Running GreenFuelBreak_Pipeline.ipynb (miniconda + JupyterLab)

0. What you need

Your project folder (26X_GFB_Data/) should contain:

26X_GFB_Data/
├── GreenFuelBreak_Pipeline.ipynb   <- the notebook you run
├── environment.yml                 <- conda environment spec (provided)
├── inputs/                         <- all input data (road_segments, NKSK-boundaries,
│                                       plant-shapefiles, hawaii_dem, Penman/AET rasters, xlsx)
└── notebooks/                      <- the 5 original notebooks (not needed to run)

The notebook auto-detects inputs/ when you launch Jupyter from this folder, and it also has your ~/Desktop/26X_GFB_Data/inputs path hard-coded as a fallback, so either way it will find the data.

---

1. Open a terminal and confirm conda works

conda --version

If "command not found", open the Anaconda Prompt (Windows) or run source ~/miniconda3/bin/activate (macOS/Linux) first.

2. Create the environment (one time)

From inside the project folder:

cd ~/Desktop/26X_GFB_Data
conda env create -f environment.yml

This makes a conda environment named gfb with the full geospatial stack (GDAL/PROJ come bundled via conda-forge, so there is nothing else to compile). It takes a few minutes.

Prefer to build it by hand instead of the file? This is the equivalent:

conda create -n gfb -c conda-forge python=3.11 geopandas rasterio shapely pyproj \
  contextily mapclassify pandas numpy matplotlib openpyxl requests beautifulsoup4 \
  jupyterlab ipykernel

3. Activate it

conda activate gfb

Your prompt should now start with (gfb).

4. Register the environment as a Jupyter kernel (one time)

python -m ipykernel install --user --name gfb --display-name "Python (gfb)"

5. Launch JupyterLab from the project folder

cd ~/Desktop/26X_GFB_Data
jupyter lab

This opens JupyterLab in your browser.

6. Open the notebook and pick the kernel

1. In the file browser, open GreenFuelBreak_Pipeline.ipynb.
2. Top-right kernel selector -> choose Python (gfb).

7. Run it

• Run everything: menu Run -> Run All Cells, or run cell by cell with Shift+Enter.
• Step 0 prints an "Input check" table. Every row should say OK. If any says MISSING, fix the path (see Troubleshooting) before continuing.

That's it. Generated files (scored shapefile, CWD rasters/tables, watering CSV, cost CSV, and all maps) are written to a new gfb_outputs/ folder inside the project folder.

---

Good to know

• Step 1 (species scraper) is optional and normally skips itself. It only runs if inputs/gonative_combined_scored_3_13.xlsx is missing. You do not need internet for the rest of the pipeline.
• Map basemaps need internet. The Esri shaded-relief topo layer is fetched online. With no connection the maps still render (border, roads, colors), just without the topographic backdrop.
• The first code cell (0a) runs pip install .... Inside the gfb env everything is already present, so pip just reports "already satisfied" and changes nothing. You can comment that cell out if you prefer.
• Re-running: outputs overwrite cleanly. Safe to Run All again.

Changing where inputs/outputs live

Open the master config cell (Step 0b) and set the paths directly, e.g.:

INPUT_DIR  = "/Users/garyding/Desktop/26X_GFB_Data/inputs"
OUTPUT_DIR = "/Users/garyding/Desktop/26X_GFB_Data/gfb_outputs"

Troubleshooting

• Input check shows MISSING — you launched Jupyter from a different folder. Either cd into 26X_GFB_Data before jupyter lab, or hard-set INPUT_DIR in Step 0b.
• Kernel "Python (gfb)" not listed — re-run step 4, then refresh JupyterLab.
• conda env create fails on solving — update conda first: conda update -n base conda, then retry. Make sure you are on the conda-forge channel (the yml already sets this).
• Want to start over — conda env remove -n gfb and repeat from step 2.