CAN Scope

A portable, zero-install Windows tool for loading CAN/automotive measurement files, decoding signals with DBC or ARXML (AUTOSAR) databases, and plotting them interactively.

---

Supported Formats

Format	Extension	Database Required
Vector Binary Logging Format	.blf	Yes
Vector CANalyzer ASCII Log	.asc	Yes
ASAM MDF4 — bus logging (raw CAN frames)	.mf4	Yes — auto-detected
ASAM MDF4 / MDF3 — pre-decoded signals	.mf4 / .mdf	No
CSV (narrow or wide columnar)	.csv	No

MDF auto-detection: CAN Scope automatically probes each MDF file on open. If it contains raw CAN bus frames (CAN_DataFrame.* channel groups) it routes through the bus-logging pipeline and requires a database file (.dbc or .arxml). Pre-decoded MDF files load directly without a database.

Database Formats

Both .dbc and .arxml (AUTOSAR System Description) files are accepted wherever a CAN database is required.

• Open an .arxml in the Database Manager exactly as you would a .dbc.
• Mixed configurations are supported: some CAN channels can use a .dbc and others an .arxml.
• Supported ARXML variants: AUTOSAR 3.x and AUTOSAR 4.x.
• To convert an existing .dbc to .arxml for testing: cantools convert input.dbc output.arxml

---

Screenshots

Stacked plot:

Multiple Axis plot:

CAN Trace / Raw Frame viewer:

---

Features

File & Decoding

• Multi-format loading — BLF, ASC, MF4 (bus-logged and pre-decoded), MDF3, and CSV in one tool
• Multi-channel database mapping — assign a different .dbc or .arxml to each CAN channel via the Database Manager; channel 0 acts as an "All Channels" fallback
• Database match quality indicators — the Database Manager shows per-channel decode coverage bars with J1939 PGN fallback scoring
• Streaming decode — signal tree and plots update in real time while large files are still decoding
• Persistent channel config — database-to-channel assignments saved as .canscope_ch JSON and reloaded automatically between sessions

Plot Modes

• Normal — all signals share one Y axis
• Multi-Axis — independent Y axis per signal, all on the same X axis
• Stacked — INCA/CANdb-style lanes, one row per signal with independent Y scaling

Cursors & Navigation

• Dual cursors — draggable C1 and C2 with live value readout in the signal table
• ΔT display — time delta between C1 and C2 shown in the toolbar
• Fit to Window — rescales X and Y to the full extent of all visible data
• Fit Vertical — rescales Y only to match data in the current X view (X range unchanged)
• Both fit operations work correctly across all three plot modes including multi-axis floating ViewBoxes

Signal Table

• 5-column table — Visibility ☑ | Signal | Cursor 1 | Cursor 2 | Unit
• Signal grouping — assign signals to named groups; collapse/expand groups with the arrow; toggle group visibility via the checkbox in the same column
• Drag-and-drop reorder — drag signals or use Ctrl+Up / Ctrl+Down
• Visibility toggle — check/uncheck individual signals or an entire group; cursor position and all other signal values remain stable during toggle
• Signal name display options — optionally prefix signal names with channel number and/or message name

Configuration

• Save / Load configuration — fully persists: file paths, channel-DBC mapping, plotted signal list, group assignments, per-signal visibility, signal colors, plot background color, plot mode, cursor state, table column widths, signal name display flags
• Undo — up to 3 levels of undo for plot actions (Ctrl+Z)

CAN Trace (Raw Frame Viewer)

• On-disk indexed store — no frame cap; 18 B/frame RAM, 64 B/frame on disk with memory-mapped O(1) access
• Sliding window display — shows 5 000 frames at a time with scroll
• Filter by ID and channel — vectorised numpy mask, no disk access for filtering
• On-demand signal decode — click a frame row to see decoded signals using the per-channel DBC

