PubDAS support

[d-chambers]

A few notes on supporting PubDAS formats:

For the original 6-8 data sets:

• We need SEGY support. From our meeting notes @aaronjgirard is going to work on this using segyio which we will have to take as a new dependency (or write our own segy reader, yuck)

• FORSEE will be a little tricky to support. Looking at the files in PubDAS there is no metadata attached. Just the raw (pre-processed from TDMS file) data and timestamp arrays. The other necessary information is found in its readme file. I suppose we could hardcode these values and make a special "FORESEE" parser, but to be safe we can't support automatic format detection. We don't want another file with "raw" and "timestamp" to have the hardcoded gauge-length and such applied.

• Valencia is also a bit odd, but might be do-able. This is the HDF5 structure of a random file I selected:

Unfortunately, there is no easy signature (eg a string that says "FEBUS" as a root attribute). I am also not sure if sub-folders will always start with "fa1" or not. Does anyone have a contact at Febus we could ask?

For the DAS data datasets

I haven't gone through all of these yet but a few observations:

• There are several miniseed-based datasets. We could support this if we use ObsPy (or PyRocko) as an optional dependency.