Metadata-Version: 2.0
Name: s11-classifier
Version: 2.4.2
Summary: Classifier
Home-page: https://gitlab.com/satelligence/classifier
Author: Rens Masselink
Author-email: rens.masselink@satelligence.com
License: Apache-2.0
Platform: UNKNOWN
Requires-Python: >=3.5
Description-Content-Type: text/markdown
Requires-Dist: Fiona (<1.9.0,>=1.8.3)
Requires-Dist: boto3 (<1.10.0,>=1.9.4)
Requires-Dist: click (<6.8.0,>=6.7)
Requires-Dist: folium (<0.7.0,>=0.6.0)
Requires-Dist: geojson (<2.5.0,>=2.4.0)
Requires-Dist: geopandas (<0.5.0,>=0.4.0)
Requires-Dist: matplotlib (<2.1.0,>=2.0.0)
Requires-Dist: numpy (<1.14.0,>=1.13.1)
Requires-Dist: pandas (<0.23.0,>=0.22.0)
Requires-Dist: pylint (<1.9.0,>=1.8.2)
Requires-Dist: pytest (<5.3.0,>=5.2.0)
Requires-Dist: rasterio (<1.1.0,>=1.0.11)
Requires-Dist: rasterstats (<0.14.0,>=0.13.0)
Requires-Dist: rtree (<0.9.0,>=0.8.3)
Requires-Dist: scikit-learn (<0.22.0,>=0.21.3)
Requires-Dist: xgboost (<0.82,>=0.81)

# Classifier

A scalable land cover classification tool for humans

Go to [the documentation](https://satelligence.gitlab.io/classifier) for more
 info on how to install and run Classifier!

## Note on performance

For now, the multithreading of XGboost is not properly implemented. If you have
a lot of input files the XGBoost method is MUCH slower than e.g. scikit-learn's
random forest. With single input files (so not a lot of IO operations), the
difference in performance is small.

## issues and bugs

Issues are tracked on the [issue list of this
repo](https://gitlab.com/satelligence/classifier/issues).

## Development

Please follow the [Satelligence development guidelines](https://gitlab.com/satelligence/workflow/blob/master/dev_workflow.md)
when adding features or fixing bugs.

(Fork) Clone this repo:

```sh
git clone git@gitlab.com:satelligence/classifier.git
```

Make sure you are running the latest docker image.

Run docker with a binding to the source files:

```sh
docker run  -v $(pwd):/app -t -i classifier
```

