chess-spectral 1.2.5

Spectral lattice-fermion encoder for 2D and 4D chess (D4 / B4 symmetry, graph-Laplacian eigenbasis, 640-dim / 45 056-dim HDC)

chess_spectral (Python)

Python reference implementation of the 640-dim (2D) and 45 056-dim (4D) spectral chess encoders, sibling of the C17 port in ../src/. Use this for REPL / LLM / notebook analysis; use the C binary for batch throughput.

Install

From PyPI (recommended):

pip install chess-spectral

The base install pulls only numpy and scipy. If you need PGN ingest via chess_spectral.corpus, request the [corpus] extra to add python-chess:

pip install "chess-spectral[corpus]"

Package page: https://pypi.org/project/chess-spectral/

From source

Editable install from a local checkout:

pip install -e docs/chess-maths/chess-spectral/python/

From a git URL (pin a commit in production):

pip install "git+https://github.com/lemonforest/mlehaptics.git@COMMIT#subdirectory=docs/chess-maths/chess-spectral/python"

After install, two console scripts are on your $PATH:

chess-spectral --help            # 2D CLI (formerly `python spectral_py.py`)
chess-spectral-4d --help         # 4D CLI

Both packages also expose __version__:

>>> import chess_spectral, chess_spectral_4d
>>> chess_spectral.__version__, chess_spectral_4d.__version__
('1.1.3', '1.1.3')

In-place (no install)

The legacy workflow still works: every test and analysis script uses sys.path.insert to bootstrap off the python/ directory, so pytest docs/chess-maths/chess-spectral/python/tests/ runs without any install.

Output is byte-identical to the C CLI — the spectral csv command here produces the same bytes as the C spectral csv does on the same input file. The Python CLI (chess-spectral, entry point chess_spectral.cli:main) mirrors the C CLI subcommand-for-subcommand.

Layout

chess_spectral/
  encoder.py             # encode_640(pos) → np.ndarray(640,)
  frame.py               # v2 .spectral[z] binary I/O + transparent gzip
  csv_export.py          # dist_prev / cos_prev / energies CSV
  cli.py                 # `chess-spectral csv file.spectralz` (2D CLI)
  phase_operators/       # §11 phase-space move generator + is_check (1.2.0+)
chess_spectral_4d/
  cli.py                 # `chess-spectral-4d tables-verify --phase all`
pyproject.toml           # PEP 621 packaging metadata
tests/
  test_parity.py         # Python output == C output (5 tests)
  phase_operators/       # §11 validation suite (92 tests; 1.2.0+)

Quick start

>>> from chess_spectral import (
...     encode_640, channel_energies, read_encodings, fen_to_pos,
... )

>>> pos = fen_to_pos("rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/RNBQKBNR w KQkq - 0 1")
>>> enc = encode_640(pos)
>>> enc.shape
(640,)

>>> channel_energies(enc)
{'A1': 0.0, 'A2': 19.845, 'B1': 45.2825, 'B2': 45.2825,
 'E': 322.57, 'F1': 88.77, 'F2': 1851.01, 'F3': 1507.65,
 'FA': 19.92, 'FD': 0.0}

# Read a whole game that was encoded by either C or Python
>>> hdr, arr = read_encodings("game.spectralz")  # transparent gzip
>>> arr.shape
(161, 640)

CLI

python spectral_py.py csv        game.spectralz -o game.csv
python spectral_py.py encode     -i game.ndjson -o game.spectralz -z
python spectral_py.py encode-fen --fen "..."   -o single.spectral
python spectral_py.py version

Phase operators (§11)

chess_spectral ships a phase-space move generator and check detector as the chess_spectral.phase_operators subpackage (added in 1.2.0). The primitives compute all moves and check relationships as modular arithmetic on a single integer per square — phi(r, c) = r·67 + c·7 mod 640 — rather than geometric coordinates. They are a drop-in equivalent to python-chess's pseudo_legal_moves + is_check, validated at 100% on the reference corpus, and compose naturally with the spectral encoder's coprime phase structure.

import chess
from chess_spectral.phase_operators import (
    occupation_aware_moves_c,   # pseudo-legal dests from a square
    available_castles,          # legal castles for side-to-move
    phasecast_is_check,         # is the mover's king attacked?
    move_leaves_king_in_check,  # would this move expose our king?
)

board = chess.Board()

# Pseudo-legal destinations for the knight on b1 (the 1,2 indexed
# from bottom-left: row 0, col 1). mover_charge is +1 for white.
dests = occupation_aware_moves_c(board, "N", 0, 1, +1)
# -> frozenset({(2, 0), (2, 2)})   (a3 and c3)

# Check detection
phasecast_is_check(board)  # False on the starting position

Validation coverage (see PHASE_OPERATOR_SUPPLEMENT.md for the full research record):

• Empty-board pseudo-legal destinations for every piece type at every square — 416 / 416 pairs (§11.3).
• Occupation-aware pseudo-legal moves including en passant and castling — 1153 / 1153 at n=100 positions (§11.4).
• is_check on arbitrary pseudo-legal transitions — 3393 / 3393 (§11.5 path 1).

When to use phase operators vs python-chess

Both give you the same answers. The phase-space formulation is faster than python-chess's pseudo_legal_moves in pure Python (~7 µs/piece for the fastest solution vs ~15–20 µs for python-chess's equivalent), and slower than python-chess's bitboard is_check (phasecast_is_check runs ~138 µs vs ~22 µs for python-chess). The subpackage exists because the phase-space formulation is the substrate the §11/§12 spectral research is built on — not because it universally out-performs python-chess. If you're writing a search engine hot loop, python-chess's bitboards are the right choice. If you're doing phase-space research (similarity, partition detection, encoder composition) or need the operators as primitives in C via a future codegen port, the subpackage is the right choice.

Parity with C

The encoder uses the same tables (PAWN_ANTI_FIBER, DIAG_DEV) as the codegen that feeds the C side. Tables are rebuilt at import time from encoder_512 primitives plus the directed pawn adjacency from chess_pawn_laplacian — so both implementations can be verified from first principles.

Run the parity suite:

python tests/test_parity.py

The critical test is test_csv_matches_c_byte_for_byte — reads the C- produced .spectral file, runs the Python CSV exporter, asserts the output equals the C CSV bit-for-bit.

Why two implementations

	C	Python
Throughput	µs/encode	ms/encode
REPL / notebooks	✗	✓
LLM-pasteable	binary	code
Scipy / numpy exploration	✗	✓
Embeds in mobile / web	✓	✗
Exact numerical reference	tables baked at build	rebuilt from primitives

Develop new channels in Python first (faster iteration, scipy.linalg at hand, no rebuild loop). Once the math is frozen, port to C and verify parity via the test suite.