csdigit package

Submodules

csdigit.cli module

Command-line interface for CSD conversion.

Provides a CLI wrapper around the csdigit library, allowing conversion between decimal numbers and Canonical Signed Digit (CSD) representation from the terminal.

Supported conversions:

  • Decimal float to CSD (–to_csd / -c)

  • Decimal float to CSD with limited non-zero digits (–to_csdnnz / -f)

  • CSD string to decimal (–to_decimal / -d)

Usage:

python -m csdigit.cli –to_csd 28.5 python -m csdigit.cli –to_decimal “+00-00.+” python -m csdigit.cli –to_csdnnz 28.5 –nnz 4

csdigit.cli.main(args: List[str]) → None[source]

Wrapper allowing main() to be called with string arguments in a CLI fashion

Instead of returning the value from main(), it prints the result to the stdout in a nicely formatted message.

Parameters:

args (List[str]) – command line parameters as list of strings (for example ["--verbose", "42"]).

csdigit.cli.parse_args(args: List[str]) → Namespace[source]

Parse command line parameters

Parameters:

args (List[str]) – command line parameters as list of strings (for example ["--help"]).

Returns:

command line parameters namespace

Return type:

argparse.Namespace

csdigit.cli.run() → None[source]

Calls main() passing the CLI arguments extracted from sys.argv

This function can be used as entry point to create console scripts with setuptools.

csdigit.cli.setup_logging(loglevel: int) → None[source]

Setup basic logging

Parameters:

loglevel (int) – minimum loglevel for emitting messages

csdigit.csd module

Canonical Signed Digit (CSD) conversion utilities.

Provides functions to convert between decimal numbers and Canonical Signed Digit (CSD) representation — a signed-digit representation using only +, -, and 0 symbols with no consecutive non-zero digits.

Functions:

to_csd: Convert a decimal float to CSD string with specified precision. to_csd_i: Convert an integer to CSD string. to_decimal: Convert a CSD string to a decimal number. to_decimal_using_pow: Convert a CSD string to decimal using pow() (deprecated). to_csdnnz: Convert a decimal float to CSD with limited non-zero digits. to_csdnnz_i: Convert an integer to CSD with limited non-zero digits.

csdigit.csd.to_csd(decimal_value: float, places: int) → str[source]

The to_csd function converts a given decimal number to its Canonical Signed Digit (CSD) representation with a specified number of decimal places.

Original author: Harnesser <https://sourceforge.net/projects/pycsd/> License: GPL2

Parameters:

  • decimal_value (float) – The decimal_value parameter is a double precision floating-point number that represents the value to be converted to CSD (Canonical Signed Digit) representation

  • places (int) – The places parameter in the to_csd function represents the number of decimal places to include in the CSD (Canonical Signed Digit) representation of the given decimal_value

Returns:

The function to_csd returns a string representation of the given decimal_value in Canonical Signed Digit (CSD) format.

Examples

>>> to_csd(28.5, 2)
'+00-00.+0'
>>> to_csd(-0.5, 2)
'0.-0'
>>> to_csd(0.0, 2)
'0.00'
>>> to_csd(0.0, 0)
'0.'
csdigit.csd.to_csd_i(decimal_value: int) → str[source]

The to_csd_i function converts a given integer into a Canonical Signed Digit (CSD) representation.

Original author: Harnesser <https://sourceforge.net/projects/pycsd/> License: GPL2

Parameters:

decimal_value (int) – The decimal_value parameter is an integer that represents the decimal value to be converted to CSD format

Returns:

The function to_csd_i returns a string containing the CSD (Canonical Signed Digit) value.

Examples

>>> to_csd_i(28)
'+00-00'
>>> to_csd_i(-0)
'0'
>>> to_csd_i(0)
'0'
csdigit.csd.to_csdnnz(decimal_value: float, nnz: int) → str[source]

The to_csdnnz function converts a given decimal number into a Canonical Signed Digit (CSD) representation with a specified number of non-zero digits.

Original author: Harnesser <https://sourceforge.net/projects/pycsd/> License: GPL2

Parameters:

  • decimal_value (float) – The decimal_value parameter is a double precision floating-point number that represents the input value for conversion to CSD (Canonic Signed Digit) fixed-point representation

  • nnz (int) – The parameter nnz stands for “number of non-zero bits”. It represents the maximum number of non-zero bits allowed in the output CSD (Canonical Signed Digit) representation of the given decimal_value

Returns:

The function to_csdnnz returns a string representation of the given decimal_value in Canonical Signed Digit (CSD) format.

Examples

>>> to_csdnnz(28.5, 4)
'+00-00.+'
>>> to_csdnnz(-0.5, 4)
'0.-'
>>> to_csdnnz(0.0, 4)
'0'
>>> to_csdnnz(0.5, 4)
'0.+'
csdigit.csd.to_csdnnz_i(decimal_value: int, nnz: int) → str[source]

The to_csdnnz_i function converts a given integer into a Canonical Signed Digit (CSD) representation with a specified number of non-zero digits.

