Avoid silent overwrite of existing spectrum/SRM FITS files

[Copilot]

Repeated processing of the same observation produces identical default filenames (stx_spectrum_<uid>.fits / stx_srm_<uid>.fits), silently clobbering any previously written files.

Changes

• New stx_unique_filename.pro — helper function that returns the input path unchanged if free, or finds the next available name by appending _1, _2, … before the extension. Extension-agnostic (uses last . in the path).

• stx_convert_science_data2ospex.pro — after the output paths are finalised in fits_info_params, both are resolved through stx_unique_filename. If a name changes, the updated path is written back to fits_info_params (so the FITS writer and OSPEX object stay in sync) and a message is printed unless silent is set.

Example behaviour

stx_spectrum_12345.fits        → already exists
stx_spectrum_12345_1.fits      → already exists
stx_spectrum_12345_2.fits      → free  ✓

; Console output:
; Spectrum file stx_spectrum_12345.fits already exists. Saving as: stx_spectrum_12345_2.fits
; SRM file stx_srm_12345.fits already exists. Saving as: stx_srm_12345_2.fits

[ennosigaeus]

You should be using the IDL routine "file_basename" to check for the name of the file and extension and not do that much manual work in stx_unique_filename. I would assume the entire routine become much shorter, so you may be able to integrate it with the main routine.

[Copilot]

You should be using the IDL routine "file_basename" to check for the name of the file and extension and not do that much manual work in stx_unique_filename. I would assume the entire routine become mu...

Refactored in 82e1881. stx_unique_filename now uses file_basename(filename, '.fits') and file_dirname to split path/base/extension, removing all the manual strpos/strmid work. The active code body is down to ~10 lines.

[Copilot]

Pull request overview

This PR prevents repeated processing runs from silently overwriting previously written spectrum/SRM FITS outputs by introducing a helper that selects the next available filename and applying it during OSPEX conversion.

Changes:

• Added stx_unique_filename.pro to find an available output filename by appending an incrementing _<N> suffix before the extension.
• Updated stx_convert_science_data2ospex.pro to run the finalized spectrum/SRM output paths through stx_unique_filename, updating fits_info_params and optionally printing a notice (unless silent is set).

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 1 comment.

File	Description
stix/idl/processing/spectrogram/stx_unique_filename.pro	New helper function to pick a non-existing output filename by applying a numeric suffix.
stix/idl/processing/spectrogram/stx_convert_science_data2ospex.pro	Uses stx_unique_filename to avoid clobbering existing spectrum/SRM outputs and keeps fits_info_params in sync.

Comments suppressed due to low confidence (1)

stix/idl/processing/spectrogram/stx_unique_filename.pro:56

• stx_unique_filename always appends a new _<counter> suffix to the full base name. If the input already ends with an existing numeric suffix (e.g. foo_1.fits) and that file exists, this will generate foo_1_1.fits rather than incrementing to foo_2.fits as described in the header examples. Consider detecting a trailing _<int> in the base name and incrementing from that value (or stripping it before applying the counter).

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[ennosigaeus]

I like the approach, but the stx_unique_filename is in the wrong place. Maybe better in a more generic folder, like https://github.com/i4Ds/STIX-GSW/tree/master/stix/idl/util

[Copilot]

I like the approach, but the stx_unique_filename is in the wrong place. Maybe better in a more generic folder, like https://github.com/i4Ds/STIX-GSW/tree/master/stix/idl/util

Moved to stix/idl/util/stx_unique_filename.pro in 6062f82.

[ennosigaeus]

LGTM, already gave feedback to copilot.

[ennosigaeus]

@paolomassa can you have a quick look and spin in IDL to see if this is fine?