Welcome to pySpecData’s documentation!¶
pySpecData allows you to deal with multi-dimensional spectroscopy data in an object-oriented fashion.
Please note this package is heavily utilized by three other packages that our lab manages on github:
See pySpecData examples here, and for a simple example applied to 1D data, see here.
Some more explanation¶
In practice, this means that it makes the code for processing spectral data shorter and more quickly legible. It automatically handles the following issues, without any additional code:
relabeling axes after a Fourier transformation
propagation of errors
adding units to plots
To do this, you work with a pySpecData nddata object (which includes information about dimension names, axis values, errors, and the units) rather than working directly with traditional numpy ndarray objects. (pySpecData is built on top of numpy.)
If you have ever worked with arrays in Matlab or Python before, you are familiar with the additional code needed to convert between index numbers and axis values. Using the funny notation of pySpecData, you can do this automatically.
For example, say you have loaded an nddata object called d wanted to take time-domain data, Fourier transform, select out the central 20 kHz in the frequency domain, and then inverse Fourier transform, while preserving the correct axes throughout. That looks like this:
>>> d.ft('t2', shift=True)
>>> d['t2':(-10e3,10e-3)]
>>> d.ift('t2')
Note that most pySpecData methods operate in-place on the data; it modifies the data inside the nddata, rather than making a new copy. This is because we assume that we are progressively optimizing/filtering our spectral data with each new line of code. This allows us to quickly work through many operations (like the ft here) without keeping many copies of a large dataset and with less typing for each operation. (If you every truly want to create a copy a dataset, just attach a .C)
Todo
describe updates allowing aliasing (nu2 vs. t2, etc.)
Note that
>>> plot(d)
Quickly generates a publication-quality plot; it automatically plots the data on the correct axes and includes the units and the name of dimension (t2 in this example) and its units along the x axis.
How do I generate an nddata object?¶
We primarily work in magnetic resonance,
so have written wrappers for a few different types of NMR
(nuclear magnetic resonance) and ESR (electron spin resonance)
file formats.
You can use the pyspecdata.find_file() function to automatically load them as nddata.
Additionally, we have written several classes that allow you to read nddata objects directly from e.g. an oscilloscope. These are available as separate repositories on github (please contact us for further info).
Finally, you can easily build nddata from standard arrays, as discussed in the section about nddata objects.
What can I do with nddata objects?¶
To understand how to manipulate nddata objects, head over to the section about nddata objects.
Contents:¶
There are many other features of pySpecData that govern the interaction of ndata objects with plots and, e.g. that allow you to generate a nice PDF laboratory notebook showing every step of your data processing. These and further details are covered in the various sections of the documentation:
Example Gallery