id
int64 0
5.38k
| issuekey
stringlengths 4
16
| created
stringlengths 19
19
| title
stringlengths 5
252
| description
stringlengths 1
1.39M
| storypoint
float64 0
100
|
|---|---|---|---|---|---|
799 |
DM-4236
|
11/02/2015 14:00:55
|
Specify default output location for CmdLineTasks
|
When neither {{\--output}} or {{\--rerun}} is specified as an argument to a {{CmdLineTask}}, any output from that task appears to be written back to the input repository. Note the use of the term "appears": from a preliminary inspection of the code and documentation, it's not clear if this behaviour can be overridden e.g. by environment variables. The HSC stack behaves differently, using {{$INPUT/rerun/$USER}} as a default output location. A [brief discussion|https://community.lsst.org/t/new-argument-parser-behavior-rerun-flag-introduction-discussion/345] suggests that this is the preferred behaviour. Please update the LSST stack to match the HSC behaviour.
| 2 |
800 |
DM-4237
|
11/02/2015 15:12:45
|
unable to upload images to nebula
|
I seem to be unable to upload an image to neblua from a URL via either horizon or the nova cli client. The request seems to queue briefly and then reports a status of {{killed}}. Eg {code} glance image-create --name "centos-7.1-vagrant" --disk-format qcow2 --container-format bare --progress --copy-from http://sqre-kvm-images.s3.amazonaws.com/centos-7.1-x86_64 --is-public False --min-disk 8 --min-ram 1024 {code}
| 2 |
801 |
DM-4238
|
11/02/2015 15:30:19
|
Fix integer casting error in numpy version 1.10 in obs subaru
|
Fix type casting in obs_subaru in lates numpy in obs_subaru
| 0.5 |
802 |
DM-4239
|
11/02/2015 15:43:47
|
Identify Qserv areas affected by secondary index
|
Evaluate Qserv software for the Czars and workers to identify where an interface to the secondary index will be required for efficient operation.
| 5 |
803 |
DM-4245
|
11/02/2015 16:30:48
|
Image Viewer memory leak
|
When reloading the same 500MB RAFT image into an image viewer (see the script below), it was discovered that single node Firefy server with 3G memory runs out of memory after ~15 reloads Test case: keep reloading the html file with the following Javascript, creating an image viewer with 500MB image: function onFireflyLoaded() { var iv2= firefly.makeImageViewer("plot"); iv2.plot({ "Title" :"Example FITS Image'", "ColorTable" :"16", "RangeValues":firefly.serializeRangeValues("Sigma",-2,8,"Linear"), "URL" :"http://localhost/demo/E000_RAFT_R01.fits"}); } Follow up: The bug was traced to java.awt.image.BufferedImage objects not being evicted from VIS_SHARED_MEM cache. Further search showed that java.awt.image.BufferedImage (along with java.io.BufferedInputStream) is in src/firefly/java/edu/caltech/ipac/firefly/server/cache/resources/ignore_sizeof.txt, which lists the classes that have to be ignored when calculating the size of cache. Testing on single node server (VIS_SHARED_MEM cache is not replicated), using [host:port]/fftools/admin/status page: BEFORE (java.awt.image.BufferedImage was commented out in ignore_sizeof.txt) After 14 reloads: Memory - Used : 3.7G - Max : 3.55G - Max Free : 488.0M - Free Active : 488.0M - Total Active : 3.55G Caches: VIS_SHARED_MEM @327294449 Statistics : [ Size:15 Expired:0 Evicted:0 Hits:246 Hit-Ratio:NaN Heap-Size:1120MB ] OUT OF MEMORY on next reload AFTER THE CHANGE (Commented java.awt.image.BufferedImage in ignore_sizeof.txt) After 36 reloads: Memory - Used : 1672.9M - Max : 3.55G - Max Free : 1968.0M - Free Active : 1468.0M - Total Active : 3.6G Caches: VIS_SHARED_MEM @201164543 Statistics : [ Size:3 Expired:0 Evicted:34 Hits:659 Hit-Ratio:NaN Heap-Size:1398MB ]
| 2 |
804 |
DM-4251
|
11/04/2015 10:29:07
|
Please include obs_subaru in CI
|
{{obs_subaru}} should be included in the CI system.
