📄 Description of the file final_candidates_catalogue_2026_clean_with_suffixes.csv This catalogue contains transient candidate detections identified in the Chandra X-ray archive using a custom pipeline (including the novel CuDiDi method). Each row corresponds to one candidate event. The candidates are named using the convention: XT YYMMDD[a-z],where YYMMDD represents the date of the observation, and the optional lowercase letter (a, b, c, …) distinguishes multiple candidates from the same observation. The table includes the following columns: 🔢 General Columns CANDIDATE:Name of the candidate in the format XT YYMMDD, with an optional alphabetical suffix (e.g., XT 230615b) if multiple candidates were found in the same observation. OBSID:Chandra Observation ID in which the candidate was found. SOURCE:Internal source number within the observation (from source detection). RA / DEC:Right Ascension and Declination of the candidate (J2000 coordinates). 🚩 Diagnostic Flags (Boolean) These flags help assess the reliability of each candidate and identify potential artefacts or known sources. A value of True indicates that the corresponding condition applies: non_unique:True if the transient is matched with a source detected at low significance in other Chandra observations (sources detected at >3σ in more than one observation are already excluded). snr_match:True if the candidate falls within an observation known to contain a Galactic supernova remnant (e.g., Cassiopeia A, Tycho), based on cross-matching with SNR catalogs. extended_fov:True if the source lies within or near an extended structure (e.g., diffuse Galactic emission) identified by manual inspection of the field of view. readout_streak:True if the source overlaps with a known instrumental readout streak, often caused by bright sources like Cyg X-1. edge_affected:True if the detected source region is only partially covered by the detector due to dithering, which may result in unreliable flux or light curve measurements. bkg_correlated:True if the source light curve is strongly correlated (Pearson coefficient > 0.15) with the background during the T90 interval, indicating likely background flaring. start_or_end:True if the start (T5) or end (T95) of the transient lies within the first or last 5% of the observation. Such events may be truncated by the observational window. Of 1420 candidates, 765 have no flags set to true and are considered the 'golden sample'.