fix: prevent tdengine raw frame overwrite
All checks were successful
ci/woodpecker/push/woodpecker Pipeline was successful

This commit is contained in:
lingniu
2026-06-29 18:41:58 +08:00
parent 5c1c61f29b
commit 8804c01d99
9 changed files with 422 additions and 6 deletions

View File

@@ -225,15 +225,23 @@ JT808 真实转发链路已验证:
仓库提供 `tools/jt808_e2e_smoke.py`,用于重复验证 TCP `808` 到 TDengine history 的闭环:
```bash
export TDENGINE_REST_URL='http://<tdengine-host>:6041/rest/sql/vehicle_ts'
export TDENGINE_USERNAME='root'
export TDENGINE_PASSWORD='<tdengine-password>'
python3 tools/jt808_e2e_smoke.py \
--connection-mode session \
--start-phone 13079962000 \
--frames 3 \
--expect-history-count 3 \
--verify-pagination \
--archive-root "$PROJECT_ROOT/data/archive-jt808"
--archive-root "$PROJECT_ROOT/data/archive-jt808" \
--tdengine-rest-url "$TDENGINE_REST_URL" \
--tdengine-username "$TDENGINE_USERNAME" \
--tdengine-password "$TDENGINE_PASSWORD"
```
默认 `--connection-mode per-frame`,适合稳定 smoke。需要暴露长连接连续帧边界时可切换 `--connection-mode session`。脚本会输出发送数量、history 可见数量、分页验证结果raw archive 检查结果。
默认 `--connection-mode per-frame`,适合单帧连接 smoke。生产验收建议使用 `--connection-mode session` 覆盖同一连接内连续位置帧。脚本会输出发送数量、history 可见数量、分页验证结果raw archive 检查结果`tdengineRawFrames``tdengineRawFrames` 表示本次可见位置记录中的唯一 `rawUri` 有多少个已确认写入 TDengine `raw_frames`
- 合成 0x0100 注册帧终端号 `13079969999` 已验证TCP `808` 返回 `0x8100` 注册 ACK`raw_frames.metadata_json` 包含 `jt808.register.province``jt808.register.city``jt808.register.maker``jt808.register.deviceType``jt808.register.deviceId``jt808.register.plateColor``jt808.register.plate`,对应 raw archive 文件存在。
复查 SQL
@@ -277,13 +285,29 @@ SELECT COUNT(*) FROM raw_frames
WHERE protocol = 'GB32960' AND vin = 'LTEST202606290001';
SELECT COUNT(*) FROM telemetry_fields
WHERE protocol = 'GB32960' AND vin = 'LTEST202606290001';
SELECT event_time, vin, raw_uri, raw_id
SELECT event_time, vin, raw_uri, frame_id
FROM raw_frames
WHERE protocol = 'GB32960' AND vin = 'LTEST202606290001'
ORDER BY event_time DESC
LIMIT 5;
```
也可以直接运行仓库 smoke 工具验证 TCP `32960`、history records、TDengine 字段、TDengine `raw_frames` 和 raw archive
```bash
export TDENGINE_REST_URL='http://<tdengine-host>:6041/rest/sql/vehicle_ts'
export TDENGINE_USERNAME='root'
export TDENGINE_PASSWORD='<tdengine-password>'
python3 tools/gb32960_e2e_smoke.py \
--archive-root "$PROJECT_ROOT/data/archive-jt808" \
--tdengine-rest-url "$TDENGINE_REST_URL" \
--tdengine-username "$TDENGINE_USERNAME" \
--tdengine-password "$TDENGINE_PASSWORD"
```
预期输出包含 `records >= 1``fieldCounts` 中关键字段均大于 0、`tdengineRawFrames` 等于本次可见唯一 `rawUri` 数量、`archiveChecked >= 1`
复查 API
```bash

