i2pc / scipion Goto Github PK

• Delete this param in all protocols (relion, highRes, frealign, etc)
• Check if this param is copied with copyInfo function.

Allow the possibility to have a file where the user can take project notes.
Also configure which viewer will be used to open the notes file.

Suggested by @rmarabini

scipion viewer movie.mrcs fails if movie.mrcs is greater than 2 Gb.

The _extended attribute can only hold and extended item id inside a set or output result (as outputParticles)
but not both. Which is the case when selecting items from output sets.
This issue also provokes other problems:

1. Define relation fails when using an item from a set as input
2. Merge fails when picking items from sets (maybe the same reason as in 1)

Install it on folder /gpfs/fs1/apps/scipion/ with user xmipp.
First we got an error cause git was old. Jesus installed a new git (2.4) but to use it you need to source . /gpfs/fs1/home/bashrc on your env.
Then we got an error trying to install scipy and opencv so we installed it without those modules.
Then we run some tests and found that Eman 2.1 does not work cause it has been precompiled with a libc library wich is not the one on crunchy system. According to Jesus this cannot be changed without reinstalling crunchy.
Relion movie alignment does not work cause it needs opencv.
Resmap cannot be installed either cause it depends on scipy.
After updating scipion when run scipion install it failed installing pcre-8.36. IT gives following error:

make1: Entering directory /gpfs/fs1/apps/scipion/software/tmp/pcre-8.36' CXXLD libpcrecpp.la /usr/lib64/gcc/x86_64-suse-linux/4.1.2/../../../../x86_64-suse-linux/bin/ld: .libs/libpcrecpp_la-pcrecpp.o: relocation R_X86_64_PC32 againststd::basic_string<char, std::char
_traits, std::allocator >::~basic_string()@@GLIBCXX_3.4' can not be used when making a shared object; recompile with -fPIC
/usr/lib64/gcc/x86_64-suse-linux/4.1.2/../../../../x86_64-suse-linux/bin/ld: final link failed: Bad value
collect2: ld returned 1 exit status
make1: *** [libpcrecpp.la] Error 1

The problem was that pcre lib is compiled with flag fPIC while libpcrecpp_la-pcrecpp.o (which is linked byt it) is not. Jordi fixed it.

import a set of movies
try to visualize them

If you enter a SetOfAverages as input, the output is SetOfPartices.
Check how it is done in xmipp - preprocess. Add the corresponding test validating that if enter averages, you get averages at the end.

Some mrc volumes are read as stacks, but when importing mask, we can consider always as volumes.

After finished, the time of each protocol continues counting....I think @ldelcano fixed a related problem some time ago.

The idea is to create a mechanism that allows modifying the conventions followed by Scipion while keeping backwards compatibility in the sqlite files

When creating a subset (let say of averages out of some classes) and in a further
step I try to select recently created averages, the object appears duplicated....

@ldelcano , could you help @airenzp with this issue so she can go back to web as soon as possible?

Perform a protocol execution in a remote host.
For this we need:

• Detect required files for the protocol
• Store protocol in a separate mapper.
• Copy all files to remote host(user and pass in configuration)
• Launch a remove command to execute the protocol.
• Remotely monitor the job (this is related to Job Monitoring task)

and show @tranquilinho how to do it :)

PSDs produced by the align movie protocol should be centered
Right now they are centered in the corners. Really added by @rmarabini in Redmine.

For instance, phase flipping on Xmipp Extract Particles protocol should not have default value to force user to set it (it has a deep impact on further processment).

It would be very useful to launch individual programs from scipion.
Scipion will setup the required environment on-demand without any .bashrc change.

Now, xmipp viewer is the central viewer shere showj is called to show scipion's objets.
We need to move to a more general viewer (outside of xmipp) for all scipion's objects viewer. Take into account this after the release.

Micrograph path on images.xmd points to tmp folder that seems to be emptied after protocol finishes.
This has been found when using this metadata to import particles on anothre project.

Create at least two test: one for converting from pdb id and another one from file.
Check that the sampling rate is properly set (when changing sampling rate and dimensions)

Bshow needs a star file with a specific format (see attached file).
Image file needs to be a stack (either a spider stack or a mrc stack).
If dealing with individual images we need a block per micrograph (image).
Create a bsoft.convert to create the required star file

It is not that bshow requires a star file to open the image.
Bshow can open spider or mrc stacks and single img files but if we want to open many imgs they cannot be opened together on a single viewer (like xmipp does), we need to create a stack or a star file.

Xmipp metadata_sql module does not accept metadata labels with a 'dot' in the middle.
Jose Miguel suggest maybe method label2Str should be changed to return the string already with quotes.

Remove debug prints in showj:
column _filename allow render true is image true

And check for others.

@cossorzano I think after commit ad721c0 the test is broken.

scipion test em.protocols.test_protocols_xmipp_3d.TestXmippCreateMask3D

Error:
...
AttributeError: 'XmippProtCreateMask3D' object has no attribute 'outputMask'

Apply geo by default

The user should be able to create a project on a different location that ScipionUserData. In this case a simbolic link should be created there pointing to the real location.

Showj does not correctly computes the Avg & Std images when using transformation matrix.

To see the problem open the XmippTuturial project and check the output SetOfParticles in cl2d only align
and also the output Xmipp md file.

after instalation of motion_cor the program dosefgpu_driftcorr is not in the path
therefore movie_alignment protocol will fail ;-)

First version exists that performs an automatic gaussian picking for all micrographs when it has been performed by the user on one of them.
Pending to ask user if she wants to do this.
And investigate if this can be implemented on a different protocol that uses a wizard (that could be also used by dogpicker) and xmipp gui to show coordinates.

In the project graph view, if we click and move very fast a node "box", it seems
that some Exception happens and the node coordinates go to infinity....

Reported by Roberto:
If I threshold a mask (pre-process volume) want I get is a volume but not a mask.
Therefore, I can not use mask in any other program

std::cerr commands in C code appears in run.stdout instead run.stderr in protocols

At least this hapends with program

gnedIntermediate.xmd --sampling 1.340000 --cropULCorner -1 -1 --cropDRCorner -1 -1 --frameRange 2 24 --max_shift 16 --oavg Falcon_2015_05_21-15_06_09_0_aligned.mrc --oaligned movie_000001_kk.mrcs

out put:
Frame 2 to Frame 3 -> (0.967423,-0.0141962)

In summary the first frame is 1, in parameter help the first frame is 0

Also keep statistics of downloads (use jplot)

Movie alignment
Ransac
Resmap wizard for different volumes
Review getFiles method for each object

Also temporal images are generated on project folder, instead of protocol tmp.

After discussing with Vahid I think the PSD computed by the movie alignment protocol is divided in two halves showing the PSD before and after running optical flow. May be we should modify it so it shows the psd before dosegpu is used (if selected in the protocol)

I experienced this error if the micrograph filename is not shown by default after the ctffind4 protocol.
If I go to View -> Columns and select to display (not even render) the _micObj._filename. Then anytime that I right-click -> Open, it crashes in the Java virtual machine

Also the elements ended with _filename are rendered following the function _loadTable at the line 338 in the file dataset.py, but in the case of the rot.spectra this doesnt make sense.