https://github.com/teleprint-me/float_is_close.c
Compare floating-point numbers with a given tolerance in C
https://github.com/teleprint-me/float_is_close.c
c floating-point python
Last synced: 3 months ago
JSON representation
Compare floating-point numbers with a given tolerance in C
- Host: GitHub
- URL: https://github.com/teleprint-me/float_is_close.c
- Owner: teleprint-me
- License: agpl-3.0
- Created: 2024-08-18T18:40:00.000Z (9 months ago)
- Default Branch: main
- Last Pushed: 2024-08-20T00:44:44.000Z (9 months ago)
- Last Synced: 2025-01-03T10:45:48.189Z (4 months ago)
- Topics: c, floating-point, python
- Language: C
- Homepage:
- Size: 81.1 KB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# float_is_close.c
**Compare floating-point numbers with a given tolerance in C**
## Overview
`float_is_close.c` is a C library designed to accurately compare floating-point numbers with a configurable tolerance. It provides functions to compare both single-precision and double-precision floating-point numbers. This library is intended to be lightweight, efficient, and easy to integrate into existing C projects.
## Directory Structure
```plaintext
.
├── auto-build -> auto_build.sh
├── auto_build.sh
├── CMakeLists.txt
├── float_is_close.py
├── include
│ └── float_is_close.h
├── LICENSE
├── README.md
├── src
│ └── float_is_close.c
└── tests
└── test_float_is_close.c
```
- **`auto-build`**: A symbolic link to `auto_build.sh`, which automates the build and test process.
- **`auto_build.sh`**: Shell script for automated build and testing.
- **`CMakeLists.txt`**: CMake configuration file for building the project.
- **`float_is_close.py`**: Python proof of concept or utility script.
- **`include/float_is_close.h`**: Header file containing the public API.
- **`LICENSE`**: The project is licensed under the AGPL.
- **`README.md`**: This file.
- **`src/float_is_close.c`**: Source file implementing the floating-point comparison functions.
- **`tests/test_float_is_close.c`**: Test cases for the implemented functions.
## Building the Project
To build the project, run the `auto_build.sh` script. This script automates the cleaning, building, and testing processes.
### Steps:
1. Clone the repository:
```sh
git clone https://github.com/teleprint-me/float_is_close.c float-is-close
cd float-is-close
```
2. Run the build script:
```sh
./auto-build
```
This will configure and build the project using CMake, compile the source files, and run the tests.
### Build Output
After building, the structure of the `build/` directory will look like this:
```plaintext
build/
├── bin
│ └── test_float_is_close
├── lib
│ └── libfloat_is_close.so -> libfloat_is_close.so.0.1.0
└── ...
```
- **`build/bin/`**: Contains the test executable.
- **`build/lib/`**: Contains the shared library file.
## API Documentation
The primary API functions are defined in `include/float_is_close.h`.
### `double_is_close`
```c
bool double_is_close(double a, double b, size_t significand);
```
- **Parameters**:
- `a`, `b`: The double-precision floating-point numbers to compare.
- `significand`: The number of significant digits to consider (1-15).
- **Returns**: `true` if the numbers are close within the specified tolerance, otherwise `false`.
### `float_is_close`
```c
bool float_is_close(float a, float b, size_t significand);
```
- **Parameters**:
- `a`, `b`: The single-precision floating-point numbers to compare.
- `significand`: The number of significant digits to consider (1-7).
- **Returns**: `true` if the numbers are close within the specified tolerance, otherwise `false`.
### Macros
- **`FIC_MIN(a, b)`**: Returns the minimum of two values.
- **`FIC_MAX(a, b)`**: Returns the maximum of two values.
- **`FIC_CLAMP(value, lower, upper)`**: Clamps a value within a specified range.
## References
### Mathematical Background:
- [Calculus](https://math.libretexts.org/Bookshelves/Calculus/Calculus_3e_(Apex))
- [Discrete Mathematics](https://math.libretexts.org/Bookshelves/Combinatorics_and_Discrete_Mathematics/Discrete_Mathematics_(Levin))
### Original Papers:
- [APL comparison tolerance](https://dl.acm.org/doi/10.1145/800114.803685)
- [PEP-0485](https://peps.python.org/pep-0485/)
### Original Source Code:
- [Christopher Barker's Proof of Concept](https://github.com/PythonCHB/close_pep/blob/master/is_close.py)
## License
This project is licensed under the AGPL License - see the [LICENSE](LICENSE) file for details.