Skip to main content

Automated property interpolation and color calculations for main sequence stars

Project description

Build Status Coverage Status astropy PyPI version Requirements Status


This code provides functionality for automating calculations based on Eric Mamajek's "A Modern Mean Dwarf Stellar Color and Effective Temperature Sequence" (

The purpose of the code is to allow users to automatically interpolate the property columns of the table, and to calcualte colors based on any combination of the photometry columns.

Also provides utilities for matching a wide variety of spectral type strings.

Installation and Requirements

MeanStars requires the following python modules:

  • numpy
  • scipy
  • astropy

To install, simply run:

pip install MeanStars

Alternatively, grab a copy of the repository and run:

pip install .

To install in developer (editable) mode, grab a copy of the repository and run:

pip install -e .

This last option is useful if you want to update the data file.


To use MeanStars, you must first create a MeanStars object:

from MeanStars import MeanStars
ms = MeanStars()

This object contains a number of useful attributes, including:

  • The full dataset in astropy table format
  • ms.bands: The names of all unique bands from the dataset (in a string array)
  • ms.colors: All of the original colors from the dataset (in an nx2 string array, where n is the length of ms.bands and the color is the first column minus the second). The same information is also encoded in ms.colorstr, which is the original color name from the dataset
  • ms.noncolors: All data attributes not related to color (these include stellar masses, radii, etc.)
  • ms.SpecTypes: All of the major spectral types from the dataset (in a string array - nominally this will always be O,B,A,F,G,K,M,L,T,Y)
  • ms.colorgraph: A directed graph (encoded as a dictionary) mapping all relationships between bands established by the colors in the original dataset

Interpolating Colors

MeanStars provides two methods for interpolating colors: TeffColor and SpTColor, where the former interpolates by effective temperature, and the latter by spectral type. In general, it is unlikely that you will want to query the data by any spectral type not explicilty listed, so the SpTColor most frequently acts as a simple lookup table of the data. In each case, the methods are called by providing the two bands defining the color (called start and end in the code such that the color is start - end), and the temperature or spectral type.

So, to find the 'U-B' color of a 29000 K star, you would execute:


This particular instance corresponds to an exact entry in the data table, and the value returned should exactly match the entry. A more interesting case is the 'U-H' color of a 6000 K star:


Here, you are requesting a color not found in the table at a temperature not found in the table (but bracketed by other temperature values). You can query to find the specific sets of color combined to give this result by running:


which will return ['U', 'B', 'V', 'Ks', 'H'], meaning that the 'U-B','B-V','V-Ks', and 'Ks-H' colors were added to get the result.

Interpolating by spectral type works exactly the same way, excpet that the type is defined by two input variables representing the major and minor subtype. So, to find the 'U-H' color of a G2V dwarf, you would execute:


Each time one of these routines is called on a new color (for a given object instance), the generated interpolant is saved in ms.Teffinterps and ms.SpTinterps (as appropriate to the method call). This means that the interpolant is generated just once per object instance, speeding up subsequent computations.

Interpolating Other Properties

Just as with colors, any other property in the original data set can be interpolated as a function of effective temperature or spectral type, via methods TeffOther and SpTOther, respectively. The methods use the same syntax as their color counterparts, save that the property is defined by a single string input.

So, to find the mean solar mass of a 5500 K star, you would run:


and to find the mean radius of a K9V star, you would run:


Looking Up Nearest Data Entries

For any data columns with monotonically varying data (such that a nearest-neighbor interpolant can be defined), the nearest data entry can be identified via the tableLookup method. For example, to see which spectral type has closest log(solar luminosity) to 3, we can execute:

ind = ms.tableLookup('logL', 3.0)

which returns:

SpT logL
--- ----
B3V 2.99

Matching Spectral Types

MeanStars provides a regular expression and a utility method for parsing spectral type strings. These are intended to work with Morgan-Keenan (MK, MKK, or Yerkes) style spectral type strings containing a Spectral Class (and Subclass) and (optionally) a Luminosity Class. The default matching behavior (provided by attribute specregex) is to Letter|number|roman numeral where the letter is the spectral class (one of OBAFKGMLTY), the number is the subclass (nominally between 0, inclusive, and 10, exclusive), and the roman numeral is the luminosity type (nominally I through VII). The number can be an integer or have a decimal, there can be parentheses around the number and/or roman numeral, and there can be spaces in between any or all of the values. The luminosity class (roman numeral) and subtype (number) can be two values separated by a slash or dash. Examples of valid strings that can be matched by this regular expression include: G0V, G(0)V, G(0)(V), G0.5V, G5/6V, G(5/6)(IV/V), G 0.5 (V), G 0.5V. Both the subclass and luminosity class are treated as optional, and so we can also match things like G and G0.

The second supported case (using the same attribute) is where the type string contains multiple spectral types, spearated by a slash or dash. All of the same formatting is supported as in the previous case, and examples of valid matches include things like G8/K0IV. In these cases, as many elements as possible are extracted from both classes. That is, the regular expression always returns 6 matching groups in cases where any match is possible. In cases where only one class string is present, the second three groups are all None.

This regular expressions is wrapped by class method matchSpecType, which will attempt to return a single value for the spectral class, subclass, and luminosity class for a given string. It will also look for leading white dwarf ('D', 'WD', or 'wd') and subdwarf ('sd') patterns and set the luminosity class accordingly.

The method takes a single string input:


will return: ('G', 0.0, 'V').

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

MeanStars-3.4.0.tar.gz (32.2 kB view hashes)

Uploaded source

Built Distribution

MeanStars-3.4.0-py3-none-any.whl (32.4 kB view hashes)

Uploaded py3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page