How to construct model without downloading

[pllim]

Follow up of #1 and #6 -- What if I don't want to download from SVO? What if I already have a sampled transmission curve locally but I want to use your package to turn it into a parametrize Astropy model?

[bmorris3]

This is the section of the source where that magic happens:

tynt/tynt/download.py

Lines 143 to 162 in 04e0b7c

	diff_wl = np.diff(wl)
	delta_lambda = np.nanmedian(diff_wl[diff_wl != 0])
	lambda_0 = wl.min()
	n_lambda = len(wl)
	# Create a simplified wavelength grid:
	simplified_wavelength = np.arange(lambda_0, (n_lambda + 1) *
	delta_lambda + lambda_0,
	delta_lambda)
	tr_max = tr.max()
	# Interpolate transmittance onto simplified wavelength grid:
	tr_interp = np.interp(simplified_wavelength, wl, tr)
	# Take the DFT of the interpolated transmittance curve
	fft = np.fft.fft(tr_interp)[:n_terms]
	# Save results in a dictionary
	row = [n_lambda, lambda_0, delta_lambda, tr_max] + fft.tolist()

I agree that it would be handy to split this out of the DownloadManager object and into a standalone function, so users can tinker with it. PRs are welcome!