pyarinc is a modern, typed Python library for decoding ARINC 717 and ARINC 767 flight‑data recorder formats. It provides deterministic bit‑extraction utilities, clean parameter models, PRM/VEC configuration parsing, and end‑to‑end decoding into pandas DataFrames.

The library is designed for analysis pipelines, automated QA tooling, and research workflows that require reliable, test‑covered decoding of FDR/QAR data.

The package is not published on PyPI. Install locally:

pip install .

pip install -e .

pip install -e .[test]

Run the full test suite:

pytest
# or
pytest -vv

Tests cover:

• Bit extraction (717/767, MSB‑first, cross‑byte)
• Data‑type decoding (BNR, BCD, DISCRETE, PACKED, CHAR/ASCII, UTC, COB)
• Frame reconstruction (717) and frame parsing (767)
• Scheduling and superframes
• PRM/VEC parsing
• End‑to‑end decoding for both ARINC 717 and ARINC 767

• Detects aligned and bitstream formats
• Converts bitstream → aligned frames
• Reconstructs frames, subframes, and superframes
• Decodes BNR, BCD, DISCRETE, PACKED, CHAR/ISO, UTC
• Applies rate‑based scheduling and superframe rules
• Produces time‑indexed parameter values
• Word‑based parameter indexing via:
  • subframe
  • word
  • bit_offset
• Absolute bit indexing is not used for 717 parameters

• Frame boundary detection (sync word, header, trailer)
• Timestamp extraction with wrap‑around handling
• Parameter extraction using absolute bit indexing (start_bit)
• VEC/PRM‑based configuration
• Full data‑type support:
  • BNR (signed/unsigned)
  • BCD
  • DISCRETE
  • PACKED BITS
  • CHAR / ASCII
  • UTC
  • COB (Computed On Board) with formula evaluation
• Scheduling:
  • Timestamp‑based time axis (primary)
  • Rate‑based scheduling (fallback, 717‑compatible)
• DataFrame output with:
  • time
  • parameter_name
  • value
  • frame_index
  • frame_id
  • valid

pyarinc uses a unified Parameter class with explicit separation between ARINC 717 and ARINC 767 semantics.

• Use word‑based indexing:
  • subframe
  • word
  • bit_offset
• start_bit must remain unset (None)
• Decoded via decode_from_frame()
• Created using:

Parameter.from_717(
    name="ALT",
    bit_length=12,
    data_type="BNR",
    subframe=0,
    word=2,
    bit_offset=4,
)

• Use absolute bit indexing:
  • start_bit (0‑based MSB‑first)
• Optional frame_id_767 for multi‑frame configurations
• Decoded via decode_raw_from_bytes()
• Created using:

Parameter.from_767(
    name="IAS",
    bit_length=16,
    data_type="BNR",
    start_bit=48,
    frame_id_767=3,
)

This separation eliminates ambiguity and prevents accidental misdecoding.

The library parses PRM and VEC formats, including:

• subframe index
• word index
• bit offset
• bit length
• rate
• superframe index
• scale and offset

• start_bit (absolute bit index)
• frame_id_767
• bit length
• rate
• scale and offset
• COB formulas

Both JSON and text formats are supported.

The VEC parser supports a richer and more consistent token set across ARINC 717 and ARINC 767, improving configuration clarity while preserving full compatibility with existing decoders.

• Support for TYPE=, SIGNED=, SCALE=, OFFSET=
• Recognition of bare type tokens (BNR, BCD, CHAR)
• Parsing of CONV= and OPT= (stored in mapping for future use)
• No changes to the Parameter.from_717 API

• Support for bare type tokens (BNR, BCD, CHAR)
• Parsing of SIGNED=, SCALE=, OFFSET=
• Decimal FID= supported; hexadecimal FID=0xNN ignored
• COB formulas stored without overriding the declared data type
• Absolute bit indexing (start_bit = word * 32 + bit_offset) retained

• Unified token handling across 717/767
• Expanded test coverage for all supported tokens and legacy behaviors
• No changes to the public Parameter API

Typical decoding flow:

1. Load raw data (aligned, bitstream, or ARINC 767 frames)
2. Convert bitstream to aligned frames if needed (717)
3. Load PRM or VEC configuration
4. Construct parameters using from_717() or from_767()
5. Decode parameters using scheduling and superframe rules
6. Produce a pandas DataFrame

This project is a clean rewrite inspired by the decoding logic in the original FlightDataDecode repository:

https://github.com/osnosn/FlightDataDecode

pyarinc re‑implements the core logic with a modern architecture, strict typing, and full test coverage. No legacy scripts or .dat formats are included.

• No Lua integration
• No custom .dat format
• No print() statements — uses Python logging
• Fully typed (Python 3.12+)
• Deterministic, testable, modular design