| 1 |
805 |
DM-4252
|
11/04/2015 11:50:54
|
Create GitLFS Technical Note
|
Create a SQuaRE Technical Note describing the architecture of the GitLFS service implementation.
| 2 |
806 |
DM-4256
|
11/04/2015 15:06:53
|
Sphinx support of sqr-001 technical note
|
Support the distribution of a technical note SQR-001 - Remove oxford comma in author list (documenteer) - Solve issue where title is repeated if the title is included in the restructured text document - Solve issue where name of the HTML document is README.html
| 1 |
807 |
DM-4259
|
11/04/2015 15:45:42
|
Create a set of tests (or update the current ones) to facilitate refactoring of dipole measurement
|
This will create a test (not necessarily a unit test) that will simulate dipoles and measure them so that the measurement can be compared to truth values. This may be simply refactoring the current tests. This task should also include generating more generalizable utilities needed to create the dipoles and incorporating these and other test data into the stack so that they can be used in other studies.
| 5 |
808 |
DM-4264
|
11/05/2015 13:06:45
|
SQuaRE supertask design meeting 2
|
Hold a teleconference design discussion with members of the SQuaRE team.
| 1 |
809 |
DM-4266
|
11/05/2015 13:20:44
|
Should only read fringe data after checking the filter
|
The fringe subtraction is not necessarily performed if {{doFringe}} is True. It is only if the filter of the raw exposure is listed in config fringe.filters. Fringe data should not be read unless the filter is indicated. There are likely no such filter data and it would cause runtime errors. Seems related to changes from RFC-26 and DM-1299.
| 8 |
810 |
DM-4286
|
11/06/2015 22:35:34
|
Processing of DECAM COSMOS field - Part I
|
This story covers work on the verification plan done in October. The DECAM Cosmos field was selected, however DECAM ISR is not available so the starting point for now is Community Pipeline reduced data. Have put the data through processCcdDecam and makeCoaddTempExp but had a number of failures that we have so far not been able to pin down. A list of user experience issues is being collated and [~frossie] will generate Summer 16 cycle stories to address those that fall within SQuaRE's defined scope of activities. Story closed to fit within month, but work is ongoing.
| 20 |
811 |
DM-4295
|
11/09/2015 12:05:49
|
Run and document multinode integration tests on Openstack+Docker
|
Boot openstack machines using vagrant, then deploys docker images and finally launch multinodes tests. FYI, lack of DNS on OpenStack Cloud cause problems, but a vagrant plugin seems to solve this.
| 8 |
812 |
DM-4298
|
11/09/2015 15:13:16
|
want equiv of m1.xlarge flavor with smaller disk
|
I'd like to be able to build images with vcpus & ram from the {{m1.xlarge}} flavor that can be run on a {{m1.medium}} with it's smaller disk image, this would require a new flavor with 16GiB ram/8vcpus but only 40GiB of disk. Something along the lines of: {{openstack flavor create --ram 16384 --disk 40 --vcpus 8 ...}} Is that possible?
| 2 |
813 |
DM-4301
|
11/10/2015 12:17:50
|
Convert banner and menu to react/flux
|
Add flux data model to capture menu and banner information. Convert banner and menu UI from gwt to react. As part of this task, bring in Fetch API to simplify client/server interactions.
| 8 |
814 |
DM-4303
|
11/10/2015 17:15:45
|
re-deploy lsstsw on Jenkins
|
Pandas was added to the bin/deploy script in lsstsw to support sims development. This has already been merged to master in 4b1d1a0fa. The ticket is to ask that lsstsw be redeployed so the sims team can build branches that use pandas.
| 3 |
815 |
DM-4304
|
11/10/2015 17:38:12
|
Add unit testing into gradle build for Firefly's server-side code
|
Add a test task to Firefly's common build script. This can be used by any sub-project to run unit test. Added unit testing to Jenkins continuous integration job to ensure new code does not break unit testing.
