ome / omero-figure Goto Github PK

Dear Will,

In the export, there is a problem depending on which PDF reader you use.

For instance, when using SumatraPDF (very popular) the text layers are there but in black, whereas in Adobe, they are in the correct colors.

What could be the issue here?

Several people in our lab have let me know that it's difficult to visually correspond the channels in the #channel_sliders div with the channels in the main Webclient. They suggested adding named labels for each channel. @jaycopeland let me know that this has been discussed, but there are concerns over space-constraints. Here is the solution I've hacked together (in a very temporary way):

The text is partially occluded depending on the position of the sliders, but that seems like a minor issue compared to the utility of this feature. @will-moore if this is an approach you'd like to take, I'd be interested in working towards implementing this in a proper pull request.

Hi,

I'm trying to install omero_figure on a server ( version 5.3.0-ice36-b59)
It already provides the OMERO.web front end just fine.

As the omero user I did the "pip install omero-figure", which worked,
but when I try to add the new application to the menus I get:

~$ bin/omero config append omero.web.apps '"omero_figure"'
Cannot retrieve default value for property omero.web.apps

I'm I missing something?
Any help would be highly appreciated.

Thanks
Andreas

See figure 33236 on demo server.

20th May: "There seems to be a bug with the scale bar in OMERO.figure (everything else is working OK). When I click "Show" the web page reloads, and no scale bar appears, although the "Show" button is replaced with "Hide". What's odd is that this feature was working correctly when I demo'd it for my group ~3 weeks ago, before the upgrade."

21st May: Screenshots showing that on click of "Show", the "Leave page?" dialog box is shown. If click 'Leave Page' then the "Hide" button is shown for this panel after reload.

Urls update like:

.../file/33236
.../file/33236/
.../file/33236/?

An optional grid (think graph paper) that does not go into exported PDFs just to be used to align things.

Snap-to-grid drag-n-drop is also a coveted feature.

Hi,
I am trying to use Figure with some images of RNA FISH and need to be able to label the RNAs in italics on the images as this is the convention in Drosophila and other organisms. I can use Markdown to get italics in the Figure legend (as suggested on the Trello card) but I can't get it to work in the label. Is it also possible to get italics in the label itself or would these labels have to be added afterwards?
Thanks,
Tamsin (Davis Lab)

It is probably not intended that when right clicking on an image, the image then moves while the context menu is open when moving the mouse. I assume probably that right click should just exhibit ordinary web context menu and nothing else.

For development with a remote Omero server, I'm setting chrome to make redirects like this:

from
http://localhost:8000/demo/%7B%%20url%20'webgatewayopen_with/
to
https://omero.hms.harvard.edu/webgateway/open_with/?_=1234567891234

from
http://localhost:8000/demo/%7B%%20url%20'webgatewa/render_image/548111/0/0/?
to
https://omero.hms.harvard.edu/figure/render_scaled_region/548111/0/0/?&region=-8160,-16416,48960,48960&

To my surprise, this works. Is there a proper way to configure a plugin like figure to use an external server?

Our users are starting to use figure a lot. With an increased in the number of figures, comes an organization issue because it's hard to find the figure they want. Example, consider the following 5 figure names:

• sggmRNA_sgg_NMJ_Nerve_single plane
• sggmRNA_sgg_s5_NMJ_single planes
• sgg_sggprobes_NMJ_zoom
• sgg_sggprobes_NMJ_zoom
• sgg_Rab11_Venus_NMJ_zoom

None of them will be listed by searching for "sgg NMJ". Only one will using "sgg_NMJ". Again, none will show up with "NMJ_sgg".

I understand that the search is done in the name, I'm just arguing that a name is not enough. This becomes worse when all the figures are at the same level, there's no filesystem-like hierarchy (but I would prefer a tag system to a hierarchic organization).

In OMERO it's possible to tag images and datasets. Something like that would be pretty sweet.

When selecting an image I notice that if I have any cursor velocity at all, the image will move a tiny amount. I'm not sure how programs generally handle this, perhaps only move after a sustained period of holding down left-click?