View File

@@ -30,7 +30,7 @@ public final class TdengineEnvelopeRows {
metadata.putAll(envelope.getMetadataMap());
metadata.putAll(raw.getMetadataMap());
return Optional.of(new TdengineRawFrameRow(
instant(envelope.getEventTimeMs()),
rawFrameTs(envelope, metadata),
raw.getFrameId(),
instant(envelope.getIngestTimeMs()),
raw.getMessageId(),
@@ -142,6 +142,24 @@ public final class TdengineEnvelopeRows {
return Instant.ofEpochMilli(epochMillis);
}
private static Instant rawFrameTs(VehicleEnvelope envelope, Map<String, String> metadata) {
String rawArchiveEventId = firstNonBlank(metadata.get("rawArchiveEventId"), envelope.getEventId());
if (rawArchiveEventId == null || rawArchiveEventId.isBlank()) {
return instant(envelope.getEventTimeMs());
}
try {
long sequenceId = Long.parseLong(rawArchiveEventId);
long baseMillis = sequenceId / 1000L;
long sameMillisecondOffset = sequenceId % 1000L;
if (Math.abs(baseMillis - envelope.getEventTimeMs()) > 60_000L) {
return instant(envelope.getEventTimeMs());
}
return Instant.ofEpochMilli(baseMillis + sameMillisecondOffset);
} catch (NumberFormatException ignored) {
return instant(envelope.getEventTimeMs());
}
}
private static String protocol(VehicleEnvelope envelope) {
return firstNonBlank(envelope.getSource(), "UNKNOWN");
}

View File

@@ -51,6 +51,20 @@ class TdengineEnvelopeRowsTest {
assertThat(row.metadataJson()).contains("\"channel\":\"tcp-808\"", "\"auth\":\"passed\"");
}
@Test
void expandsRawFrameTsWithRawArchiveEventIdSequenceToAvoidSameMillisecondOverwrite() {
VehicleEnvelope first = rawFrameEnvelope("1782729152494000");
VehicleEnvelope second = rawFrameEnvelope("1782729152494001");
TdengineRawFrameRow firstRow = TdengineEnvelopeRows.rawFrame(first).orElseThrow();
TdengineRawFrameRow secondRow = TdengineEnvelopeRows.rawFrame(second).orElseThrow();
assertThat(firstRow.ts()).isEqualTo(Instant.ofEpochMilli(1_782_729_152_494L));
assertThat(secondRow.ts()).isEqualTo(Instant.ofEpochMilli(1_782_729_152_495L));
assertThat(firstRow.eventTime()).isEqualTo(Instant.ofEpochMilli(1_782_729_152_494L));
assertThat(secondRow.eventTime()).isEqualTo(Instant.ofEpochMilli(1_782_729_152_494L));
}
@Test
void mapsLocationEnvelopeToVehicleLocationRow() {
VehicleEnvelope envelope = VehicleEnvelope.newBuilder()
@@ -190,4 +204,25 @@ class TdengineEnvelopeRowsTest {
assertThat(TdengineEnvelopeRows.telemetryFields(envelope).getFirst().vehicleKey())
.isEqualTo("jt808:13079963291");
}
private static VehicleEnvelope rawFrameEnvelope(String rawArchiveEventId) {
return VehicleEnvelope.newBuilder()
.setEventId(rawArchiveEventId)
.setVin("unknown")
.setSource("JT808")
.setEventTimeMs(1_782_729_152_494L)
.setIngestTimeMs(1_782_729_152_539L)
.putMetadata("phone", "13079973200")
.putMetadata("rawArchiveEventId", rawArchiveEventId)
.setRawFrameFact(RawFrameFactPayload.newBuilder()
.setFrameId("rf-" + rawArchiveEventId)
.setPhone("13079973200")
.setMessageId(0x0200)
.setRawUri("archive://2026/06/29/JT808/unknown/" + rawArchiveEventId + ".bin")
.setChecksum("sha256:" + rawArchiveEventId)
.setRawSizeBytes(41)
.setParseStatus(ParseStatusProto.PARSE_STATUS_NOT_PARSED)
.putMetadata("rawArchiveEventId", rawArchiveEventId))
.build();
}
}

