DiagLogViewer — Vehicle Diagnostic Log Analyzer

Desktop GUI tool for parsing, filtering, and analyzing vehicle diagnostic logs
AUTOSAR DLT | CAN Bus | UDS | OBD-II | J1939 | C++ Frame Parser | Anomaly Detection

---

Overview

Working with automotive diagnostic logs during my MSc, I kept running into the same problem: no good open-source tool existed for visually parsing and filtering AUTOSAR DLT and CAN bus logs together. Most workflows involved grep, manual scripts, and Excel — slow and error-prone.

DiagLogViewer fixes that. It's a PyQt5 desktop app that parses DLT and CAN .asc logs, decodes UDS service calls, detects anomalies automatically, and generates HTML reports — all in one window.

---

Features

Feature	Details
DLT Log Parser	AUTOSAR Diagnostic Log and Trace format with UDS service decoding
CAN Bus Parser	Vector CANalyzer .asc format, OBD-II PID decoding, J1939 PGN decoding
C++ Frame Decoder	Native C++17 parser for UDS/OBD-II/J1939 — called from Python via subprocess
Anomaly Detection	Log gaps, error bursts, OTA failures, SecurityAccess violations
Performance Analysis	Mean/P95/P99 inter-message gap, message rate, per-ECU breakdown
Visualization	Embedded matplotlib charts (timing series, message rate histogram)
HTML Report Export	Jinja2-rendered professional diagnostic report
Real-time Filtering	ECU ID, App ID, Log Level, Keyword, Time Range

---

Architecture

DiagLogViewer/
├── main.py                    # PyQt5 GUI main window
├── parser/
│   ├── dlt_parser.py          # AUTOSAR DLT format parser + UDS service decoder
│   ├── can_parser.py          # CAN .asc parser + Python-side OBD-II decode
│   └── can_frame_parser.cpp   # C++17 low-level CAN frame decoder (UDS/OBD/J1939)
├── analyzer/
│   ├── anomaly_detector.py    # Gap, error burst, OTA, security anomaly detection
│   └── performance_analyzer.py # Timing stats: mean, P95, P99, per-ECU rate
├── reporter/
│   └── html_reporter.py       # Jinja2 HTML diagnostic report generator
├── tests/
│   ├── test_dlt_parser.py     # 10 unit tests for DLT parser
│   └── test_anomaly_detector.py # 6 unit tests for anomaly detection
└── .github/workflows/ci.yml   # GitHub Actions CI (lint + test + C++ compile)

---

Quick Start

# 1. Install Python dependencies
pip install PyQt5 matplotlib jinja2

# 2. Compile the C++ CAN frame parser (optional but recommended)
cd parser
g++ -O2 -std=c++17 -o can_frame_parser can_frame_parser.cpp

# 3. Run the GUI
python main.py

The app loads with demo DLT data automatically — no log file needed to explore features.

---

Usage

1. Open DLT Log — load an AUTOSAR .dlt or .txt log file
2. Open CAN Log — load a Vector CANalyzer .asc file
3. Load Demo — explore with realistic synthetic vehicle diagnostic data
4. Filter — filter by ECU ID, App ID, Log Level, keyword, or time range
5. Anomalies tab — view detected anomalies with severity ratings
6. Performance Stats tab — view timing metrics (mean, P95, P99, per-ECU rate)
7. Charts — visualize inter-message gap or message rate over time
8. Export HTML Report — generate a professional diagnostic report

---

UDS Services Decoded

Service ID	Name
0x10	DiagnosticSessionControl
0x22	ReadDataByIdentifier
0x27	SecurityAccess
0x2E	WriteDataByIdentifier
0x34	RequestDownload (OTA)
0x36	TransferData (OTA)
0x7F	NegativeResponse
...	15+ services total

---

Testing

pip install pytest pytest-cov
pytest tests/ -v --cov=.

16 unit tests covering parser, anomaly detection, and performance analysis.

---

Author

Jattin Shah — MSc Applied AI, TU Dresden
github.com/J4jatin | linkedin.com/in/jattin-shah