Table Of Contents

This Page

Welcome to fitsblender’s documentation!

fitsblender.fitsblender(headers, spec)

Given a list of FITS headers, aggregates values and creates a table made up of values from a number of headers, according to the given specification.


  • headers is a sequence where each element is either:

    • a pyfits.Header instance
    • a 2-tuple of the form (filename, extension), where filename is a path to a FITS file and extension is an extension number.
  • spec is a list defining which keyword arguments are to be aggregated and how. Each element in the list should be a sequence with 2 to 5 elements of the form:

    (src_keyword, dst_name, function, error_type, error_value)

    • src_keyword is the keyword to pull values from. It is case-insensitive.

    • dst_name is the name to use as a dictionary key or column name for the destination values.

    • function (optional). If function is not None, the values from the source are aggregated and returned in the aggregate_dict. If function is None (or the tuple contains only 2 elements), all values are stored as a column with the name dst_name in the result table.

      If not None, function should be a callable object that takes a sequence of values and returns an aggregate result. If the function returns None, no values will be added to the aggregate dictionary. There are many functions in Numpy that are directly useful as an aggregating function, for example:

      Lambda functions are also often useful:

      • first: lambda x: x[0]
      • last: lambda x: x[-1]

      Additionally, function may be a tuple, where each member is itself a callable object. The result will be a tuple containing results from each of the given functions. For instance, to aggregate a range of values, i.e. both the minimum and maximum values, use the following as function: (numpy.min, numpy.max).

    • error_type (optional) defines how missing or syntax-errored values are handled. It may be one of the following:

      • ‘ignore’: missing or unparsable values are ignored. They are not included in the list of values passed to the aggregating function. In the result table, missing values are masked out.
      • ‘raise’: missing or unparsable values raise a ValueError exception.
      • ‘constant’: missing or unparsable values are replaced with a constant, given by the error_value field.
    • error_value (optional) is the constant value to be used for missing or unparsable values when error_type is set to ‘constant’. When not provided, it defaults to NaN.


A 2-tuple of the form (aggregate_dict, table) where:

  • aggregate_dict is a dictionary of where the keys come from dst_name and the values are the aggregated values as run through function.
  • table is a masked Numpy structured array where the column names come from dst_name and the column contains the values from src_keyword for all of the given headers. Missing values are masked out.

Indices and tables