Original author: Harnesser <https://sourceforge.net/projects/pycsd/> License: GPL2

Parameters:

  • decimal_value (int) – The decimal_value parameter is an integer that represents the decimal value to be converted to CSD format

  • nnz (int) – The parameter nnz stands for “number of non-zero bits”. It represents the maximum number of non-zero bits allowed in the output CSD (Canonical Signed Digit) representation of the given decimal_value

Returns:

The function to_csdnnz_i returns a string representation of the given decimal_value in Canonical Signed Digit (CSD) format.

Examples

>>> to_csdnnz_i(28, 4)
'+00-00'
>>> to_csdnnz_i(-0, 4)
'0'
>>> to_csdnnz_i(0, 4)
'0'
>>> to_csdnnz_i(37, 2)
'+00+00'
>>> to_csdnnz_i(158, 2)
'+0+00000'
csdigit.csd.to_decimal(csd: str) → float[source]

Convert CSD string to decimal number.

Original author: Harnesser <https://sourceforge.net/projects/pycsd/> License: GPL2

Parameters:

csd (str) – The csd parameter is a string containing the CSD (Canonical Signed Digit) value that we want to convert to a decimal number

Examples

>>> to_decimal("+00-00.+")
28.5
>>> to_decimal("0.-")
-0.5
>>> to_decimal("0")
0
>>> to_decimal("0.0")
0.0
>>> to_decimal("0.+")
0.5
csdigit.csd.to_decimal_using_pow(csd: str) → float[source]

The to_decimal_using_pow function converts a Canonical Signed Digit (CSD) string to a decimal number using the pow function.

Deprecated since version 0.1.0: Use to_decimal instead.

Original author: Harnesser <https://sourceforge.net/projects/pycsd/> License: GPL2

Parameters:

csd (str) – The csd parameter is a string containing the CSD (Canonical Signed Digit) value

Examples

>>> to_decimal_using_pow("+00-00.+")
28.5
>>> to_decimal_using_pow("0.-")
-0.5
>>> to_decimal_using_pow("0")
0.0
>>> to_decimal_using_pow("0.0")
0.0
>>> to_decimal_using_pow("0.+")
0.5

csdigit.csd_multiplier module

CSD Multiplier — Verilog Code Generation

Translates a Canonical Signed Digit (CSD) string into a synthesizable Verilog module that implements constant multiplication using only shift-and-add/subtract operations. When the CSD string contains repeated patterns, LCSRe (Longest Common Substring with Repeated Elements) is used to share hardware via a sub-expression wire.

csdigit.csd_multiplier.generate_csd_multiplier(csd: str, input_width: int, max_power: int) → str[source]

Generate Verilog code for a CSD multiplier module with LCSRe optimization.

When the CSD string contains repeated non-overlapping patterns (detected via the longest_repeated_substring algorithm), the generated Verilog factors out a shared sub-expression wire — reducing the number of adders.

Parameters:

  • csd – CSD string (e.g., “+00-00+0+”)

  • input_width – Bit width of the input signal x

  • max_power – Highest power of two in the CSD (must be len(csd)-1)

Returns:

Verilog module code as a string

Raises:

ValueError – If csd length doesn’t match max_power+1 or the string contains characters other than ‘+’, ‘-’, ‘0’

csdigit.csd_multiplier.generate_csd_multipliers(coeffs: list[tuple[str, str, int, int]], module_name: str = 'csd_filter') → str[source]

Generate a Verilog module with multiple CSD multipliers and cross-CSE.

When the same CSD substring appears in multiple coefficients, a shared sub-expression wire is created — reducing total adder count across the entire filter.

All coefficients must share the same input_width and max_power so that the same bit position encodes the same power of two across all multipliers. If a coefficient is narrower, pad its CSD with leading '0' characters.

Parameters:

  • coeffs – List of (output_name, csd_str, input_width, max_power) tuples. All entries must share the same input_width and max_power.

  • module_name – Name for the generated Verilog module.

Returns:

Verilog module code as a string

csdigit.lcsre module

Longest Repeated Non-Overlapping Substring (LCSRe) finder.

Finds the longest substring that appears at least twice in the input string without overlapping occurrences. Uses dynamic programming with a space-optimized 2-row DP table (O(n) memory instead of O(nu00b2)).

This is used internally by the CSD multiplier generator to detect repeated patterns in CSD strings, enabling hardware sharing via sub-expression wires.

Reference:

https://www.geeksforgeeks.org/longest-repeating-and-non-overlapping-substring/

csdigit.lcsre.longest_repeated_substring(csd_string: str) → str[source]

Longest repeated non-overlapping substring

The longest_repeated_substring function takes a string as input and returns the longest repeated substring in the string.

Parameters:

csd_string (str) – The parameter csd_string is a string containing the CSD value

Returns:

The function longest_repeated_substring returns a string, which is the longest repeated substring in the given input string csd_string.

Examples

>>> longest_repeated_substring("+-00+-00+-00+-0")
'+-00+-0'

Module contents