https://github.com/cheind/py-lapsolver
Fast linear assignment problem (LAP) solvers for Python based on c-extensions
https://github.com/cheind/py-lapsolver
linear-assignment-problem python solver
Last synced: 9 months ago
JSON representation
Fast linear assignment problem (LAP) solvers for Python based on c-extensions
- Host: GitHub
- URL: https://github.com/cheind/py-lapsolver
- Owner: cheind
- License: mit
- Created: 2018-02-07T05:42:25.000Z (about 8 years ago)
- Default Branch: master
- Last Pushed: 2024-11-29T07:44:14.000Z (about 1 year ago)
- Last Synced: 2025-05-16T05:03:45.860Z (9 months ago)
- Topics: linear-assignment-problem, python, solver
- Language: Python
- Homepage:
- Size: 214 KB
- Stars: 160
- Watchers: 3
- Forks: 24
- Open Issues: 3
-
Metadata Files:
- Readme: ReadMe.md
- License: LICENSE
Awesome Lists containing this project
README
## py-lapsolver
**py-lapsolver** implements a Linear sum Assignment Problem (LAP) solver for dense matrices based on shortest path augmentation in Python. In practice, it solves 5000x5000 problems in around 3 seconds.
### Install
```
pip install [--pre] lapsolver
```
Windows binary wheels are provided for Python 3.5/3.6. Source wheels otherwise.
### Install from source
Clone this repository
```
git clone --recursive https://github.com/cheind/py-lapsolver.git
```
Then build the project and exectute tests
```
python setup.py develop
python setup.py test
```
Executing the tests requires `pytest` and optionally `pytest-benchmark` for generating benchmarks.
### Usage
```python
import numpy as np
from lapsolver import solve_dense
costs = np.array([
[6, 9, 1],
[10, 3, 2],
[8, 7, 4.]
], dtype=np.float32)
rids, cids = solve_dense(costs)
for r,c in zip(rids, cids):
print(r,c) # Row/column pairings
"""
0 2
1 1
2 0
"""
```
You may also want to mark certain pairings impossible
```python
# Matrix with non-allowed pairings
costs = np.array([
[5, 9, np.nan],
[10, np.nan, 2],
[8, 7, 4.]]
)
rids, cids = solve_dense(costs)
for r,c in zip(rids, cids):
print(r,c) # Row/column pairings
"""
0 0
1 2
2 1
"""
```
### Benchmarks
Comparisons below are generated by scripts in [`./lapsolver/benchmarks`](./lapsolver/benchmarks).
Currently, the following solvers are tested
- `lapjv` - https://github.com/gatagat/lap
- `munkres` - http://software.clapper.org/munkres/
- `ortools` - https://github.com/google/or-tools **
- `scipy` - https://github.com/scipy/scipy/tree/master/scipy
- `lapsolver` - this project
**reduced performance due to costly dense matrix to graph conversion. If you know a better way, please let me know.
Please note that the x-axis is scaled logarithmically. Missing bars indicate excessive runtime or errors in returned result.
#### Additional Benchmarks
Berhane performs an in depth analysis of Python3 linear assignment problem solver at https://github.com/berhane/LAP-solvers
### References
**py-lapsolver** heavily relies on [code](https://github.com/jaehyunp/stanfordacm/blob/9e1375cd4eba68a59dd7b1e2f81692653e9908a9/code/MinCostMatching.cc) published by @jaehyunp at https://github.com/jaehyunp/