Frap script create figure

[will-moore]

This adds OMERO.figure creation to a new simple_frap_with_figure.py script.

Time indexes are currently hard-coded (don't span whole movie) and figure has scalebars and timestamp labels.

I've also updated both simple_frap scripts to use roi_service.getShapeStatsRestricted() instead of getting numpy data (as in Basel day-2 walkthrough).
Also the matplotlib -> PIL image method is changed simply to allow it to run on a Mac. This creates a temp file which isn't deleted (will be cleaned-up when the temp script folder is cleaned up).

To test, use RCC1-FRAP images at http://web-dev-merge.openmicroscopy.org/webclient/?show=dataset-25602 (user-3). These are my own images, so these can be made public.
The jpeg plots come from the Deltavision's own analysis - can be used to tell which image (of same name) gives a good FRAP curve.

• Choose 2 or more images and small Rectangle on the "bleach spot" where you see the intensity go down in the 4th time point.
• Select the images and run script (on web-dev-merge under practical > python > server > simple_frap_with_figure
• Script success message gives the figure ID. Open new tab at /figure/file/ID/ to see it. (screenshot in comment below)

[jburel]

We should probably keep 2 scripts: frap and frap with figure
since the figure section is large due to the fact that we have to copy some code from gateway

[will-moore]

Added timestamp labels and scalebars:

[pwalczysko]

The script ...with figure version seems to work fine. It does create the Figure as advertised.
But probably because of the missing matplotlib on eel, we do not have the FRAP graph on the Figure. I have imported the images to outreach, and tomorrow I will install the script as well to test this, and also to check whether matplotlib is installed on outreach.

I am a bit confused what output is there from the script simple frap - maybe this is not dealt with in this PR - but if yes, why is the script there on eel with no output to show ?

[pwalczysko]

This still works fine, even on outreach, after I have manually installed the scripts there. (note that I have to solve the ansible problem on outreach tomorrow with @manics as inticated in the chat ).

The absence of the FRAP graph panel in the new figure suggests though that even on outreach, there is no matplotlib installed. Is it a question of adding the installation to the playbook via pypi @manics ?)

[pwalczysko]

The font in the created graphs (the axis labels) is really tiny - could we get a bigger one please ?
I have to zoom in the figure to the max to be able to show something

[pwalczysko]

when I start simple frap with figure script I get in the UI

That is a bit confusing (I suppose both simple frap and simple frap with figure somehow share the same UI bit ?

[will-moore]

@pwalczysko UI name fixed in ee94c1d

[pwalczysko]

RFE:

• simple frap with figure creates a figure -> intuitive
• simple frap creates a figure as well -> anything but intuitive

The difference between the two scripts is not whether or not they create an OMERO.figure, but rather, whether or not they create a new image with the FRAP graph on it (and add it to the figure).

[pwalczysko]

Other than #32 (comment) confusion, it works as expected here. Note that all should be now uploaded to outreach, no need to upload scripts manually there anymore. The playbook will be adjusted once this (and other PRs) are merged (and also when the Kymograph is sorted out)

[will-moore]

@pwalczysko Apologies, I think the confusion at #32 (comment) was because when I first opened this PR I uploaded the modified "simple_frap.py" script to outreach with figure creation. Then I removed figure creation from that script but did not update on outreach.

[will-moore]

@pwalczysko Updated simple_frap.py on outreach now.

[pwalczysko]

Works as expected. Ready to merge FMPOV