Run a Single Slice Basic Reconstruction#

Input file for This Example#

The NXtomo of this tutorial can be picked in /data/projects/tomo-sample-data/training/workspaces/[date]/[user_id]/PROCESSED_DATA/basic_reco/

Step 0: When Using ESRF Infrastructure#

Make sure you have a version of tomotools available:

module load tomotools

Step 1: Launch Tomwer Canvas#

Once on a cluster-node you can run:

tomwer canvas

[1]:

from tomwer.resources.video import Video

Video("rnsMClEMNcY", height=500, width=800)

[1]:

This tutorial focuses on datasets coming from BLISS. Therefore, we will only consider the NXtomo case.

However, be aware that processing can also be done with Spec-EDF, either natively or once the Spec-EDF has been converted to an NXtomo.

Step 2: Load the Workflow to Process Reconstructed Slices#

From the tomwer canvas top menu select the workflow already defined for reconstructing a slice locally:

Examples -> simple use cases -> simple slice reconstruction

[2]:

from tomwer.resources.video import Video

Video("RB2yF4s_jS4", height=500, width=800)

[2]:

This workflow contains the following widgets:

  • Scan Selector: scan selector

    Allows selection of one or multiple scans (either NXtomo or Spec-EDF).

  • Reduced Darks and Flats: reduced dark flat icon

    Computes the mean of dark frames and the median of flat frames to enable flat field normalization later in the workflow.

  • Center of Rotation: (see the ‘cor_search’ notebook for more information) center of rotation icon

    Computes the center of rotation (CoR) of the scan either manually or automatically.

  • Nabu Slice: nabu slice

    Used to reconstruct specific slices. You can specify reconstruction parameters here.

  • Data Viewer: data viewer

    Displays raw data and reconstructed slice(s).

Note: You can add a new widget from the left panel by left-clicking on the widget or by dragging a link from the node output to a downstream widget.

Step 3: Fine-Tune the Workflow#

  • Center of Rotation: Select the sliding-window algorithm and lock it lock button.

  • Nabu Slices: Add some Paganin phase retrieval.

    nabu phase

Step 4: Add the NXtomo to be processed#

Add the NXtomo contained in /data/projects/tomo-sample-data/training/workspaces/[date]/[user_id]/PROCESSED_DATA/basic_reco/ to be processed to the Scan selector either by:

  • clicking ‘Add’ from the interface

  • or drag and drop a file to the interface

  • or copy / paste some file path / scan identifier

scan selector

Note: The same works for other format like EDF/SPEC.

Step 5: Run the Workflow#

  • From the Scan Selector, select the scan and press send selected. This will trigger the workflow.

  • Wait for the different steps to be processed and check the outputs.

Step 6: Reconstruct a subset of the volume#

  • Add the nabu volume reconstruction to the canvas

  • Define the subset to be reconstructed

volume subset

  • Connect it to the workflow

    Connect the volume reconstruction to the workflow only once the subset is defined. Else you might reconstruct the entire volume.

  • (Optionaly) connect the volume viewer to the nabu volume viewer

Step 7: Save (and Share) the Workflow#

Once you are happy with your workflow, you can save it (Ctrl+S) and load it again later.

Note: The dataset used will also be saved, so this can be a good way to share your data processing with colleagues or report a bug that is easy to reproduce.

Step 8 (Optional): Launch the Same Processing on Slurm#

You can launch the same processing on Slurm by:

  • Adding a Cluster Configuration widget as the input to Nabu Slices.

  • Adding a Cluster Monitor widget as the output of Nabu Slices.

  • Retriggering the processing.

[3]:

from tomwer.resources.video import Video

Video("sxGarNwmGRE", height=500, width=800)

[3]:

[ ]:

Step 9 (Optional): Add a Widget to Convert from BLISS Acquisition to NXtomo#

9.1 add the nxtomomill-h52nx widget#

Currently, you only have a BLISS-HDF5 file that needs to be converted to the NXtomo format. To achieve this, you can use the nxtomomill h52nx tool.

Follow these steps:

  • Add a nxtomomill-h52nx widget. h52nx icon

  • Connect it to the Scan Selector.

[4]:

from tomwer.resources.video import Video

Video("LrlwwNUwR1g", height=500, width=800)

[4]:

9.2: Convert BLISS Raw Data#

  • Double-click on the nxtomomill h52nx widget.

  • Add the BLISS-HDF5 file: /data/projects/tomo-sample-data/training/workspaces/[date]/[user_id]/RAW_DATA/basic_reco/bambou_hercules_0001.h5.

    parameters → Custom output folder

    output folder → Custom output folder

  • Select the file(s) to be converted and press send selected.

Note: Usually, you will already have the NXtomo (e.g., from the data listener data listener) or you will process data from the RAW_DATA folder. But for this training you need to specify the output folder. An alternative is to use nxtomomill nx-copy CLI. For example:

nxtomomill nx-copy /data/projects/tomo-sample-data/training/part1_basic_reco/bambou_hercules_0001.nx [output_folder]

[ ]: