O2K Near-IR Reduction Pipeline

This folder contains the Python-based O2K imaging pipeline plus a few helper tools for registration fixes and WCS solving.

Main workflow

From the top-level O2000_rex folder, reduce one observing night with:

python reduce.py YYYYMMDD

This does three things:

1. reads the raw FITS files in YYYYMMDD
2. organizes them into the MMDD reduction folder and SN subfolders
3. runs the full Python reduction and then the HTML/JPG QA step

Example:

cd /Users/lluisgalbany/Desktop/O2000_rex
python reduce.py 20260504

The reduced products are written into the MMDD folder, for example 0504.

What the pipeline does

The default Python engine replaces the old cl < analyze.cl IRAF step. It performs:

• dark and flat creation
• bad-pixel correction using bpm2012.fits
• dark and flat correction of science frames
• sky-frame construction and subtraction
• negative outlier replacement
• simple cosmic-ray cleaning
• image registration
• final min/max-rejected combination
• header EXPTIME update
• QA image and HTML generation

Optional flags

Skip the QA preview step:

python reduce.py YYYYMMDD --skip-check

Use the legacy IRAF engine instead of the Python reducer:

python reduce.py YYYYMMDD --engine cl

This only works on a machine with IRAF cl available.

Registration rescue tools

If the automatic registration is not good enough, these helpers can be used on one SN folder at a time.

Measure residual registration quality:

python qc_registration.py /path/to/MMDD/SNNAME H
python qc_registration.py /path/to/MMDD/SNNAME J

Measure manual shifts by clicking the same target in each frame:

python click_register.py /path/to/MMDD/SNNAME H --prefix cr

Apply those manual shifts, rebuild xobj*, and recombine the final image:

python apply_manual_shifts.py /path/to/MMDD/SNNAME H

If the xobj* frames are already acceptable and you only want to recombine after editing listobj*, use:

python recombine_registered.py /path/to/MMDD/SNNAME H

WCS step

After reduction, copy the final science images you want to solve into finalimages/, then run:

cd /Users/lluisgalbany/Desktop/O2000_rex/finalimages
python solve_wcs.py SN2022ea_20260504_H

The script:

• extracts the SN name from the filename
• queries TNS for the SN coordinates
• matches catalog and image sources interactively
• writes a solved FITS file named <input>_wcs.fits

If needed, you can bypass TNS and pass coordinates directly:

python solve_wcs.py SN2022ea_20260504_H --ra 208.173538708 --dec 59.1578780404

Required files kept in this folder

• reduce.py
• organize.py
• python_reduction.py
• check.py
• qc_registration.py
• click_register.py
• apply_manual_shifts.py
• recombine_registered.py
• bpm2012.fits
• finalimages/solve_wcs.py

Python dependencies

The pipeline expects a Python environment with the packages used by the scripts, including:

• numpy
• scipy
• scikit-image
• astropy
• photutils
• matplotlib
• astroquery
• requests

Notes

• The reduction creates fresh calibration and intermediate products inside each night folder, so old top-level generated files are not needed.
• OBJECT names like SN2023_vcx are normalized automatically during organization.
• If a run stops midway, rerunning reduce.py starts that night again from the organized reduction stage rather than doing a full checkpoint resume.