https://github.com/evanwporter/cloth
Sloth + C++ = Cloth
https://github.com/evanwporter/cloth
Last synced: about 1 month ago
JSON representation
Sloth + C++ = Cloth
- Host: GitHub
- URL: https://github.com/evanwporter/cloth
- Owner: evanwporter
- License: mit
- Created: 2024-06-16T20:05:06.000Z (10 months ago)
- Default Branch: main
- Last Pushed: 2024-12-30T20:23:55.000Z (4 months ago)
- Last Synced: 2025-01-21T10:07:30.720Z (3 months ago)
- Language: C++
- Size: 2.43 MB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE.txt
Awesome Lists containing this project
README
# Cloth
Cloth is a high performance C++ implementation of the Pandas API.
In exchange for a fraction of the features of Pandas, Cloth has a fraction of the execution time and a fraction of the resources of Pandas.
To obtain such speeds Cloth is implemented using Eigen to hold the underlying data and Nanobinds for the fast C++ -> Python bindings. Note that because the data is stored within Eigen, it must be homogenous. This is akin to the `DataMatrix` in very early versions of Pandas.
Syntax within Cloth are implemented and binded in such a way that is (nearly) identical to Pandas. Check out tests/test_cloth.py for an example of how to use Cloth in Python.
The biggest advantage to using Cloth as opposed to Pandas, is the `TimeSeries` and `TimeFrame` datatypes. These are specific implementations of the `Series` and `DataFrame` meant for Financial Data over time. Specifically these implementations offer a fast method of performing decimal calculations--without the inaccuracies brought on by floating point numbers (a necessity when dealing with money)--via the [decimal](https://github.com/vpiotr/decimal_for_cpp) datatype.
# Frames
* `cloth.Series`
* `cloth.DataFrame`
# Locational Indexing
* `cloth.Frame.loc[string]`
* `cloth.Frame.loc[slice]`
* `cloth.Frame.loc[cloth.slice]`
# Integer Indexing
* `cloth.Frame.iloc[int]`
* `cloth.Frame.iloc[slice]`
* `cloth.Frame.iloc[cloth.slice]`
# Boolean Indexing
* `cloth.Frame.where(boolean array)`
* `boolean array -> cloth.Series > value`
# Column Indexing
* `cloth.DataFrame[string]`
# DateTime
* `cloth.datetime(ISO string)`
# Timedelta
* `cloth.timedelta(interval)`
# Resampling
* `cloth.frame.resample(timedelta)`