We have a figure where one of the panels has a dx and dy value of
null. This makes FigureExport.build_figure method to fail like so:

  File "Figure_To_Pdf.py", line 801, in build_figure
    self.add_panels_to_page(panels_json, image_ids, page)
  File "Figure_To_Pdf.py", line 1586, in add_panels_to_page
    image, pil_img = self.draw_panel(panel, page, i)
  File "Figure_To_Pdf.py", line 1408, in draw_panel
    pil_img = self.get_panel_image(image, panel, orig_name)
  File "Figure_To_Pdf.py", line 1326, in get_panel_image
    cx += dx
TypeError: unsupported operand type(s) for +=: 'long' and 'NoneType'

Seems it goes like this:

1. panels have a dx and dy value of null
2. python's json module converts those values to None
3. those values are accessed internally like panel.get('dx', DEFAULT_OFFSET)
4. because those keys do exist in the key, the DEFAULT_OFFSET is never used

I'm using the Figure_To_Pdf script like so:

from Figure_To_Pdf import FigureExport
ann = conn.getObject("FileAnnotation", id)
json = "".join(ann.getFileInChunks())
script_params = {
    'Figure_JSON' : json,
    'Webclient_URI': 'https://omero1.bioch.ox.ac.uk/webclient/',
    'Export_Option' : 'PDF',
}
fig_export = FigureExport(conn, script_params, export_images=False)
fig_export.build_figure()

A possible solution would be to have json.loads ignore values with a
null value (possible via the object_hook argument).

For what is worth:

• this seems to only happen when using the script programatically.
  When using the 'Export PDF' button on the omero figure application
  it works. I'm guessing something else on the GUI is fixing the
  nulls. That something else is also doing it when choosing the
  'File > Export as JSON' menu. So maybe that thing could be in the
  omero_figure package?

• I'm not sure how this image ended up with a null value there. We
  had a list of 350 figures to convert, and only one of them had this
  issue. So maybe this is not an issue on reading the JSON, but on
  how that JSON got generated in first place.

It would be very useful to have a way to add one image several times (showing different frames, and aligned on a row / column) for time lapse datasets. A syntax like ID:start:end:by might do the job; one might want to distinguish between column and row alignement (using a different separator? or an optional argument? or give a matrix dimension?).
Thinking about it, it might be easier to have dedicated dialog to select the frames and set the layout (in which case there would be a "add series" button next to the "add image" one). This could extended to handle any dimension of the dataset (channels, slice, time)…
Does this makes sense?

Is there a plan in the works to have tags available in the same way [channels] etc are currently? I think it would be useful for fast consistent annotation.

Generally useful I think, but particularly I noticed it when replacing one of the images in my figure and realising that the view was then totally unsuitable for the new image.

From @dpwrussell:
I noticed that when I pasted the JSON from figure into the JSON text entry in the script for generating PDFs, then I got this error:

Traceback (most recent call last):
  File "./script", line 633, in <module>
    runScript()
  File "./script", line 620, in runScript
    fileAnnotation = create_pdf(conn, scriptParams)
  File "./script", line 547, in create_pdf
    addInfoPage(conn, scriptParams, c, panels_json, figureName)
  File "./script", line 490, in addInfoPage
    c.save()
  File "/usr/local/lib/python2.7/dist-packages/reportlab/pdfgen/canvas.py", line 1215, in save
    self._doc.SaveToFile(self._filename, self)
  File "/usr/local/lib/python2.7/dist-packages/reportlab/pdfbase/pdfdoc.py", line 219, in SaveToFile
    data = self.GetPDFData(canvas)
  File "/usr/local/lib/python2.7/dist-packages/reportlab/pdfbase/pdfdoc.py", line 244, in GetPDFData
    return self.format()
  File "/usr/local/lib/python2.7/dist-packages/reportlab/pdfbase/pdfdoc.py", line 424, in format
    IOf = IO.format(self)
  File "/usr/local/lib/python2.7/dist-packages/reportlab/pdfbase/pdfdoc.py", line 869, in format
    fcontent = format(self.content, document, toplevel=1)   # yes this is at top level
  File "/usr/local/lib/python2.7/dist-packages/reportlab/pdfbase/pdfdoc.py", line 80, in format
    f = element.format(document)
  File "/usr/local/lib/python2.7/dist-packages/reportlab/pdfbase/pdfdoc.py", line 1600, in format
    return D.format(document)
  File "/usr/local/lib/python2.7/dist-packages/reportlab/pdfbase/pdfdoc.py", line 676, in format
    L = [(format(PDFName(k),document)+b" "+format(dict[k],document)) for k in keys]
  File "/usr/local/lib/python2.7/dist-packages/reportlab/pdfbase/pdfdoc.py", line 80, in format
    f = element.format(document)
  File "/usr/local/lib/python2.7/dist-packages/reportlab/pdfbase/pdfdoc.py", line 1776, in format
    if f is None: raise ValueError("format not resolved %s" % self.name)
ValueError: format not resolved None?show=image-2

It works fine when I run export-pdf from figure though.

