https://github.com/developmentseed/label-maker-dask

Library for running label-maker as a dask job
https://github.com/developmentseed/label-maker-dask

dask machine-learning microsoft osm

Last synced: about 1 month ago
JSON representation

Library for running label-maker as a dask job

• Host: GitHub
• URL: https://github.com/developmentseed/label-maker-dask
• Owner: developmentseed
• License: mit
• Created: 2021-08-31T13:25:28.000Z (over 3 years ago)
• Default Branch: main
• Last Pushed: 2022-05-06T15:37:06.000Z (over 2 years ago)
• Last Synced: 2024-10-14T09:40:47.374Z (2 months ago)
• Topics: dask, machine-learning, microsoft, osm
• Language: Jupyter Notebook
• Homepage:
• Size: 85.9 KB
• Stars: 7
• Watchers: 4
• Forks: 3
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# label-maker-dask

Library for running [label-maker](https://github.com/developmentseed/label-maker/) as a [dask](https://dask.org/) job

# Acknowledgements

This work was undertaken with support from Microsoft to be run on the [Planetary Computer](https://planetarycomputer.microsoft.com/). With access to the Planetary Computer Hub, you can find an interactive notebook tutorial for running this library.

# Basic Example

Instantiate a distributed dask cluster

```python

from dask.distributed import Client

cluster = ...

client = Client(cluster)

```

Create a label maker job

```python

from label_maker_dask import LabelMakerJob

lmj = LabelMakerJob(

    zoom=13,

    bounds=[-44.4836425781, -23.02665962797, -43.412719726, -22.5856399016],

    classes=[

        { "name": "Roads", "filter": ["has", "highway"] },

        { "name": "Buildings", "filter": ["has", "building"] }

      ],

    imagery="http://a.tiles.mapbox.com/v4/mapbox.satellite/{z}/{x}/{y}.jpg?access_token=ACCESS_TOKEN",

    ml_type="segmentation",

    label_source="https://qa-tiles-server-dev.ds.io/services/z17/tiles/{z}/{x}/{y}.pbf"

)

```

Build & execute the job

```python

lmj.build_job()

lmj.execute_job()

```

View or otherwise use the results (by passing to a machine learning framework)

```python

for result in lmj.results:

    ...

```