View File

@@ -6,6 +6,7 @@ from __future__ import annotations
import argparse
import datetime as dt
import json
import os
import pathlib
import socket
import time
@@ -14,6 +15,11 @@ import urllib.request
from dataclasses import dataclass
from typing import Any
try:
from tools import tdengine_smoke
except ModuleNotFoundError:
import tdengine_smoke
SHANGHAI = dt.timezone(dt.timedelta(hours=8))
DEFAULT_SAMPLE = (
@@ -144,6 +150,25 @@ def wait_for_fields(
return last_items
def wait_for_tdengine_raw_frames(
rest_url: str,
username: str,
password: str,
sql: str,
minimum: int,
timeout_seconds: float,
request_timeout: float,
) -> int:
deadline = time.monotonic() + timeout_seconds
last_count = 0
while time.monotonic() < deadline:
last_count = tdengine_smoke.query_count(rest_url, username, password, sql, request_timeout)
if last_count >= minimum:
return last_count
time.sleep(1)
return last_count
def parse_field_keys(value: str) -> list[str]:
return [item.strip() for item in value.split(",") if item.strip()]
@@ -172,6 +197,50 @@ def check_archive(root: pathlib.Path, records: list[dict[str, Any]], limit: int)
return checked
def raw_frame_count_sql(info: FrameInfo, records: list[dict[str, Any]]) -> str:
sqls = raw_frame_count_sqls(info, records)
if not sqls:
raise ValueError("no raw archive uri found in gb32960 records")
return sqls[0]
def raw_frame_count_sqls(info: FrameInfo, records: list[dict[str, Any]]) -> list[str]:
seen: set[str] = set()
sqls: list[str] = []
for record in records:
uri = record.get("rawArchiveUri") or ""
if not uri or uri in seen:
continue
seen.add(uri)
sqls.append(tdengine_smoke.raw_frame_count_sql(
protocol="GB32960",
vehicle_key=info.vin,
vin=info.vin,
raw_uri=uri,
))
return sqls
def verify_tdengine_raw_frames(
rest_url: str,
username: str,
password: str,
sqls: list[str],
timeout_seconds: float,
request_timeout: float,
) -> int:
if not sqls:
raise RuntimeError("no raw archive uri available for tdengine raw_frames verification")
verified = 0
for sql in sqls:
count = wait_for_tdengine_raw_frames(
rest_url, username, password, sql, 1, timeout_seconds, request_timeout)
if count < 1:
raise RuntimeError(f"tdengine raw_frames row not visible for query: {sql}")
verified += 1
return verified
def run(args: argparse.Namespace) -> dict[str, Any]:
frame = read_hex_frame(pathlib.Path(args.frame_hex_file))
info = parse_frame_info(frame)
@@ -212,6 +281,17 @@ def run(args: argparse.Namespace) -> dict[str, Any]:
if args.archive_root:
archive_checked = check_archive(pathlib.Path(args.archive_root), records, args.archive_check_limit)
tdengine_raw_frames = None
if args.tdengine_rest_url:
tdengine_raw_frames = verify_tdengine_raw_frames(
args.tdengine_rest_url,
args.tdengine_username,
args.tdengine_password,
raw_frame_count_sqls(info, records),
args.history_timeout,
args.http_timeout,
)
return {
"tcpHost": args.tcp_host,
"tcpPort": args.tcp_port,
@@ -221,6 +301,7 @@ def run(args: argparse.Namespace) -> dict[str, Any]:
"ackBytesHex": ack.hex(),
"records": len(records),
"fieldCounts": field_counts,
"tdengineRawFrames": tdengine_raw_frames,
"archiveChecked": archive_checked,
"firstRawUri": first_raw_uri(records),
}
@@ -245,6 +326,9 @@ def parser() -> argparse.ArgumentParser:
p.add_argument("--expect-record-count", type=int, default=1)
p.add_argument("--archive-root", default="")
p.add_argument("--archive-check-limit", type=int, default=1)
p.add_argument("--tdengine-rest-url", default=os.environ.get("TDENGINE_REST_URL", ""))
p.add_argument("--tdengine-username", default=os.environ.get("TDENGINE_USERNAME", "root"))
p.add_argument("--tdengine-password", default=os.environ.get("TDENGINE_PASSWORD", "taosdata"))
p.add_argument("--require-ack", action="store_true", default=True)
p.add_argument("--tcp-timeout", type=float, default=5)
p.add_argument("--http-timeout", type=float, default=5)

