ZTF forced-photometry light curves of nuclear transient candidates

ZTF forced-photometry light curves in g and r, and in i where available, for 12,251 nuclear transient candidates, spanning 2018 March to 2024 September. The preliminary quality cuts described in Masci et al. (2023) have been applied, followed by Galactic extinction correction and baseline correction. Two catalogue tables give positions, extinction coefficients, classifications where available, and classifier probabilities for the unclassified sources. Released with Anilkumar et al. (2026). Contents Light curves are distributed as four ZIP archives. Each expands to a directory of CSV files named <ZTF_ID>.csv agn_sources.zip — 5,113 AGN sne_sources.zip — 1,042 Supernovae tde_sources.zip — 66 TDEs unknown_sources.zip — 6,030 Unclassified sources classified_sources.csv — Catalogue of the 6,221 classified sources unknown_sources.csv — Catalogue of the 6,030 unclassified sources The two tables are disjoint and cover all 12,251 sources. Light curve files Each file starts with a '#' header block giving the ZTF ID, TNS name, RA/Dec (J2000, degrees), classification with the DOI of its source catalogue, extinction coefficients (A_r, A_g, A_i) and classifier probabilities. The data table follows: JD [days] — Julian Date at start of exposure flux [μJy] — extinction- and baseline-corrected flux flux_unc [μJy] — uncertainty on `flux`, baseline uncertainty propagated filter — g.ztf, r.ztf or i.ztf flux_org [μJy] — extinction-corrected flux before baseline correction flux_unc_org [μJy] — uncertainty on `flux_org` zpdiff [mag]— photometric zeropoint of the difference image Catalogue tables classified_sources.csv — ZTF_ID, TNS_Name, RA, DEC, classification, A_r, A_g, A_i, Bibcode. Bibcode gives the DOI of the catalogue the label came from, comma-separated where a source appears in more than one Supernova labels keep their published subtypes rather than being collapsed to a single class unknown_sources.csv — ZTF_ID, TNS_Name, RA, DEC, A_r, A_g, A_i and six probability columns Notes Class probabilities are given only for sources without a classification, i.e. in unknown_sources.csv and the headers of unknown_sources.zip. Of the 6,030 unclassified sources, 5,367 have probabilities. The remaining 663 failed feature extraction and were omitted at the prediction stage; their fields are empty in the CSV and shown as '-' in the file headers. SN and TDE labels come from the ZTF Bright Transient Survey (Perley et al. 2020; Fremling et al. 2020), the Transient Name Server, and the manyTDE repository (Mummery et al. 2024). AGN labels come from Liu et al. (2019), the WISE AGN catalogue (Assef et al. 2018), and the Milliquas catalogue (Flesch 2023), and are catalogue-based rather than spectroscopic classifications of the transient itself. The TDE archive holds 66 light curves. Seven are included for completeness but were not used in the accompanying paper: five repeating TDE candidates (TDE2022exr, TDE2021mhg, TDE2020acka, AT2018mac, TDE2021uqv) and two events falling below the 30-epoch cut (TDE2019cmw, TDE2018hyz). The unclassified sample is not a class. It is the set of candidates for which no classification was found. Selection, cuts, corrections and the classifiers are described in Anilkumar et al. (2026). Citation If you make use of this dataset, please cite both this Zenodo record and the paper it accompanies, Anilkumar et al. (2026), in which the sample, the corrections applied to the light curves, and the classifiers are described in full. Anilkumar et al. (2026) References - Assef, R. J., et al. 2018, ApJS, 234, 23- Flesch, E. W. 2023, OJAp, 6, 49- Fremling, C., et al. 2020, ApJ, 895, 32- Liu, H.-Y., et al. 2019, ApJS, 243, 21- Masci, F. J., et al. 2023, arXiv:2305.16279- Mummery, A., et al. 2024, MNRAS, 527, 2452- Perley, D. A., et al. 2020, ApJ, 904, 35- Transient Name Server: https://www.wis-tns.org/- manyTDE: https://github.com/sjoertvv/manyTDE