Portable Build

• Single-folder .exe — no Python installation required on the target machine
• Splash screen — shows loading progress; minimisable, appears in taskbar, responds to Win+D / Show Desktop
• Auto-build CI — GitHub Actions builds and uploads a release ZIP on every v*.*.* tag push

---

Getting Started

Run from Source

git clone https://github.com/dinacaran/canscope.git
cd canscope
pip install -r requirements.txt
python app.py

For MDF bus-logging support (raw CAN frames inside MF4) also install:

pip install canmatrix

Download Portable .exe

Go to Releases and download the latest CANScope_vXX.XX.XX_Windows.zip. Unzip anywhere and run CANScope.exe — no installer needed.

---

Usage

Step	Action
1	Open File — select .blf, .asc, .mf4, .mdf, or .csv
2	Open Database — required for BLF, ASC, and MF4 bus-logged files (.dbc or .arxml)
3	Database Manager (optional) — assign a different database per CAN channel
4	Load + Decode — background decode; signal tree populates live
5	Double-click or drag a signal to plot it, or select and press Space
6	Drag Cursor 1 to read time and signal value in the table
7	Toggle Cursor 2 to measure ΔT between two points
8	Switch between Normal, Multi-Axis, and Stacked plot modes
9	Fit to Window or Fit Vertical to rescale the view
10	Save Config (Ctrl+S) — persists file paths, database mapping, plotted signals, and all view state
11	Export CSV to save selected signal data to a time-aligned file
12	CAN Trace to browse raw CAN frames (BLF / ASC / MF4 bus-log)

Keyboard Shortcuts

Key	Action
F	Fit to Window — rescale X and Y to all data
V	Fit Vertical — rescale Y only, keep current X
Space	Plot selected signal(s) from the signal tree
Delete	Remove selected signal from plot
Ctrl+Z	Undo last plot action (up to 3 levels)
Ctrl+S	Save configuration
Ctrl+Up	Move selected signal up
Ctrl+Down	Move selected signal down
Ctrl+Shift+R	Open Raw CAN Frame viewer

---

Build Portable .exe

pip install pyinstaller
pyinstaller CANScope.spec

Output in dist/CANScope/. The GitHub Actions workflow (.github/workflows/build.yml) builds and uploads the ZIP automatically on every tag push matching v*.*.*.