Hi,

Using figure 1.2.1 in OMERO 5.2.3:

When adding a scalebar to be 'inside' top-right, or bottom-right, and changing the font size, the scalebar moves outside the image.

The scalebar should move towards the centre of the image when it's increased in size, rather than it being clipped and leaving the image.

Perhaps a separate issue, but it seems once I've chosen to place the scalebar inside the image, not outside the image, it's no longer possible to place it outside again. The only choices are which corner.

The Figure_To_Pdf.py script could be used by a client to generate PDF or TIFF from the figure json like so:

script_params = {
    'Figure_JSON' : figure_json,
    'Export_Option' : 'TIFF',
    'Webclient_URI': 'https://omero1.bioch.ox.ac.uk/webclient/',
}
fig_export = TiffExport(conn, script_params, export_images=False)
fig_export.build_figure()

However, this method ends by uploading the built figure as a file annotation to the omero server. Would it be possible to have a method that only does the building of the figure?

Some users wanted to change the color of the channels.

The pre-defined colors are the standard but offer very little flexibility in terms of coloring.

Perhaps a color picker like in OMERO view figure could be used?

Example: Image name LymphknotenNonParametric_4µ-pixel_spacing.dcm

Traceback:
File "/Users/andreas/omero/openmicroscopy/dist/lib/python/django/core/handlers/base.py" in get_response
  114.                     response = wrapped_callback(request, *callback_args, **callback_kwargs)
File "/Users/andreas/omero/openmicroscopy/dist/lib/python/omeroweb/decorators.py" in wrapped
  440.             retval = f(request, *args, **kwargs)
File "/Users/andreas/omero/figure/views.py" in save_web_figure
  162.     figureJSON = str(figureJSON)

Exception Type: UnicodeEncodeError at /figure/save_web_figure/
Exception Value: 'ascii' codec can't encode character u'\xb5' in position 70: ordinal not in range(128)

Exact measurements for both formatting and display that show the pixel area represented by an image viewport. E.g. If I had 2 same sized viewports on the same image but one was much more zoomed in that the other, the pixel measurements on that one would be less.

This is not a big issue, it's only the redudancy on the client that has to do:

TiffExport(conn, {'Export_Option' : 'TIFF', ...}, ...)

The option Export_Option in the Figure_To_Pdfscript is used to create a FigureExport or TiffExport object. It is then used again in FigureExport.build_figure to decide whether to create a zip file (it will create a zip file if exporting to a tiff file and there's multiple pages). However, that information is already in whether this is a TiffExport object. Having a class property like 'supports_multipage_per_file' would avoid that redudancy (although the TIFF does support multiple pages per file).

When we have time-lapse images in a figure, the figure itself could be exported as a movie, with all the included images playing simultaneously.
This would allow users to compose a specific layout for a movie just as they can do for a figure.
Time-stamp labels should probably show delta-T from start of movie.
Time offset of individual movies with respect to one-another will be preserved.
Possible export option would allow to choose start/end of the time-course.
Export options could be presented in a dialog box when user chooses export as movie.

Probably need to support "crop page to panels" first

Figure requirement markdown is installing latest Pillow, if that is done in the wrong order (that is likely as figure is installed after server installation) this will cause https://trello.com/c/jZjPZf8K/637-attributeerror-numpy-ndarray-object-has-no-attribute-tobytes.

Once we moved figure to become PyPi package I removed logic to read figure version from package.json. That's because package.json is not a part of the app and won't be included. I provided a workaround to read the version from resources like:

OMERO_FIGURE_VERSION = pkg_resources.require("omero-figure")[0].version

but I wasn;t sure if that is the best solution.

At the moment Figure version is in 2 places:

• https://github.com/ome/omero-figure/blob/master/package.json#L3
• https://github.com/ome/omero-figure/blob/master/omero_figure/settings.py#L8

Would be nice to unify that.

cc @jburel

Not strictly a Figure issue as everyone would love to see it in the clients as well, but it was mentioned so recording it here.

Basically everyone wants what softWoRx offers with regards to the histogram.

Will,

It would be great if we could use the current color wheel to set the background color of the Canvas.

Indeed, sometimes it is preferable to have a black background for instance instead of white.

What say you?

We've recently run into an issue (figure 3.2.1 on a 5.4.5-ice36-b83 server) with tiled images that cannot be imported into OMERO.figure and give the error "Image '{NAME}' is too big for OMERO.figure.

