CTA IRAP Project Gateway: Issueshttps://cta-redmine.irap.omp.eu/https://cta-redmine.irap.omp.eu/favicon.ico?14312453732023-03-27T15:14:56ZCTA IRAP Project Gateway
Redmine ctools - Bug #4239 (Feedback): Mac OS X binary build fails due to access denied exceptionhttps://cta-redmine.irap.omp.eu/issues/42392023-03-27T15:14:56ZKnödlseder Jürgenjurgen.knodlseder@irap.omp.eu
<p>The Mac OS X binary build fails with the following message:<br /><pre>
java.io.IOException: remote file operation failed: /Users/jenkins/jenkins/workspace/ctools-cr-osx at hudson.remoting.Channel@54e7e181:Mac OS 10.11: java.nio.file.AccessDeniedException:
</pre><br />It turned out that the <code>ctools.egg-info</code> directory in the <code>ctools-cr-osx</code> is owned by <code>root</code> while all other files are owned by <code>jenkins</code>:<br /><pre>
$ ls -l pkg_build/src/ctools-2.1.0.dev/pyext/
total 232
-rw-r--r-- 1 jenkins staff 27643 27 mar 14:31 Makefile
-rwxr-xr-x 1 jenkins staff 9094 27 mar 09:46 Makefile.am
-rw-r--r-- 1 jenkins staff 29847 27 mar 09:46 Makefile.in
drwxr-xr-x 6 root staff 204 27 mar 14:33 ctools.egg-info
-rwxr-xr-x 1 jenkins staff 2642 27 mar 09:46 ctphase.i
-rwxr-xr-x 1 jenkins staff 2607 27 mar 09:46 ctprob.i
-rwxr-xr-x 1 jenkins staff 2682 27 mar 09:46 ctpsfcube.i
-rwxr-xr-x 1 jenkins staff 2624 27 mar 09:46 ctselect.i
-rwxr-xr-x 1 jenkins staff 2847 27 mar 09:46 ctskymap.i
-rwxr-xr-x 1 jenkins staff 2707 27 mar 09:46 cttsmap.i
-rwxr-xr-x 1 jenkins staff 2741 27 mar 09:46 ctulimit.i
-rw-r--r-- 1 jenkins staff 7739 27 mar 14:31 setup.py
-rwxr-xr-x 1 jenkins staff 8044 27 mar 09:46 setup.py.in
</pre></p> ctools - Bug #3766 (New): conda install on Ubuntu 20.04 does not work for Python 3.8https://cta-redmine.irap.omp.eu/issues/37662021-09-06T10:55:46ZKnödlseder Jürgenjurgen.knodlseder@irap.omp.eu
<p>Monica reports the following:</p>
<p>I have updated to Ubuntu 20.04 now and have found out the following:</p>
<p>1) I was hitting a conda bug not specific to ctools which created many <br /> conflicts.</p>
<p>I saw a similar bug reported here:<br /><a class="external" href="https://github.com/nanoporetech/medaka/issues/117">https://github.com/nanoporetech/medaka/issues/117</a></p>
<p>This was fixed by modifying ~/.condarc and removing strict:<br />channel_priority: strict<br />Using flexible worked!</p>
<p>2) I was not able to install ctools with python 3.8 like in my MAC, but <br />python 3.7 worked with the standard instructions in Ubuntu.</p> ctools - Feature #3713 (New): Check file extension for output observationshttps://cta-redmine.irap.omp.eu/issues/37132021-05-28T07:57:40ZTibaldo Luigiluigi.tibaldo@irap.omp.eu
<p>Several tools, e.g., ctobssim or ctselect, can provide output observations in multiple formats (event lists in FITS, observations lists in XML). If the user specifies an output file name with the wrong extension (for instance, .fits for an observation list) the tools runs without any warning or exception, but the file then is unreadable by other tools.<br />I think we should add an exception or at least a specific warning in the log when this occurs.</p> ctools - Action #3633 (New): Make sure that all FITS files have unitshttps://cta-redmine.irap.omp.eu/issues/36332021-05-01T21:22:48ZKnödlseder Jürgenjurgen.knodlseder@irap.omp.eu
<p>It was reported that exposure cubes have no units. We should verify that all files produced by ctools come with a correction description of the units of the content.</p> ctools - Action #3610 (New): Add spatial butterflyhttps://cta-redmine.irap.omp.eu/issues/36102021-04-22T15:30:06ZKnödlseder Jürgenjurgen.knodlseder@irap.omp.eu
<p><code>ctbutterfly</code> computes the spectral butterfly diagram for a fitted model. The same approach can be applied to the spatial parameters, providing the error ellipse. Similar butterflies may be generated for other spatial parameters.</p>
<p>Probably the best is to add this capabilities to <code>ctbutterfly</code>, but one could also envision an additional tool.</p> ctools - Support #3539 (In Progress): cutlimit to averaged velocity cross sectionhttps://cta-redmine.irap.omp.eu/issues/35392021-02-14T13:05:17Zmarafatto lorenzolmaraf@yahoo.com
<p>I modelled DM profiles using CLUMPY and get J-Factor and spectrum<br />then I used ctobssim to get events and then ctulimit to get the limits<br />I need the <sigma v> for my research. I used standard 3e-26 value for calculating values from CLUMPY.<br />how can I convert ctulimit final value into <sigma v> velocity averaged cross section?<br />I add the fits file and spectrum from CLUMPY and events and ctulimit.log from ctools <br />many thanks!</p> ctools - Action #3537 (New): Speed-up tutorials https://cta-redmine.irap.omp.eu/issues/35372021-02-04T22:58:45ZKnödlseder Jürgenjurgen.knodlseder@irap.omp.eu