---

Project Structure

canscope/
├── app.py                        # Entry point, APP_NAME="CAN Scope", APP_VERSION
├── config/
│   └── diagnostics/
│       ├── motor_control.yaml    # Fault rules for motor/inverter domain (user-editable)
│       └── README.md             # Rule authoring guide
├── core/
│   ├── models.py                 # RawFrame, DecodedSignalSample dataclasses
│   ├── channel_config.py         # ChannelConfig: {channel → DBC or ARXML}, decoder cache, save/load .canscope_ch
│   ├── load_worker.py            # QThread: CAN-raw / bulk-array / sample-loop decode paths
│   ├── signal_store.py           # SignalStore, SignalSeries (array.array storage)
│   ├── raw_frame_store.py        # On-disk indexed store: 18 B/frame RAM + 64 B/frame disk, mmap
│   ├── dbc_decoder.py            # DBCDecoder with 3-level cache — accepts .dbc and .arxml
│   ├── vectorized_decoder.py     # Vectorised BLF/ASC decode + cached series refs
│   ├── blf_reader.py             # BLFReaderService (wraps python-can)
│   ├── export.py                 # CSV export
│   ├── readers/
│   │   ├── __init__.py           # reader_factory() + dbc_required_for() — format detection
│   │   ├── db_format.py          # SUPPORTED_DB_SUFFIXES, is_database_file(), db_format_label()
│   │   ├── base.py               # MeasurementReader protocol
│   │   ├── blf_can_reader.py     # BLF + DBC/ARXML pipeline
│   │   ├── asc_can_reader.py     # ASC + DBC/ARXML pipeline
│   │   ├── mdf_reader.py         # MF4/MDF pre-decoded via asammdf + is_bus_logging() probe
│   │   ├── mdf_can_reader.py     # MDF4 bus logging via python-can MF4Reader + DBC/ARXML
│   │   └── csv_reader.py         # Wide and narrow CSV
│   └── diagnostics/              # AI-powered diagnostics engine (Ctrl+Shift+A)
│       ├── __init__.py
│       ├── config_loader.py      # YAML domain parser — infers rule type, auto-generates id/title
│       ├── context.py            # DiagnosticContext — read-only SignalStore adapter for rule processors
│       ├── engine.py             # DiagnosticEngine — orchestrates rule runs and evidence building
│       ├── evidence.py           # EvidenceBuilder — reduces large data to <5 KB LLM snippets
│       ├── models.py             # Finding, Severity, AnalysisResult dataclasses
│       ├── rules/
│       │   ├── __init__.py       # RULE_PROCESSORS dispatch table
│       │   ├── expression.py     # Free-form condition evaluation (>, <, =, !=, and, or)
│       │   ├── fault_signal.py   # fault_when operator evaluation
│       │   ├── range_check.py    # min/max boundary check
│       │   └── message_loss.py   # Gap detection between CAN samples
│       └── llm/
│           ├── __init__.py
│           ├── client.py         # GitHubModelsClient — streaming OpenAI-compatible REST
│           └── prompts.py        # Analysis and chat follow-up prompt builders
├── gui/
│   ├── main_window.py            # MainWindow, toolbar, config save/load, plot_finding()
│   ├── plot_widget.py            # PlotPanel: normal / multi-axis / stacked, dual cursors, zoom_to_time()
│   ├── signal_tree.py            # SignalTreeWidget with live search
│   ├── raw_frame_dialog.py       # Sliding-window CAN Trace (RawFrameStore, no cap)
│   ├── dbc_manager.py            # Database Manager dialog: per-channel DBC/ARXML, match quality bars
│   ├── splash.py                 # CANScopeSplash — minimisable, taskbar-visible splash screen
│   └── diagnostics/              # Diagnostics UI (non-modal window)
│       ├── activation.py         # Wires Ctrl+Shift+A shortcut in MainWindow
│       ├── window.py             # DiagnosticsWindow — domain selector, run controls, auto-plot on fault
│       ├── findings_panel.py     # Left panel — severity-coloured finding list + details pane
│       ├── chat_panel.py         # Right panel — streaming LLM chat with status label
│       └── worker.py             # AnalysisWorker / LLMWorker (background thread helpers)
├── resources/
│   ├── splashscreen.png          # 1635 × 962 splash image
│   ├── CANScope_ICON.png         # 1254 × 1254 app icon source
│   └── app_icon.ico              # Multi-resolution ICO (256/128/64/48/32/16 px)
├── requirements.txt
├── CANScope.spec                 # PyInstaller spec, bundles resources/ and config/
└── .github/workflows/build.yml  # Auto-build on v*.*.* tag push

---

Dependencies

Package	Version	Purpose	License
PySide6	≥ 6.7	GUI framework	LGPL v3
pyqtgraph	≥ 0.13	Interactive plots	MIT
python-can	≥ 4.6	BLF / ASC / MF4 bus-log reading	LGPL v3
cantools	≥ 39.4	DBC + ARXML (AUTOSAR 3/4) decoding	MIT
numpy	≥ 1.26	Array operations	BSD
asammdf	≥ 7.0	MF4 / MDF pre-decoded signals	MIT
canmatrix	any	MF4 bus-log support (optional)	LGPL v3
pyinstaller	≥ 6.0	Portable .exe build	GPL v2 (build-only)

---

Code Signing Policy

⚠️ Release builds are currently unsigned. Windows SmartScreen may show a warning on first run — click More info → Run anyway.

---

License

MIT — see LICENSE.