One image in question is 8460 x 8460 pixels of 16bit data and 4 channels. I can't find any documentation about a file size limit other than issue 37 which led me to try a larger paper size. Going up to A0 still gives the same error.

The only workaround I've found is to downsample the images and keep a high-res for analysis and low res for figure, but this seems a bit perverse.

From Sandra:
Needs the pixel size for the selection frame when cropping
Wants to be able to see the size and whether square etc. when cropping an image so can do same size-shaped crop on another image to enable comparison.

See http://lists.openmicroscopy.org.uk/pipermail/ome-devel/2015-July/003391.html

The permissions system in the read-write group should allow true figure collaboration to be feasible.

Hello,

The Davis lab have been trialling OMERO figure and have come up with a list of comments / queries that I hope we can discuss at the meeting:

• can you use Gamma as well as max min with contrasting images in figure?
• can you “copy and paste” contrasting settings between images in figure or type in specific contrast values so different images get the same contrast?
• Could there be both max and average projection options for Z-stacks?
• As well as rotate could there also be flip (or mirror) horizontal and vertical?
• When you add images to a figure the most recent image can sit “on top” of another image as an overlay. Is there a way of doing this so you can get multiple image insertson figure independent of the order they were added (Layers)?
• Is there a way of adding arbitrarily placed characters e.g. N for nucleus the same way you can add an arrow anywhere, or even an arbitrary txt box - e.g. for making posters?
• Can you copy multiple image ID's at once from OMERO to paste into OMERO figure?

Many thanks

Josh

1. Tool tips for icons in the preview pane.
2. Demo figure : Show in webclient could point at the JCB figure.
3. Import as : (paste from local backup of the figure). --> Kinda acts like a template object and helps in omero independent archiving.
4. Delete attachment (throw a pop-up with a warning message "this is an omero figure , and deleting this would delete the figure permanently".
5. Maximum Intensity projection : Default option normally would be zmin to zmax. Explain the use of the two sliders via a tooltip?

Minor Bug :
Thumbnails not showing the "Edit ID" window.

I can't stress enough how important this is. Everyone I showed it to immediately asked for this.

If the users embrace this they will quickly gather a huge number of saved Figures that need to be categorisable and browseable. They want to be able to have stuff like: Paper->Experiment->Figure1 etc, etc.

It is very useful when working with something that should end up perfectly lined up to be able to move objects along one axis while locking the other.

For example, if I have four images in a square and wish to move the bottom two down a bit, it would be good to be able to hold down a modifier key while moving them so that they moved in the y axis only, x being locked.

Rotation is very nice but for consistency sometimes you need to mirror an image (eg to get all images in the same orientation). Can you add a mirror function, combined with rotation this allows all possible orientations.

At this early stage it would probably be a good idea to have a feedback link straight to GitHub issues on the main bar. Honestly, if it takes more than 1 click to get to the place to give feedback, most users will not ever get there.

Hi,

While fiddling with the allowed image size, we noticed that if the image is very big, the panel which holds the image in the figure gets created but not populated. The image successfully loads in the Preview, but not in the figure view.

Looking at the console, there are some JS errors coming from raphael.js having to do with negative coordinates on x-y system.

This bug is in principle not an issue when the pixel size is restricted to 5000x5000 but if this setting was arbitrary, there might still be a bug to be fixed here.

Figure 1.2.1, server "OME OMERO demo server"

I managed to come across the following error while trying to export a figure I made with a few images and a bundle of ROIs drawn on the images.

I can't replicate the error, but perhaps there's a range check or something which needs to be handled?

/usr/lib64/python2.6/site-packages/reportlab/lib/utils.py:653: DeprecationWarning: tostring() is deprecated. Please call tobytes() instead.
  self._data = im.tostring()
Traceback (most recent call last):
  File "./script", line 1716, in <module>
    runScript()
  File "./script", line 1703, in runScript
    fileAnnotation = export_figure(conn, scriptParams)
  File "./script", line 1660, in export_figure
    return figExport.buildFigure()
  File "./script", line 688, in buildFigure
    self.add_panels_to_page(panels_json, imageIds, page)
  File "./script", line 1343, in add_panels_to_page
    image, pilImg = self.drawPanel(panel, page, i)
  File "./script", line 1160, in drawPanel
    pilImg = self.getPanelImage(image, panel, origName)
  File "./script", line 1080, in getPanelImage
    cx += dx
TypeError: unsupported operand type(s) for +=: 'long' and 'NoneType'

Figures belong to both a user and a group. The first is pretty obvious and when searching for existing figures, one can filter by user. However, figure group permissions are never referred on the GUI. Since a user can belong to multiple groups, it's not obvious to the user which group his figures are.

