search

alisterburt / tomo_preprocessing

Other
4 stars 1 forks source link

Issue writing aligned_tilt_series.star #40

Closed EuanPyle closed 2 years ago

EuanPyle commented 2 years ago

When running wrapper outside of relion (ie on the command line with relion_tomo_align_tilt_series) a aligned_tilt_series.star is not written. With latest update to tomo_model, aligned_tiltseries.star is written (mostly correctly) but is titled data as opposed to data_global. Not sure if this is an issue on our end or rln end

alisterburt commented 2 years ago

any ideas why?

this is the conditional for when that func is run https://github.com/alisterburt/tomo_preprocessing/blob/273787ce7901a1259704424fe02c2ae959d8d8d9/tomography_preprocessing/tilt_series_alignment/aretomo/batch.py#L68-L73

that runs this https://github.com/alisterburt/tomo_preprocessing/blob/273787ce7901a1259704424fe02c2ae959d8d8d9/tomography_preprocessing/tilt_series_alignment/aretomo/_utils.py#L6-L19

which should actually run https://github.com/alisterburt/tomo_preprocessing/blob/273787ce7901a1259704424fe02c2ae959d8d8d9/tomography_preprocessing/tilt_series_alignment/_job_utils.py#L91-L114

So we should update there...

data_ as opposed to data_global. Not sure if this is an issue on our end or rln end

This is an issue on the rln end -> let's update that ourselves :)

EuanPyle commented 2 years ago

Woops, re-writing whole comment, I'm out of it today. The error must be in tomo_preprocessing/tomography_preprocessing/tilt_series_alignment/_job_utils.py as it's not writing an aligned_tilt_series.star output even if I do IMOD:patch-tracking. Weirdly, it works fine when I use RELION so I'm mostly just confused

EuanPyle commented 2 years ago

Found the problem finally with our wrapper not generating aligned_tilt_series.star. Had a bit of an unsuccessful day, so very happy with this small win!

alisterburt commented 2 years ago

what was the issue?!

RELION writes that file itself actually which means there is some irritating decoupling of where things are defined vs where they are interpreted - I made our version a CLI program that could be called at the end but sjors seemed to prefer keeping that in the cpp stuff 🤷🏼‍♂️

We should probably remove this from our code and make ours capable only of aligning a single ts at a time