Quickstart

To install and run the vNext pipeline, follow the quickstart guide below.

Installation

The vNext pipeline runs in a dedicated conda environment (kpfpipe, Python 3.14). Clone the repository, create the environment, and install the package in editable mode:

git clone -b kpf-next https://github.com/Keck-DataReductionPipelines/KPF-Pipeline.git
cd KPF-Pipeline
conda env create -f environment.yml
conda activate kpfpipe
pip install -e .

Running the DRP

The easiest way to run the pipeline is via the command line interface (CLI).

The kpfpipe command has two main entry points: masters builds nightly calibration products (bias, dark, flat, wavelength solution), and science reduces science exposures into RVs. Each takes input/output data directories and a datecode (masters) or obs_id (science) to act on.

Input/output directories specify the root directory for the data files:

Input L0 data files should be placed in {input_dir}/L0/
Reduced L1/L2/L4 science files will output to {output_dir}/L{n}/
Masters files will output to {output_dir}/masters/
Logs will output to {output_dir}/logs/
Quicklook plots will output to {output_dir}/QLP

Masters

Build master calibrations for one or more nights (identified by a YYYYMMDD datecode) directly in the CLI:

kpfpipe masters \
    --input_dir /path/to/input/ \
    --ouput_dir /path/to/output/ \
    --dates 20240405 20250912 20250111

Or by passing a plain text file which lists a single datecode per line:

kpfpipe masters \
    --input_dir /path/to/input/ \
    --ouput_dir /path/to/output \
    --dates /path/to/datecodes.txt

You can also generate masters over a range of nights:

kpfpipe masters \
    --input_dir /path/to/input/ \
    --ouput_dir /path/to/output/ \
    --date_range 20240405 20240418

Science

Reduce one or more science frames (identified by their obs_ids):

kpfpipe science \
    --input_dir /path/to/input/ \
    --ouput_dir /path/to/output/ \
    --obs_ids KP.20240405.40113.57 KP.20240912.84491.73

Or by passing a plain text file which lists a single obs_id per line:

kpfpipe science \
    --input_dir /path/to/input/ \
    --ouput_dir /path/to/output/ \
    --obs_ids /path/to/obs_ids.txt

Make sure that --output_dir specifies the same directory you used to run the masters pipeline so the science pipeline knows where to look for masters calibrations.

Timeseries

To conveniently process all masters and science for a single star over a given date range, use the timeseries command:

kpfpipe timeseries \
    --target 10700 \
    --input_dir /path/to/input/ \
    --ouput_dir /path/to/output/ \
    --date_range 20240405 20240418