MStoh5

[sensingcosmos]

HI Ludwig,
I find a script which can convert the .h5 files to MS to do some processing on a MeerKat data. I now need some script which can convert back from MS to .h5 files. Is there some script like this?

Cheers,
Abhik

[ludwigschwardt]

Unfortunately this is currently a one-way process...

[sensingcosmos]

Hi Ludwig, Thanks. Are you working with (or you know some people in SKA office who are) MeerKat single dish data ? I have just started to look into few data sets (1472018997.h5, 1474580092.h5, 1474580580.h5) few days ago, I am mainly interested in initially to study the noise properties and the calibration issue for future Intensity mapping problem. Is there some automatic flagging software available to flag the TOD data (Time Vs Frequency) for MeerKat autocorrelation? Thanks for your help. Cheers, Abhik

[ludwigschwardt]

Yes I am familiar with the single-dish data. We have a script (katsdpscripts/RTS/RFI_Report/rfi_report.py I think) that can flag autocorrs (although it is tuned for crosscorrs). Ask @mauch about it. Another option is to look in the archive on the page where you download the h5 file. Click on "Show Reductions" and then "Show Files" next to "Weather and RFI Report". There you'll find an h5 file containing flags that could be applied to the autocorrs.

[AlecThomson]

Hi @ludwigschwardt,

Just jumping on this thread - is there any scope for producing a mstoh5 script? This would be particularly useful for integrating with @mauch 's https://github.com/mauch/KATObitPipe pipeline.

[ludwigschwardt]

That's an interesting idea! The main challenge is to provide a ChunkStore interface to CASA tables (might be easier in MSv3). Parallel reads would be nice but not absolutely necessary since you are typically working off a directory on a single disk and the disk read speed would be the limiting factor.

If you still have the original RDB file lying around, that could serve up the metadata in the usual way.

Irregular timestamps and/or baselines might require some extra thought.