The problem we just had were two users that belong to different primary groups and a same secondary group. They wanted to share figures with images from this secondary group. Because they were not aware of group permissions for figures, some of them ended up on their primary group and so were invisible to the other user.

The GUI should make some reference to the group a figure is in. I would guess a 'Move to group' button on the 'File' menu may also be useful.

Hi,

first of all: Thank you for this very useful plugin, it's one of the things our users like most about working with OMERO.

Now the bug report: Exporting multiple pages to PDF at our system with the latest OMERO.figure from GIT (2012-12-15) will only include the first page. Is this only us or is this functionality missing?

As a side node: The ability to add a label to the scale bar is quire crucial for our users to really adopt the plugin.

Thanks,

Tris

Scale bars are displayed the correct length in the web UI, but in the exported file they are too long, apparently by a factor of 300/72, i.e. there might be an extra or missing call to scale_to_export_dpi.

Left is web UI, right is exported file.

Hi,

I am working with some of my users whom are using Figure. We are importing '.dv' files into figure and then doing maximum projection. We have found that by adjusting the min/max brightness settings in the OMERO.web client these settings are imported to Figure. This is good as it means we can have the same range in each image when we import each to Figure. We have found that by shift-clicking all the images within Figure, we can globally tune the min/max settings, but we have noticed a bug. When we move the max bar, the range we set in OMERO.web is lost and it defaults to the max image intensity level. This results in a loss of synchronisation between the grouped images and chaotic min/max control if the images have different max levels. We can work around this by just controlling all brightness settings in OMERO.web and then importing, but it is partly the power of OMERO.web to do this within this software package. Can you advise about this and potentially fix the issue. It would be best if in Figure it did not default to the max intensity of the image but retained the preset imported from OMERO.web.

Best,
Dominic

I have 3 data sets that I wish to compare, they are from different systems with different pixel sizes. I am able to label them all with scales bars (of the right scale!) and have the same viewport size is pixels (eg 200x200) , al of this is brilliant. My question is can I make them all have the same effective magnification. eg the one with 10 nm pixels needs to be 4 times smaller than the one with 40 nm pixels.

Thanks, Ian

What is the best way to download several Omero.Figure files from the webclient?

-Would like to retrieve them by their given file name rather than the file id number.
-Download a single zip containing multiple PDFs
-Script can be run from webclient / commandline/ or client-side python, whichever is easiest

Many thanks,
J

$ git diff package.json
diff --git a/package.json b/package.json
index ca8e65a..2293118 100644
--- a/package.json
+++ b/package.json
@@ -1,6 +1,6 @@
 {
   "name": "figure",
-  "version": "1.0.0-beta1",
+  "version": "1.1.1",
   "description": "OMERO figure creation app",
   "main": "index.js",
   "scripts": {
@@ -8,7 +8,7 @@
   },
   "repository": {
     "type": "git",
-    "url": "git://github.com/will-moore/figure.git"
+    "url": "git://github.com/ome/figure.git"
   },
   "keywords": [
     "backbone",
@@ -18,7 +18,7 @@
   "author": "William Moore",
   "license": "GPL",
   "bugs": {
-    "url": "https://github.com/will-moore/figure/issues"
+    "url": "https://github.com/ome/figure/issues"

would be good if the release process could include an auto-bump.

When we import an image from OMERO 5.1, display the units 'symbol' in the scalebar form and label (and save units enumeration, E.g. "MICROM"?).

Later we may want to allow users to choose the units of a scalebar. E.g. if an EM image has been imported with microns, you may want to show a scalebar in Angstroms etc. Will need conversion methods available somehow.

Will,

Our uses would love to have the possibility to place arrows, or other shapes, but mostly arrows on their Figure.

That way it would reduce the necessity to do this in another program later on.

What do you think?

Linked to that, our group here sometimes draws a box around the panel labels so that it looks more "straight". Having the possibility for shapes would allow to both annotate the panels and their labels nicely.

What is the best way to input a list of file names and return the file ID numbers?

Thanks!
-J

Hi,

I have used the [image-name] to automatically add a label to an image and as you can see it has weird characters, especially the double quotes.

However, once it is added, the label cannot be deleted, or the color changed.

I guess it has something to do with the "".

Best,

Hi some journals like scale bars labelled with a size whereas other require the size in the ledgend. Can we have a option to have labelled or unlabeleld scale bars.

Thanks,

Ian