Custom Templates¶
Add your own spectra to the template library for classification.
Requirements¶
Data Quality¶
| Requirement | Minimum | Recommended |
|---|---|---|
| Signal-to-Noise | S/N ≥ 10 | S/N ≥ 20 |
| Wavelength Coverage | 70% of 3500–9000 Å | 90% |
| Spectral Resolution | R ≥ 500 | R ≥ 1000 |
Metadata¶
Required fields: - Name: Object designation (e.g., "SN2024abc") - Type: Main classification (Ia, Ib, Ic, II, SLSN, etc.) - Subtype: Detailed classification (norm, 91bg, IIP, etc.) - Age: Days from maximum light - Redshift: Host redshift
Creating Templates¶
GUI Method¶
- Open Template Manager (Tools → Template Manager)
- Go to the Create tab
- Load spectrum file (supports .dat, .txt, .csv, .fits)
- Fill metadata:
- Name, Type, Subtype
- Age (days from max)
- Redshift
- Preview the processed template
- Save to your User Templates folder
CLI Method¶
# Single template
sage templates import-csv data\spectrum.csv --dest User_templates\ `
--name-column object_name --path-column spectrum_file_path `
--type-column type --subtype-column subtype `
--age-column age --redshift-column redshift
Batch Import¶
Import many templates at once from a CSV/TSV file.
CSV Format¶
| object_name | spectrum_file_path | type | subtype | age | redshift |
|---|---|---|---|---|---|
| SN2024abc | spectra/sn2024abc_day5.dat | Ia | norm | 5.0 | 0.023 |
| SN2024abc | spectra/sn2024abc_day10.dat | Ia | norm | 10.0 | 0.023 |
| SN2024def | spectra/sn2024def.dat | II | IIP | 15.0 | 0.015 |
- Multiple rows with the same
object_namecreate a multi-epoch template - Paths can be relative to the CSV file location
- Column names are case-insensitive
GUI Batch Import¶
- Template Manager → Manage tab
- Click Batch Import
- Select your CSV file
- Choose destination folder
- Review and confirm
CLI Batch Import¶
sage templates import-csv data\my_templates.csv `
--dest C:\User_templates `
--name-column object_name `
--path-column spectrum_file_path `
--type-column type `
--subtype-column subtype `
--age-column age `
--redshift-column redshift
File Formats¶
Supported spectrum formats:
- .dat, .txt, .ascii - Whitespace-separated (wavelength, flux)
- .csv - Comma-separated with optional headers
- .fits - Standard FITS format
Wavelength must be in Angstroms. Flux units are normalized internally.
User Templates Location¶
Templates you create are stored in your User Templates folder:
User_templates/
├── template_index.user.json # Your template index
├── templates_Ia.user.hdf5 # Your Type Ia templates
├── templates_II.user.hdf5 # Your Type II templates
└── ...
Set or change this folder: - Template Manager → left panel → Change User Folder - Settings → Paths → User Templates Directory
Override Behavior¶
When you have a user HDF5 for a type (e.g., templates_Ia.user.hdf5), it replaces the built-in templates for that type during analysis. This ensures your custom templates are used.
Best Practices¶
- Use high S/N spectra - Better templates give better classifications
- Accurate metadata - Correct type, subtype, and age are essential
- Complete wavelength coverage - Gaps reduce correlation quality
- Multiple epochs - Time series improve age constraints
- Version control - Keep your User Templates folder backed up
See Also¶
- Template Library - Built-in template bank
- Templates Manager (GUI) - GUI details
- Supported Formats - Input file formats