indicio

indicio is a python package that provides offline access to the refractiveindex.info database of optical constants. Contrary to other similar packages, the full upstream database ships inside the package, so there is no need to download it separately: just pip install indicio and you are ready to go.

indicio has zero runtime dependency, not even YAML because the database is stored internally in a custom sqlite format that python can decode natively. This also reduces the size of the database to ~16MB.

The API returns structured descriptions of dispersion models. It does not implement formulas to compute n(λ) or k(λ). That job is left to the consumer, who can pick whatever numerical stack suits the use case (numpy, jax, torch, etc.).

indicio exists primarily for integration in Torch Lens Maker, but you are welcome to use it for anything else!

Installation

pip install indicio

Quickstart

import indicio

SiO2 = indicio.get_material("main", "SiO2", "Malitson")
print(SiO2.n)

>>> Sellmeier(c1=0.0, coefficients=((0.6961662769317627, 0.06840430200099945), (0.40794259309768677, 0.11624140292406082), (0.8974794149398804, 9.896161079406738)), wavelength_range=(0.21, 6.7))

API

indicio.get_material(shelf: str, book: str, page: str) -> MaterialEntry
indicio.has_material(shelf: str, book: str, page: str) -> bool
indicio.shelves() -> tuple[str, ...]
indicio.books(shelf: str) -> tuple[str, ...]
indicio.pages(shelf: str, book: str) -> tuple[str, ...]
indicio.iter_materials() -> Iterator[MaterialEntry]
indicio.search(query: str) -> tuple[tuple[str, str, str], ...]

indicio.__version__               # library version, e.g. "1.0.0"
indicio.__database_version__      # upstream commit hash baked into the data

The data model

@dataclass(frozen=True)
class MaterialEntry:
    shelf: str
    book: str
    page: str
    name: str | None
    references: str | None
    comments: str | None
    n: NModel | None              # real part — None if upstream has only k
    k: KModel | None              # imaginary part — None if upstream has only n

• For the dispersion model n, there are 10 possible cases: 9 closed-form formulas (following the upstream dispersion-formula doc), or tabulated data.

• For the extinction coefficient k, the data is always tabulated.

For detailed data description of each model see src/indicio/models.py.

Examples

The examples/ directory contains complete evaluation scripts covering every formula type:

• examples/example_evaluation_stdlib.py: uses only the standard library (math, array, bisect).
• examples/example_evaluation_numpy.py: uses numpy; evaluators are vectorized over a wavelength array.
• examples/example_plot.py: uses numpy + matplotlib to plot n(λ) and k(λ) over the validity range.

python examples/example_evaluation_stdlib.py main SiO2 Malitson
python examples/example_evaluation_numpy.py main Au Johnson
python examples/example_plot.py main BaF2 Bosomworth-300K

Conventions

• Wavelengths are micrometers everywhere, matching the convention of refractiveindex.info.
• Tabulated arrays are raw bytes, packed as little-endian IEEE 754 float32. See examples above for decoding.

Licensing and attribution

The refractiveindex.info database is distributed under CC0 1.0 (public domain). When publishing results that rely on a particular material entry, please cite the original authors. Their citation strings are preserved verbatim in MaterialEntry.references.

print(indicio.get_material("main", "SiO2", "Malitson").references)

The indicio library code itself is distributed under its own LICENSE in the source tree.