View File

@@ -6,6 +6,7 @@ from __future__ import annotations
import argparse
import datetime as dt
import json
import os
import pathlib
import socket
import time
@@ -13,6 +14,11 @@ import urllib.parse
import urllib.request
from typing import Any
try:
from tools import tdengine_smoke
except ModuleNotFoundError:
import tdengine_smoke
def bcd(value: str) -> bytes:
digits = "".join(ch for ch in value if ch.isdigit())
@@ -247,6 +253,69 @@ def archive_path(root: pathlib.Path, uri: str) -> pathlib.Path:
return root / uri[len(prefix):]
def raw_frame_count_sql(phone: str, items: list[dict[str, Any]]) -> str:
sqls = raw_frame_count_sqls(phone, items)
if not sqls:
raise ValueError("no raw uri found in jt808 location items")
return sqls[0]
def raw_frame_count_sqls(phone: str, items: list[dict[str, Any]]) -> list[str]:
seen: set[str] = set()
sqls: list[str] = []
for item in items:
raw_uri = item.get("rawUri") or ""
if not raw_uri or raw_uri in seen:
continue
seen.add(raw_uri)
sqls.append(tdengine_smoke.raw_frame_count_sql(
protocol="JT808",
vehicle_key=item.get("vehicleKey") or ("jt808:" + phone),
phone=phone,
raw_uri=raw_uri,
))
return sqls
def verify_tdengine_raw_frames(
rest_url: str,
username: str,
password: str,
sqls: list[str],
timeout_seconds: float,
request_timeout: float,
) -> int:
if not sqls:
raise RuntimeError("no raw uri available for tdengine raw_frames verification")
verified = 0
for sql in sqls:
count = wait_for_tdengine_raw_frames(
rest_url, username, password, sql, 1, timeout_seconds, request_timeout)
if count < 1:
raise RuntimeError(f"tdengine raw_frames row not visible for query: {sql}")
verified += 1
return verified
def wait_for_tdengine_raw_frames(
rest_url: str,
username: str,
password: str,
sql: str,
minimum: int,
timeout_seconds: float,
request_timeout: float,
) -> int:
deadline = time.monotonic() + timeout_seconds
last_count = 0
while time.monotonic() < deadline:
last_count = tdengine_smoke.query_count(rest_url, username, password, sql, request_timeout)
if last_count >= minimum:
return last_count
time.sleep(1)
return last_count
def wait_for_locations(
history_base_url: str,
phone: str,
@@ -285,6 +354,13 @@ def generated_phone(start_phone: str, offset: int) -> str:
return str(int(start_phone) + offset)
def expected_history_count(args: argparse.Namespace) -> int:
minimum = min(args.frames, args.expect_history_count)
if args.verify_pagination and args.frames > 1:
minimum = max(minimum, 2)
return minimum
def run(args: argparse.Namespace) -> dict[str, Any]:
event_time = parse_event_time(args.event_time)
date_from = (event_time - dt.timedelta(minutes=5)).isoformat()
@@ -339,12 +415,12 @@ def run(args: argparse.Namespace) -> dict[str, Any]:
first_phone,
date_from,
date_to,
min(args.frames, args.expect_history_count),
expected_history_count(args),
args.history_timeout,
args.http_timeout,
)
items = page.get("items") or []
if len(items) < min(args.frames, args.expect_history_count):
if len(items) < expected_history_count(args):
raise RuntimeError(f"history rows not visible for {first_phone}: got {len(items)}")
archive_checked = 0
@@ -356,6 +432,17 @@ def run(args: argparse.Namespace) -> dict[str, Any]:
raise RuntimeError(f"raw archive missing: {path}")
archive_checked += 1
tdengine_raw_frames = None
if args.tdengine_rest_url:
tdengine_raw_frames = verify_tdengine_raw_frames(
args.tdengine_rest_url,
args.tdengine_username,
args.tdengine_password,
raw_frame_count_sqls(first_phone, items),
args.history_timeout,
args.http_timeout,
)
pagination_ok = None
if args.verify_pagination and args.frames > 1:
pagination_ok = verify_pagination(args.history_base_url, first_phone, date_from, date_to, args.http_timeout)
@@ -372,6 +459,7 @@ def run(args: argparse.Namespace) -> dict[str, Any]:
"sendElapsedSeconds": round(elapsed, 3),
"sendLocationsPerSecond": round(sent_locations / elapsed, 2),
"historyRowsForFirstPhone": len(items),
"tdengineRawFrames": tdengine_raw_frames,
"archiveChecked": archive_checked,
"paginationOk": pagination_ok,
"firstFrameId": items[0].get("frameId") if items else None,
@@ -396,6 +484,9 @@ def parser() -> argparse.ArgumentParser:
p.add_argument("--verify-pagination", action="store_true")
p.add_argument("--archive-root", default="")
p.add_argument("--archive-check-limit", type=int, default=3)
p.add_argument("--tdengine-rest-url", default=os.environ.get("TDENGINE_REST_URL", ""))
p.add_argument("--tdengine-username", default=os.environ.get("TDENGINE_USERNAME", "root"))
p.add_argument("--tdengine-password", default=os.environ.get("TDENGINE_PASSWORD", "taosdata"))
p.add_argument("--require-register-ack", action="store_true", default=True)
p.add_argument("--tcp-timeout", type=float, default=2)
p.add_argument("--http-timeout", type=float, default=5)

