ASCII notebooks

[mdmueller]

These will be used for documentation in astropy/astropy#2716.

[taldcroft]

@amras1 - for consistency with the astropy hierarchy I would suggest putting these under io/ascii instead of just ascii.

[mdmueller]

@taldcroft - That makes more sense, I fixed it in the last commit.

[taldcroft]

@amras1 - the notebooks look OK. The only thing that would be good is to put a title and description / summary of each notebook at the top. E.g.

Compare table read speed for io.ascii, Pandas, and np.genfromtxt

This notebook provides speed comparisons for reading ASCII data tables using a few different methods:

• io.ascii.read using the legacy pure-Python reader
• io.ascii.read using the fast C-based reader with the standard string-to-double converter
• io.ascii.read using the fast C-based reader with a faster string-to-double converter
• pandas.read_csv
• numpy.genfromtxt

The data tables themselves are space-delimited tables with variations in length, number of columns, and composition (how many columns of each data type: int, float, str).

Summary

• The io.ascii fast-C reader with standard conversion is about 3 to 7 times faster than the pure-Python version.
• Using the fast string-to-double converter makes io.ascii reading comparable in speed to Pandas and sometimes faster.

[mdmueller]

@taldcroft - Good idea! The format looks great, I just used your summary above and added similar introductions for the other notebooks.