| 1 |
816 |
DM-4305
|
11/10/2015 19:34:35
|
assembleCoadd broken
|
A recent update to assembleCoadd to bring over changes to do clipped coadds breaks coadd generation. There are two specific problems. 1. There is an infinite recursion because of SafeClipAssembleCoaddTask calling its own constructor in the \_\_init\_\_ method. 2. The overridden assemble method does not adhere to the original assemble call signature, so when the default run method is called by ParseAndRun, it raises an exception. Additionally I find the flow fairly confusing as the overridden assemble method is called by the default run method which then calls the default assemble method on the parent class.
| 1 |
817 |
DM-4307
|
11/11/2015 08:24:15
|
Please add HSC tests to CI
|
In DM-3663 we (= [~price]) provided an integration test for processing HSC data through the stack with the intention that it should be integrated with the CI system. Having this test available and regularly run would be enormously helpful with the HSC port -- we've already run into problems which it could have helped us avoid (DM-4305).
| 0.5 |
818 |
DM-4310
|
11/11/2015 10:53:36
|
Missing Doxygen documentation
|
As of [2015-11-10 02:53.26|https://lsst-web.ncsa.illinois.edu/doxygen/xlink_master_2015_11_10_02.53.27/] there were 19 "mainpages in subpackages" available through Doxygen. In the next build, [2015-11-10 21:16.19|https://lsst-web.ncsa.illinois.edu/doxygen/xlink_master_2015_11_10_21.16.19/], most of them have vanished and we only provide links for {{ndarray}} and {{lsst::skymap}}. As of filing this issue, they were still missing from the [latest build|https://lsst-web.ncsa.illinois.edu/doxygen/x_masterDoxyDoc/]. Please bring them back!
| 1 |
819 |
DM-4311
|
11/11/2015 11:54:03
|
Oct. on-going support to Camera team in UIUC
|
Attend UIUC weekly meeting and give support as needed.
| 2 |
820 |
DM-4312
|
11/11/2015 11:55:52
|
Nov. on-going support to Camera team in UIUC
|
Attend UIUC weekly meeting and give support as needed.
| 2 |
821 |
DM-4323
|
11/11/2015 16:56:17
|
Replace fitsthumb in obs_subaru (port HSC-1196)
|
{{fitsthumb}} is now obsolete; all the functionality we need is available in {{afw}}. Further, we want to drop it as a dependency to make the job of integrating {{obs_subaru}} with CI easier.
| 1 |
822 |
DM-4329
|
11/12/2015 09:18:31
|
Coadd_utils tests should run and skip if afwdata is missing
|
Currently, the {{coadd_utils}} tests are completely skipped at the scons layer if afwdata can not be located. This is bad for two reasons: 1. Are there any tests that can be run even if afwdata is missing?. 2. When we switch to a proper test harness (e.g. DM-3901) an important metric is the number of tests executed compared with the number of tests skipped. Each test file (or even each test) should determine itself whether it should be skipped based on afwdata availability. This should not be a global switch.
| 1 |
823 |
DM-4344
|
11/13/2015 08:58:40
|
pipe_tasks 11.0-14-ga314014+5 fails a test on os/x 10.10.5
|
Running {code} SCONSFLAGS="-j 6 opt=3" eups distrib install pipe_tasks 11.0-14-ga314014+5 {code} on my laptop running ox/x 10.10.5 results in a test failure: {code} F. ====================================================================== FAIL: testEdge (__main__.interpolationTestCase) Test that we can interpolate to the edge ---------------------------------------------------------------------- Traceback (most recent call last): File "tests/testInterpImageTask.py", line 127, in testEdge validateInterp(miInterp, useFallbackValueAtEdge, fallbackValue) File "tests/testInterpImageTask.py", line 90, in validateInterp self.assertAlmostEqual(val0, fallbackValue, 6) AssertionError: 2.5515668 != 2.5515660972664267 within 6 places {code}
| 0 |
824 |
DM-4347
|
11/13/2015 10:31:46
|
dax_imgserv 2015_10.0 build error
|
{{2015_10.0}} has a build error under a current {{lsstsw/bin/deploy}} environment. Current speculation is that this is related to the conda version of numpy being upgraded to {{1.10.1}}.
