How to run a single slice basic reconstruction

Raw data for this example

Raw data for this example can be found in /scisoft/tomo_training/part1_basic_reco You must copy them, and you can also remove the .nx file contained here (they will be created during step 1).

Step 0: Launch tomwer canvas

When you are on the cluster node after sourcing the Python virtual environment, launch the canvas:

tomwer canvas

from IPython.display import Video

Video("video/start_tomwer_canvas.mp4", embed=True, height=500)

There are two use cases for now:

• From a Bliss acquisition. Then we will convert on step 1 from Bliss acquisition to NXtomo.

• From EDF. In this case, you can use the native EDF folder as a NXtomo or convert it to NXtomo first (looking at XXX for example) and then proceed to step 2.

Step 1: From Bliss acquisition to NXtomo

For this, we need to use the bliss(HDF5) - nxtomomill alias h52nx-nxtomomill widget.

• Add the widget from the left panel with a click on the left panel item or by copy/pasting.

• Open the widget dialog by double-clicking on the canvas widget freshly created.

• Then add the Bliss file from the select button or copy/paste to the dedicated dialog.

• You should see one line added to the dialog per NXtomo that will be created.

• (Optional: Connect it with another widget like data list or data selector to see the output and make sure all the conversion went well.)

• Press ‘Send All’ / ‘Send Selected’ to trigger the conversion to Bliss file (some other dialog might appear if the NXtomo already exists).

Note 1: Keep in mind that this can also be done manually using the command line interface (CLI) of nxtomomill. For more details, please seeh52nx nxtomomill tutorial.

Note 2: For converting from EDF to NXtomo, you can do the same operation but using the ``edf2nx-nxtomomill`` widget. Have a look atEDF to NXtomo CLI tutorialfor CLI or advanced usage.

Note 3: If no configuration file is provided, then the default parameters will be used. Otherwise, you can provide a configuration file. For details, see theedf2nx widget video tutorialwhich explains how to provide a file (GUI mechanism is the same for EDF and HDF5).

from IPython.display import Video

Video("video/h52nx_example.mp4", embed=True, height=500)

Hands-on Exercise A

• Copy data from `to a local workspace (like/tmp_14_days/{your_name}`) or reuse them if they already exist.

• Launch tomwer canvas.

• Convert the Bliss .h5 file to a NXtomo (.nx) using the appropriate widget.

Step 2: Define a Basic Workflow

From raw data we will need:

• Reduced dark/flat widget

  NXtomo is expected to contain dark and flat frames. In order to compute the flat field, we need to compute reduced darks and flats.

• Center of rotation (see ‘cor_search’ notebook for more information).

  For the training, we will use the ‘sino-coarse-to-fine’ algorithm which provides a good estimation of the center of rotation. In the video, we will do it manually, but we could lock the algorithm to avoid validating the value found.

• Nabu ‘slice’

  To reconstruct one slice. In the example, we will also request a Paganin phase to have a better reconstruction.

• Data viewer

  To display the reconstructed slice (and browse the dataset).

Note: The widget can be created from the left panel (with a left mouse click on the widget) or by creating a link from a left click on the node output and releasing it downstream).

from IPython.display import Video

Video("video/create_dummy_workflow.mp4", embed=True, height=500)

Step 3: Run the Workflow

Now that you have an input and a basic workflow, we can process it.

For this, simply ‘select’ the NXtomo from the scan selector that was created during step 1. Processing should start; you have to wait until all processes are finished.

from IPython.display import Video

Video("video/execute_dummy_workflow.mp4", embed=True, height=500)

Note: You can control advancement from the ‘Object Supervisor’ on the bottom of the window. If it is not visible, you can display/hide it from the view/object supervisor option as shown in the video.

from IPython.display import Video

Video("video/object_supervisor_display.mp4", embed=True, height=500)

Hands-on Exercise B

Reconstruct a slice of the NXtomo created during Exercise A.

Save and Share the Workflow

Once you are happy with your workflow, you can save it (Ctrl+S) and you will be able to load it next time.

Note: Datasets used will also be saved, so it can be a good way to share data processing with colleagues or report a bug easily reproducible.