61
tools/tdengine_smoke.py Normal file
View File

@@ -0,0 +1,61 @@
#!/usr/bin/env python3
"""Small TDengine REST helpers for local smoke verifiers."""
from __future__ import annotations
import base64
import json
import urllib.request
from typing import Any
def sql_literal(value: str) -> str:
return "'" + (value or "").replace("'", "''") + "'"
def raw_frame_count_sql(
*,
protocol: str,
vehicle_key: str = "",
vin: str = "",
phone: str = "",
raw_uri: str = "",
) -> str:
if not protocol or not protocol.strip():
raise ValueError("protocol is required")
filters = ["protocol = " + sql_literal(protocol.strip().upper())]
optional_filters = {
"vehicle_key": vehicle_key,
"vin": vin,
"phone": phone,
"raw_uri": raw_uri,
}
present = False
for column, value in optional_filters.items():
if value and value.strip():
filters.append(column + " = " + sql_literal(value.strip()))
present = True
if not present:
raise ValueError("at least one identity or raw_uri filter is required")
return "SELECT COUNT(*) FROM raw_frames WHERE " + " AND ".join(filters)
def count_from_response(response: dict[str, Any]) -> int:
if response.get("code") != 0:
raise RuntimeError(f"tdengine query failed: {response}")
data = response.get("data") or []
if not data or not data[0]:
return 0
return int(data[0][0])
def query_count(rest_url: str, username: str, password: str, sql: str, timeout: float) -> int:
request = urllib.request.Request(rest_url, data=sql.encode("utf-8"), method="POST")
if username:
token = base64.b64encode(f"{username}:{password}".encode("utf-8")).decode("ascii")
request.add_header("Authorization", "Basic " + token)
request.add_header("Content-Type", "text/plain; charset=utf-8")
request.add_header("Accept", "application/json")
with urllib.request.urlopen(request, timeout=timeout) as response:
payload = json.loads(response.read().decode("utf-8"))
return count_from_response(payload)

View File

@@ -58,6 +58,32 @@ class Gb32960E2ESmokeTest(unittest.TestCase):
self.assertEqual(uri, "archive://2026/06/29/GB32960/LTEST/frame.bin")
def test_raw_frame_count_sql_filters_by_vin_and_raw_uri(self):
info = smoke.FrameInfo("LTEST202606290001", 0x02, None)
sql = smoke.raw_frame_count_sql(info, [
{"rawArchiveUri": "archive://2026/06/29/GB32960/LTEST202606290001/frame.bin"},
])
self.assertEqual(
sql,
"SELECT COUNT(*) FROM raw_frames WHERE protocol = 'GB32960' "
"AND vehicle_key = 'LTEST202606290001' "
"AND vin = 'LTEST202606290001' "
"AND raw_uri = 'archive://2026/06/29/GB32960/LTEST202606290001/frame.bin'",
)
def test_raw_frame_count_sqls_include_each_unique_record_raw_uri(self):
info = smoke.FrameInfo("LTEST202606290001", 0x02, None)
sqls = smoke.raw_frame_count_sqls(info, [
{"rawArchiveUri": "archive://2026/06/29/GB32960/LTEST202606290001/frame-1.bin"},
{"rawArchiveUri": "archive://2026/06/29/GB32960/LTEST202606290001/frame-2.bin"},
{"rawArchiveUri": "archive://2026/06/29/GB32960/LTEST202606290001/frame-1.bin"},
])
self.assertEqual(len(sqls), 2)
self.assertTrue(sqls[0].endswith("frame-1.bin'"))
self.assertTrue(sqls[1].endswith("frame-2.bin'"))
def test_check_archive_ignores_duplicate_uris(self):
with tempfile.TemporaryDirectory() as temp_dir:
root = pathlib.Path(temp_dir)

View File

@@ -1,4 +1,5 @@
import pathlib
import types
import unittest
from tools import jt808_e2e_smoke as smoke
@@ -33,6 +34,47 @@ class Jt808E2ESmokeTest(unittest.TestCase):
self.assertEqual(path, pathlib.Path("/tmp/archive/2026/06/29/JT808/unknown/frame.bin"))
def test_raw_frame_count_sql_filters_by_vehicle_key_phone_and_raw_uri(self):
sql = smoke.raw_frame_count_sql("13079961234", [
{
"vehicleKey": "jt808:13079961234",
"rawUri": "archive://2026/06/29/JT808/13079961234/frame.bin",
},
])
self.assertEqual(
sql,
"SELECT COUNT(*) FROM raw_frames WHERE protocol = 'JT808' "
"AND vehicle_key = 'jt808:13079961234' "
"AND phone = '13079961234' "
"AND raw_uri = 'archive://2026/06/29/JT808/13079961234/frame.bin'",
)
def test_raw_frame_count_sqls_include_each_unique_visible_raw_uri(self):
sqls = smoke.raw_frame_count_sqls("13079961234", [
{
"vehicleKey": "jt808:13079961234",
"rawUri": "archive://2026/06/29/JT808/13079961234/frame-1.bin",
},
{
"vehicleKey": "jt808:13079961234",
"rawUri": "archive://2026/06/29/JT808/13079961234/frame-2.bin",
},
{
"vehicleKey": "jt808:13079961234",
"rawUri": "archive://2026/06/29/JT808/13079961234/frame-1.bin",
},
])
self.assertEqual(len(sqls), 2)
self.assertTrue(sqls[0].endswith("frame-1.bin'"))
self.assertTrue(sqls[1].endswith("frame-2.bin'"))
def test_expected_history_count_waits_for_pagination_sample(self):
args = types.SimpleNamespace(frames=3, expect_history_count=1, verify_pagination=True)
self.assertEqual(smoke.expected_history_count(args), 2)
if __name__ == "__main__":
unittest.main()

View File

@@ -0,0 +1,35 @@
import unittest
from tools import tdengine_smoke
class TdengineSmokeTest(unittest.TestCase):
def test_sql_literal_escapes_single_quotes(self):
self.assertEqual(tdengine_smoke.sql_literal("VIN'001"), "'VIN''001'")
def test_raw_frame_count_sql_filters_by_protocol_vehicle_and_raw_uri(self):
sql = tdengine_smoke.raw_frame_count_sql(
protocol="JT808",
vehicle_key="jt808:13079962000",
raw_uri="archive://2026/06/29/JT808/frame.bin",
)
self.assertEqual(
sql,
"SELECT COUNT(*) FROM raw_frames WHERE protocol = 'JT808' "
"AND vehicle_key = 'jt808:13079962000' "
"AND raw_uri = 'archive://2026/06/29/JT808/frame.bin'",
)
def test_raw_frame_count_sql_requires_at_least_one_identity_or_raw_uri_filter(self):
with self.assertRaises(ValueError):
tdengine_smoke.raw_frame_count_sql(protocol="GB32960")
def test_count_from_response_reads_first_cell(self):
response = {"code": 0, "data": [[3]], "rows": 1}
self.assertEqual(tdengine_smoke.count_from_response(response), 3)
if __name__ == "__main__":
unittest.main()