Column("payload_json", sa.JSON(), nullable=False),
sa.Column("created_at", sa.DateTime(timezone=True), nullable=False),
sa.ForeignKeyConstraint(["raw_id"], ["terminal_message_raw.id"]),
sa.PrimaryKeyConstraint("id"),
)
op.create_index(
op.f("ix_terminal_observation_event_device_id"),
"terminal_observation_event",
["device_id"],
unique=False,
)
op.create_index(
op.f("ix_terminal_observation_event_event_time"),
"terminal_observation_event",
["event_time"],
unique=False,
)
op.create_index(
op.f("ix_terminal_observation_event_event_type"),
"terminal_observation_event",
["event_type"],
unique=False,
)
op.create_index(
op.f("ix_terminal_observation_event_raw_id"),
"terminal_observation_event",
["raw_id"],
unique=False,
)

View File

@@ -126,9 +126,7 @@ class TerminalMessageRaw(SQLModel, table=True):
payload_sha256: str = Field(default="", max_length=64, index=True)
qos: int = 0
retain: bool = False
parse_status: str = Field(default="pending", max_length=32, index=True)
parse_attempts: int = 0
next_retry_at: datetime | None = Field(
parsed_at: datetime | None = Field(
default=None,
sa_type=DateTime(timezone=True), # type: ignore
index=True,
@@ -140,27 +138,6 @@ class TerminalMessageRaw(SQLModel, table=True):
)
class TerminalObservationEvent(SQLModel, table=True):
__tablename__ = "terminal_observation_event"
id: uuid.UUID = Field(default_factory=uuid.uuid4, primary_key=True)
raw_id: uuid.UUID = Field(
foreign_key="terminal_message_raw.id", nullable=False, index=True
)
device_id: str = Field(max_length=255, index=True)
event_time: datetime = Field(
default_factory=get_datetime_utc,
sa_type=DateTime(timezone=True), # type: ignore
index=True,
)
event_type: str = Field(max_length=64, index=True)
payload_json: dict[str, Any] = Field(default_factory=dict, sa_type=JSON)
created_at: datetime = Field(
default_factory=get_datetime_utc,
sa_type=DateTime(timezone=True), # type: ignore
)
class TerminalDeviceState(SQLModel, table=True):
__tablename__ = "terminal_device_state"

View File

@@ -1,24 +1,32 @@
from __future__ import annotations
import logging
import re
from typing import Any
import paho.mqtt.client as mqtt # type: ignore[import-not-found,import-untyped]
from sqlmodel import Session
from app.core.config import settings
from app.core.db import engine
from app.mqtt.repository import insert_raw_message
from app.mqtt.topic import extract_device_id
logger = logging.getLogger(__name__)
TOPIC_PATTERN = re.compile(r"^terminal/(?P<device_id>[^/]+)/raw$")
def extract_device_id(topic: str) -> str:
match = TOPIC_PATTERN.match(topic)
if match is None:
raise ValueError("topic must match terminal/{device_id}/raw")
return match.group("device_id")
def _on_connect(
client: mqtt.Client,
client: Any,
_userdata: object,
_flags: dict[str, int],
reason_code: int,
_properties: mqtt.Properties | None = None,
_properties: object | None = None,
) -> None:
if reason_code != 0:
logger.error("MQTT connect failed: rc=%s", reason_code)
@@ -31,9 +39,9 @@ def _on_connect(
def _on_message(
_client: mqtt.Client,
_client: Any,
_userdata: object,
message: mqtt.MQTTMessage,
message: Any,
) -> None:
try:
device_id = extract_device_id(message.topic)
@@ -52,7 +60,14 @@ def _on_message(
)
def create_client() -> mqtt.Client:
def create_client() -> Any:
try:
import paho.mqtt.client as mqtt # type: ignore[import-not-found,import-untyped]
except ModuleNotFoundError as exc: # pragma: no cover - runtime guard
raise RuntimeError(
"paho-mqtt is required to run mqtt-ingestor. Install backend dependencies first."
) from exc
client = mqtt.Client(
callback_api_version=mqtt.CallbackAPIVersion.VERSION2,
client_id=settings.MQTT_CLIENT_ID,

View File

@@ -2,45 +2,33 @@ from __future__ import annotations
import logging
import time
from datetime import timedelta
from sqlmodel import Session
from app.core.config import settings
from app.core.db import engine
from app.models import TerminalMessageRaw, get_datetime_utc
from app.models import TerminalMessageRaw
from app.mqtt.parsers.mixed import parse_mixed_payload
from app.mqtt.repository import (
fetch_pending_rows_for_update,
mark_failed,
mark_processing,
mark_succeeded,
write_events_and_update_state,
update_device_state_from_events,
)
logger = logging.getLogger(__name__)
def _next_retry_delay_seconds(attempts: int) -> int:
# Bounded exponential backoff: 30s, 60s, 120s, ... capped to 15m.
return int(min(30 * (2 ** max(attempts - 1, 0)), 900))
def _process_one_row(*, session: Session, row: TerminalMessageRaw) -> None:
mark_processing(session=session, row=row)
session.commit()
try:
events = parse_mixed_payload(row.payload_bytes)
write_events_and_update_state(session=session, row=row, events=events)
update_device_state_from_events(session=session, row=row, events=events)
mark_succeeded(session=session, row=row)
except Exception as exc: # pragma: no cover - safety net path
delay = _next_retry_delay_seconds(row.parse_attempts)
mark_failed(
session=session,
row=row,
parse_error=str(exc),
next_retry_at=get_datetime_utc() + timedelta(seconds=delay),
)
session.commit()

