Direct reading of msr files

[manerotoni]

Hello,

I have some code to read directly from the msr files saved by Imspector (see https://github.com/manerotoni/minflux-analysis). This uses the specpy library shipped with imspector. Would it be interesting to include such a feature in pyMinflux?
I could work on a draft branch for it.

Greetings

Antonio Politi (MPI-NAT Goettingen)

[aarpon]

Cool, thanks!

In pyMINFLUX 0.5.0 we read confocal images and metadata from the *.msr files using a reader we developed from scratch, and in branch feature/reader_v2 (for the upcoming version 0.6.0), we added support for importing .npy, .mat, and .json files in the new Inspector m2410 format. We are now looking into importing information from the exported zarr structures.

It would be interesting to have a chat and discuss what you can contribute with your code: please reach out here: pyminflux at ethz dot ch and we can then talk directly.

[manerotoni]

Hello,
@aarpon yes we can have a chat for detail. I just add here some relevant comments.

The reasons I started to directly import zarr data structures (this is in fact what you get when using specpy) from the msr using python was:

• Directly read the data into python without conversion in imspector.
• Access the bead MBM reference coordinates that is/was otherwise hidden from the user. This was necessary when performing exchange paint and we restart the recording at each exchange. I could correct several recordings by using the same reference beads.
• During acquisition data is stored in a zarr temp directory. This is what is accessed by paraFlux. Reading the zarr would allow to do some analysis live during acquisition.

The data format you get from reading directly the msr is in fact a zarr. This is quite fast in accessing data.

Ideally I would like to use a preprocessing tool such as pyMinflux that has different importing options. I am now looking into the new data format and checking what needs to be changed.

[aarpon]

Tendentially, we would like to avoid using SpecPy and use the zarr library directly to access the relevant data; initially, we'd focus on the bead MBM reference coordinates (post acquisition).

[manerotoni]

Hello @aarpon,
I was just testing some Zarr export from the Abberior software. For the Abberior software 16.3.21317.m2410, the zarr entries for the mbm are empty, only .zattrs and .zgroups. I think this is a bug from their side. Do you see the same behavior?

If I do export directly from SpecPy I get the complete archive with mbm.

Greetings

Antonio

[javicasarias]

Hi @manerotoni ,
For iMSPECTOR-Exported zarr files after 2410, the actual mbm data is stored under zarr/grd/mbm/points. The .zattrs and .zgroups under zarr/mbm only specify the name given to the beads within the iMSPECTOR UI (R followed by a number).

Best,
Javier