<p>Some of the tutorials take quite some time, and hence are not really suited to learn using ctools. These are in particular the tutorials that are based on the 1DC data. A smaller but meaningful set of test data should be used for the tutorials so that they can quickly be reproduced.</p> ctools - Action #3536 (In Progress): Check performance of csresmap and csrespechttps://cta-redmine.irap.omp.eu/issues/35362021-02-04T13:54:10ZTibaldo Luigiluigi.tibaldo@irap.omp.eu
<p>These tools have become really slow: calculating residuals takes much longer than performing likelihood fits in some cases. We should check where the bottlenecks are, whether they can be sped up etc</p> ctools - Action #3534 (New): Try to speed-up csphagenhttps://cta-redmine.irap.omp.eu/issues/35342021-02-03T20:33:31ZKnödlseder Jürgenjurgen.knodlseder@irap.omp.eu
<p>The <code>csphagen</code> script takes quite some time for the examples on <a class="external" href="http://cta.irap.omp.eu/ctools/users/tutorials/1dc/first_onoff.html">http://cta.irap.omp.eu/ctools/users/tutorials/1dc/first_onoff.html</a>. In total 30 observations are stacked by the run. A valgrind analysis should be performed to understand where the bottlenecks are.</p> ctools - Feature #3533 (New): Nuisance parametershttps://cta-redmine.irap.omp.eu/issues/35332021-02-03T18:18:03ZHernández Cadena Sergioskerzot@ciencias.unam.mx
<p>Add a property for parameters in the likelihood to be treated as nuisance parameters, also be able to define a prior distribution for this kind of parameter (Gaussian or Uniform prior distributions). As an example, you can refer to the DM conventions document in section 6 (attached)</p> ctools - Feature #3523 (New): Add to csmodelselect the option to select models for a binned/stack...https://cta-redmine.irap.omp.eu/issues/35232021-01-30T10:53:03ZTibaldo Luigiluigi.tibaldo@irap.omp.eu
<p>Sometimes the ROI for a binned analysis is smaller than the ROI of the corresponding event lists. Right now csmodelselect only accepts event lists in input, it would be useful to have the option to use it also for binned/stacked observations.</p> ctools - Action #3455 (In Progress): Vectorise ctmodelhttps://cta-redmine.irap.omp.eu/issues/34552020-11-21T12:13:32ZKnödlseder Jürgenjurgen.knodlseder@irap.omp.eu
<p>The <code>ctmodel</code> tool should make use of the vectorised response functions.</p> ctools - Feature #3433 (New): Improve support for source localization and extension fittinghttps://cta-redmine.irap.omp.eu/issues/34332020-11-04T11:28:55ZTibaldo Luigiluigi.tibaldo@irap.omp.eu
<p>Create dedicated functions/scripts for localization and extension fitting and add dedicated tutorial</p> ctools - Action #3319 (In Progress): Shorten ctools unit testinghttps://cta-redmine.irap.omp.eu/issues/33192020-08-09T08:43:09ZKnödlseder Jürgenjurgen.knodlseder@irap.omp.eu
The ctools unit test starts to last again for too long. Here a list of specific bottlenecks:
<ul>
<li>Test cterror from Python (21 s)</li>
<li>Test cterror on command line (7.2 s)</li>
<li>Test ctobssim from Python (58 s)</li>
<li>Test ctobssim on command line (14 s)</li>
<li>Test cttsmap on command line (9.1 s)</li>
<li>Test unbinned in-memory pipeline (14 s)</li>
<li>Test unbinned pipeline with FITS file saving (14 s)</li>
<li>Test csbkgmodel from Python (3m5s) ⇐ MAJOR BOTTLENECK</li>
<li>Test csbkgmodel on command line (12s)</li>
<li>Test csbkgmodel pickeling (12s)</li>
<li>Test csiactobs from Python (9s)</li>
<li>Test cslightcrv from Python (31s)</li>
<li>Test cslightcrv on command line (10s)</li>
<li>Test cslightcrv pickeling (9.6s)</li>
<li>Test csmodelsois from Python (7.5s)</li>
<li>Test csphagen from Python (11s)</li>
<li>Test csphasecrv from Python (15s)</li>
<li>Test csscs from Python (38s)</li>
<li>Test csspec from Python (32s)</li>
<li>Test csspec on command line (14s)</li>
<li>Test csspec pickeling (14s)</li>
<li>Test cstsdist from Python in unbinned mode (29s)</li>
<li>Test cstsdist on command line (30s)</li>
<li>Test cstsdist pickeling (29s)</li>
<li>Test csworkflow from Python (29s)</li>
<li>Test csworkflow on command line (15s)</li>
</ul> ctools - Bug #3317 (New): ctbkgcube aborts if no background model is foundhttps://cta-redmine.irap.omp.eu/issues/33172020-08-08T08:00:34ZKnödlseder Jürgenjurgen.knodlseder@irap.omp.eu
<p>While it seems logical that <code>ctbkgmodel</code> aborts if no background model is found, there could be a workflow where there is no background model, and consequently, the output files should be an empty cube. This is particularly important since tools may later request a background cube, although it may not be needed.</p>
<p>The logic should be changed so that <code>ctbkgmodel</code> does not abort in the absence of a background model, producing output that can be digested by subsequent tools, such as <code>ctlike</code>.</p>