View File

@@ -1,5 +1,4 @@
from app.mqtt.parsers.mixed import parse_mixed_payload
from app.mqtt.parsers.mixed import parse_mixed_payload, summarize_rtcm_types
from app.mqtt.parsers.nmea import parse_nmea_sentences
from app.mqtt.parsers.rtcm import summarize_rtcm_types
__all__ = ["parse_mixed_payload", "parse_nmea_sentences", "summarize_rtcm_types"]

View File

@@ -3,7 +3,27 @@ from __future__ import annotations
from typing import Any
from app.mqtt.parsers.nmea import parse_nmea_sentences
from app.mqtt.parsers.rtcm import summarize_rtcm_types
def summarize_rtcm_types(payload: bytes) -> dict[int, int]:
counts: dict[int, int] = {}
i = 0
while i + 6 <= len(payload):
if payload[i] != 0xD3:
i += 1
continue
length = ((payload[i + 1] & 0x03) << 8) | payload[i + 2]
frame_end = i + 3 + length + 3
if frame_end > len(payload):
break
message_type = ((payload[i + 3] << 4) | (payload[i + 4] >> 4)) & 0x0FFF
counts[message_type] = counts.get(message_type, 0) + 1
i = frame_end
return counts
def parse_mixed_payload(payload: bytes) -> list[dict[str, Any]]:

View File

@@ -1,22 +0,0 @@
from __future__ import annotations
def summarize_rtcm_types(payload: bytes) -> dict[int, int]:
counts: dict[int, int] = {}
i = 0
while i + 6 <= len(payload):
if payload[i] != 0xD3:
i += 1
continue
length = ((payload[i + 1] & 0x03) << 8) | payload[i + 2]
frame_end = i + 3 + length + 3
if frame_end > len(payload):
break
message_type = ((payload[i + 3] << 4) | (payload[i + 4] >> 4)) & 0x0FFF
counts[message_type] = counts.get(message_type, 0) + 1
i = frame_end
return counts

View File

@@ -9,7 +9,6 @@ from sqlmodel import Session, col, select
from app.models import (
TerminalDeviceState,
TerminalMessageRaw,
TerminalObservationEvent,
get_datetime_utc,
)
@@ -34,7 +33,6 @@ def insert_raw_message(
payload_sha256=hashlib.sha256(payload).hexdigest(),
qos=qos,
retain=retain,
parse_status="pending",
created_at=ts,
)
session.add(row)
@@ -48,7 +46,8 @@ def fetch_pending_rows_for_update(
) -> list[TerminalMessageRaw]:
statement = (
select(TerminalMessageRaw)
.where(TerminalMessageRaw.parse_status == "pending")
.where(col(TerminalMessageRaw.parsed_at).is_(None))
.where(col(TerminalMessageRaw.parse_error).is_(None))
.order_by(col(TerminalMessageRaw.received_at))
.limit(batch_size)
)
@@ -61,17 +60,9 @@ def fetch_pending_rows_for_update(
return list(rows)
def mark_processing(*, session: Session, row: TerminalMessageRaw) -> None:
row.parse_status = "processing"
row.parse_attempts += 1
row.parse_error = None
session.add(row)
def mark_succeeded(*, session: Session, row: TerminalMessageRaw) -> None:
row.parse_status = "succeeded"
row.parsed_at = get_datetime_utc()
row.parse_error = None
row.next_retry_at = None
session.add(row)
@@ -80,15 +71,13 @@ def mark_failed(
session: Session,
row: TerminalMessageRaw,
parse_error: str,
next_retry_at: datetime | None,
) -> None:
row.parse_status = "failed"
row.parsed_at = get_datetime_utc()
row.parse_error = parse_error[:1024]
row.next_retry_at = next_retry_at
session.add(row)
def write_events_and_update_state(
def update_device_state_from_events(
*,
session: Session,
row: TerminalMessageRaw,
@@ -110,14 +99,6 @@ def write_events_and_update_state(
for event in events:
payload = event.get("payload", {})
event_type = str(event.get("event_type", "unknown"))
db_event = TerminalObservationEvent(
raw_id=row.id,
device_id=row.device_id,
event_time=row.received_at,
event_type=event_type,
payload_json=payload if isinstance(payload, dict) else {"value": payload},
)
session.add(db_event)
if event_type == "nmea_gga":
state.lat = payload.get("lat")

View File

@@ -1,10 +0,0 @@
import re
TOPIC_PATTERN = re.compile(r"^terminal/(?P<device_id>[^/]+)/raw$")
def extract_device_id(topic: str) -> str:
match = TOPIC_PATTERN.match(topic)
if match is None:
raise ValueError("topic must match terminal/{device_id}/raw")
return match.group("device_id")