Skip to content

Quickstart

This Quickstart gives you two ways to run the pipeline end-to-end:

  1. CLI path – run cscal-pipeline from a terminal
  2. Notebook path – run the pipeline inside Jupyter

Both produce the same ENVI, Parquet, and QA artifacts.

Install

Install from PyPI:

pip install cross-sensor-cal

For Ray support:

pip install "cross-sensor-cal[ray]"

1. CLI path

Use the CLI if you run jobs on your laptop, server, or HPC cluster.

Run a NEON flight line

Choose an output directory:

BASE=output_quickstart
mkdir -p "$BASE"

Run the pipeline:

cscal-pipeline \
  --base-folder "$BASE" \
  --site-code NIWO \
  --year-month 2023-08 \
  --product-code DP1.30006.001 \
  --flight-lines NEON_D13_NIWO_DP1_L020-1_20230815_directional_reflectance \
  --max-workers 2 \
  --engine thread

The first time this runs, it will:

  • download NEON HDF5 tiles
  • export ENVI cubes
  • apply topographic + BRDF correction
  • convolve to Landsat-style reflectance
  • write Parquet tables
  • produce QA PNG, PDF, and JSON summaries

If rerun, completed stages are skipped safely.

Inspect QA files (example on macOS)

open $BASE/NEON_D13_NIWO_DP1_L020-1_20230815_directional_reflectance/*_qa.png
open $BASE/NEON_D13_NIWO_DP1_L020-1_20230815_directional_reflectance/*_qa.pdf

2. Notebook path (Jupyter)

Use this if you want an interactive, reproducible workflow.

Run the pipeline from Python

In a notebook cell:

from cross_sensor_cal import go_forth_and_multiply

base = "output_quickstart_py"

go_forth_and_multiply(
    base_folder=base,
    site_code="NIWO",
    year_month="2023-08",
    product_code="DP1.30006.001",
    flight_lines=["NEON_D13_NIWO_DP1_L020-1_20230815_directional_reflectance"],
    max_workers=2,
    engine="thread",
)

Preview the merged Parquet table

import duckdb, os

fl = "NEON_D13_NIWO_DP1_L020-1_20230815_directional_reflectance"
merged = os.path.join(base, fl, f"{fl}_merged_pixel_extraction.parquet")

duckdb.query(f"SELECT * FROM '{merged}' LIMIT 5").df()

For a more complete notebook-style walkthrough, see:
Usage → Jupyter notebook example

Next steps