VIPCALs: a fully automated calibration pipeline for VLBI data

VIPCALs A fully automated calibration pipeline for VLBI data Diego Álvarez Ortega Institute of Astrophysics – FORTH, Crete, Greece 10th International VLBI Technology Workshop 23rd October, Gothenburg, Sweden Motivation 23rd October, 2025 10th IVTW, Gothenburg, Sweden 2 Current pipelines ●rPICARD ●VLBARUN ●EVN pipeline VLBI Data calibration ●Time consuming ●Expertise-heavy Bottleneck for projects involving large samples Still require human supervision Not suitable for all datasets SMILE: Search for Milli-Lenses SMILE Sample Archival + New VLBA Data 23rd October, 2025 10th IVTW, Gothenburg, Sweden 3 ●4968 radio loud AGNs ●Mainly Cand X-band ●~ 6000 files, > 600 projects ●> 60 TB of data ●Observed for over 30 years Need for full-automation! (or many graduate students) Carolina Casadio VIPCALs +ParselTongue GUI built with Pyside6 ●Fully automated calibration - no prior knowledge on the dataset ●Quick, robust, reproducible ●Runs on local or 23rd October, 2025 10th IVTW, Gothenburg, Sweden 4 Pre-release version available in Github: https://github.com/dalvarezo/VIPCALs/ Docs Calibration workflow 23rd October, 2025 10th IVTW, Gothenburg, Sweden 5 Pre-loading Loading and pre-calibration Amplitude & Phase calibration ●Sanity checks ●Split multiple frequencies ●Calibrator selection ●Retrieve missing tables ●Average if necessary ●Remove bad Tsys ●Choose reference antenna ●Typical AIPS VLBI workflow ●Includes ionospheric corrections ●Final fringe fit on target OR phase reference calibrator Some automation features 23rd October, 2025 10th IVTW, Gothenburg, Sweden 6 ●Pre-selection of instrumental calibrators ○Based on VLBA calibrator list fluxes VLBA Calibrator List Some automation features 23rd October, 2025 10th IVTW, Gothenburg, Sweden 6 ●Pre-selection of instrumental calibrators ○Based on VLBA calibrator list fluxes ●Retrieval of calibration tables ○For VLBA and EVN experiments Example of ANTAB file Some automation features 23rd October, 2025 10th IVTW, Gothenburg, Sweden 6 ●Pre-selection of instrumental calibrators ○Based on VLBA calibrator list fluxes ●Retrieval of calibration tables ○For VLBA and EVN experiments ●Selection of reference antenna ○Based on coverage and SNR Some automation features 23rd October, 2025 10th IVTW, Gothenburg, Sweden 6 ●Pre-selection of instrumental calibrators ○Based on VLBA calibrator list fluxes ●Retrieval of calibration tables ○For VLBA and EVN experiments ●Selection of reference antenna ○Based on coverage and SNR ●Selection of instrumental calibrators ○Baseline-based selection based on SNR Current Limitations 23rd October, 2025 10th IVTW, Gothenburg, Sweden 12 ●Non-standard or missing tables (20) ○Require manual editing ●Non-ordered IF setups (13) ●Subarrays (10) ○Only in some cases ●Wrong order information in header (2) All of them can be solved with small adjustments to the workflow Failed calibrations (45/1417) ●Total intensity - No polarization ●Continuum ●cm-wavelength ●VLBA experiments ○Code needs fine-tuning for other arrays Tested on SMILE data Future Work 23rd October, 2025 10th IVTW, Gothenburg, Sweden 13 ●Adjust and test on other arrays (EVN, VGOS) ○Different metadata, subarrays Future Work 23rd October, 2025 10th IVTW, Gothenburg, Sweden 13 Bandpass artifact ●Adjust and test on other arrays (EVN, VGOS) ○Different metadata, subarrays ●Automatic flagging of obvious bad data ○In frequency Future Work 23rd October, 2025 10th IVTW, Gothenburg, Sweden 13 VLA was auto-phasing ●Adjust and test on other arrays (EVN, VGOS) ○Different metadata, subarrays ●Automatic flagging of obvious bad data ○In frequency ○In time Future Work 23rd October, 2025 10th IVTW, Gothenburg, Sweden 13 ●Adjust and test on other arrays (EVN, VGOS) ○Different metadata, subarrays ●Automatic flagging of obvious bad data ○In frequency ○In time ●Implement opacity corrections when only weather data are available ○Opacity measurements can be fitted using Tsys and Tatm ○Important at higher frequencies (> 22 GHz) Future Work 23rd October, 2025 10th IVTW, Gothenburg, Sweden 14 Demo Avinash Kumar Felix Pötzl Alejandro Mus Imaging ●Automatic, robust, and efficient ●Multiple algorithms (CLEAN, ehtim, resolve) ●Simulating realistic data Summary 23rd October, 2025 10th IVTW, Gothenburg, Sweden 15 ●Current pipelines are not scalable to large, heterogeneous samples ○SMILE requires fully-automated calibration ●VIPCALs delivers end-to-end data calibration with minimum input ○Calibration is quick, robust, and reproducible ●Successfully tested on VLBA data ○Current version suitable for SMILE: full intensity, continuum, cm-wavelength ●Future work ○Extend to other VLBI arrays ○Implement new features https://vipcals.readthedocs.io/ https://github.com/dalvarezo/VIPCALs/ Documentation: Alvarez-Ortega et al. (2025) https://arxiv.org/abs/2508.13282 Science paper: Extra Slides 23rd October, 2025 10th IVTW, Gothenburg, Sweden - Fringe Fit Results 23rd October, 2025 10th IVTW, Gothenburg, Sweden - ●On average, FRING gives 89% of solutions with SNR > 5 ●Low fringe ratios are related to low fluxes ○Fluxes that are not accurate, as they come from VLA observations ●Manual inspection of data with low fringe ratio points towards non-detections Fringe Fit Threshold 23rd October, 2025 10th IVTW, Gothenburg, Sweden - A. R. Thompson, J. M. Moran, and G. W. Swenson Jr., Interferometry and Synthesis in Radio Astronomy