| 1 |
825 |
DM-4349
|
11/13/2015 12:52:26
|
Fix publishing script async issue and add additional release notes.
|
Async command execution causes unpredictable and unreliable results. Switches to synchronous where possible. Also add additional description to the release notes.
| 2 |
826 |
DM-4360
|
11/16/2015 12:10:02
|
obs_subaru fails to compile after DM-3200
|
Due to atypical calls in {{obs_subaru}}'s {{hsc/SConscript}} to run scripts in the {{bin}} directory, {{obs_subaru}} fails to compile after the changes made in DM-3200.
| 0.5 |
827 |
DM-4362
|
11/16/2015 14:42:56
|
SuperTask phase 1 implementation
|
This story represents the implementation of the first part of the SuperTask framework design,
| 8 |
828 |
DM-4366
|
11/17/2015 09:19:38
|
Improve overscan correction for DECam raw data
|
Currently, the default overscan correction from IsrTask is used for processing DECam raw data. Overscan subtraction is done one amplifier at a time. However, a bias jump occurs due to the simultaneous readout of the smaller ancillary CCDs on DECam, some images show discontinuity in the y direction across one amplifier, as in the example screen shot. This ticket is to improve overscan correction for DECam data so to mitigate this discontinuity in the ISR processing. Arrangement of CCDs on DECam: http://www.ctio.noao.edu/noao/sites/default/files/DECam/DECamPixelOrientation.png h3. More details: There are 6 backplanes in the readout system, shown by the colors in DECamPixelOrientation.png. In raw data files, the CCD's backplane is noted in the header keyword "FPA". Examination of some images suggests that science CCDs on orange and yellow backplanes show bias jump at 2098 pixels from the y readout. That is the y size of the focus CCDs. h3. Actions: For CCDs on the affected backplanes, divide the array into two pieces at the jump location, and do overscan correction on the upper and lower pieces separately.
| 5 |
829 |
DM-4367
|
11/17/2015 09:46:41
|
Fix bug and add unit tests for PsfShapeletApprox
|
We discovered during this Sprint that this plugin was giving us faulty values for all the models except for SingleGaussian. I will fix that bug on this issue. Obviously, a better unit test would have caught this. I am adding a DoubleGaussian unit test, plus a test that the default models provide different results. Also a timing test for all the models, as we do not really have enough information about the performance of the shapelet approximation.
| 3 |
830 |
DM-4370
|
11/17/2015 12:03:47
|
Migrate testdata for DECam from disk to git-lfs
|
There is a package full with test data for the obs_decam. It is an eups package currently, but not a git repository. I would like to migrate that into our hosted git-lfs so the obs_decam package can be built by Jenkins. [~jmatt] I'm hoping you would be willing to handle this for me. If not I can find somebody else. Thanks!
| 0.5 |
831 |
DM-4373
|
11/17/2015 13:37:36
|
HSC backport: Add tract conveniences
|
This is a port of [HSC-715: Add tract conveniences|https://hsc-jira.astro.princeton.edu/jira/browse/HSC-715]. Here is the original description: {panel} In regular HSC survey processing, we'll run with a "rings" skymap to cover the entire survey area. meas_mosaic does not currently efficiently or conveniently iterate over tracts. For example: {code} mosaic.py /tigress/HSC/HSC --rerun price/cosmos --id field=SSP_UDEEP_COSMOS filter=HSC-R {code} Note the lack of a tract in the --id specifier — we want to iterate over all tracts. This is not currently possible. Instead, if we do not know the tract of interest (which the user should not be required to know), we have to iterate over all the tracts (e.g., tract=0..12345), but the user should not be required to know the number of tracts, and this is slow (and possibly memory-hungry: currently consuming 11GB on tiger3 just for 12 exposures). We need an efficient mechanism to iterate over all tracts by not specifying any tract on the command-line. {panel} As this functionality was added specifically for {{meas_mosaic}}, it was going to be ported as part of DM-2674. Due to a recent desire to use this functionality, this ticket will be ported here.
