notes

.github/workflows/compute.yml

@@ -16,6 +16,8 @@ jobs:
         uses: actions/checkout@v3
       - name: compute
         run: python fire.py
+      - name: git-lfs
+        run: apt-get update && apt-get -y install git-lfs
       - name: push
         env:
           HF_TOKEN: ${{ secrets.HF_TOKEN }}

.gitpod.yml

@@ -0,0 +1,8 @@
+
+image:  ghcr.io/boettiger-lab/solara-geospatial:latest
+
+vscode:
+  extensions:
+    - ms-toolsai.jupyter
+    - ms-python.python
+    - ms-toolsai.jupyter-renderers

README.md

@@ -9,4 +9,19 @@ license: mit
 app_port: 8765
 ---
 
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+NOTE: The YAML header above is required for HuggingFace Spaces deployment.  Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+
+# Simple Solara demo
+
+This respository demonstrates how we can deploy an interactive map tool from a few lines of python to an app on HuggingFace spaces using Solara. 
+
+[Solara App on HuggingFace](https://huggingface.co/spaces/cboettig/solara-test)
+
+### Overview
+
+This application accesses boundary polygons from the National Parks Service and fire polygon data from CalFire to determine the location of all recorded fires in Joshua Tree national park.  We select the largest area fire since 2015 in the database (currently turns out to be Elk Trail Fire) and access all Sentinel-2 imagery from the two weeks before and after the fire alarm date. From this imagery, we compute the Normalized Burn Severity metric (NBS) around the fire polygon before and after the fire (using cloud-native approach of `pystac`, `odc.stac`, and dask), and plot this on a leaflet map overlay with splitmap and fire polygons. 
+
+
+### Code
+
+The required code for analysis is in `fire.py`, with solara visualization in `pages/01_leafmap.py`.  (For interactive )