| 1 |
832 |
DM-4375
|
11/17/2015 17:01:00
|
A slimmer testdata_decam
|
Before this ticket, the files in {{testdata_decam}} are as they are downloaded from the archive. Some are MEF, and the total is a bit big (1.2G). I made a trimmed down version of {{testdata_decam}}, 109M and available here: [https://github.com/hchiang2/testdata_decam.git] I trimmed it down by only saving the primary HDU and one data HDU. However the unit tests in {{getRaw.py}} become less meaningful and I am not sure if we really want to do this, because some complexities of DECam files are about the MEF. {{getRaw.py}} tests Butler retrieval of multiple dataset types, in particular tests if the correct HDU is retrieved. Nonetheless, {{getRaw.py}} can pass (with branch u/hfc/DM-4375 of obs_decam) Note: the old testdata_decam still live on lsst-dev:/lsst8/testdata_decam/
| 2 |
833 |
DM-4381
|
11/18/2015 10:33:53
|
"SHUTOFF" nebula instances consume core/ramIt quota
|
It appears that halted/shutoff instances have no effect on resource quota usage. Eg: {code:java} $ openstack server list +--------------------------------------+-----------------------+-------------------+----------------------------------------+ | ID | Name | Status | Networks | +--------------------------------------+-----------------------+-------------------+----------------------------------------+ ... | 1956c6d0-8aec-4f42-a781-8a68fd10179d | el7-jhoblitt | SHUTOFF | LSST-net=172.16.1.171, 141.142.208.150 | ... $ nova absolute-limits +--------------------+--------+--------+ | Name | Used | Max | +--------------------+--------+--------+ | Cores | 141 | 150 | | FloatingIps | 0 | 10 | | ImageMeta | - | 128 | | Instances | 44 | 100 | | Keypairs | - | 100 | | Personality | - | 5 | | Personality Size | - | 10240 | | RAM | 342016 | 400000 | | SecurityGroupRules | - | 20 | | SecurityGroups | 1 | 10 | | Server Meta | - | 128 | | ServerGroupMembers | - | 10 | | ServerGroups | 0 | 10 | +--------------------+--------+--------+ $ openstack server delete 1956c6d0-8aec-4f42-a781-8a68fd10179d $ nova absolute-limits +--------------------+--------+--------+ | Name | Used | Max | +--------------------+--------+--------+ | Cores | 133 | 150 | | FloatingIps | 0 | 10 | | ImageMeta | - | 128 | | Instances | 43 | 100 | | Keypairs | - | 100 | | Personality | - | 5 | | Personality Size | - | 10240 | | RAM | 325632 | 400000 | | SecurityGroupRules | - | 20 | | SecurityGroups | 1 | 10 | | Server Meta | - | 128 | | ServerGroupMembers | - | 10 | | ServerGroups | 0 | 10 | +--------------------+--------+--------+ {code}
| 2 |
834 |
DM-4383
|
11/18/2015 15:33:46
|
Avoid restarting czar when empty chunk list changes
|
Currently czar caches empty chunk list after it reads the list from file. This complicates things when we need to update the list, integration test for example has to restart czar process after it loads new data to make sure that czar updates its cached list. Would be nice to have simpler mechanism to resetting cached list in czar without restarting it completely. It could be done via special query (abusing FLUSH for example) or via sending signal (problematic if czar runs remotely). This can be potentially useful even after we replace empty chunk list file with some other mechanism as I expect that cache will stay around even for that.
| 2 |
835 |
DM-4385
|
11/18/2015 17:35:42
|
Remove stringToAny from the utils package
|
I accidentally left stringToAny in Utils.cc when implementing DM-2635, even though I removed it from Utils.h and it is not used anywhere. Remove it and mark RFC-47 Done.
| 0 |
836 |
DM-4386
|
11/19/2015 10:51:31
|
Clean up ProcessCcdDecam
|
ProcessCcdDecam needs some cleanup: * {{run}} method simply delegates to the base class * {{propagateCalibFlags}} is a no-op (deliberately in {{cab69086}}, need to explore if the original problem still exists) * The config overrides (in config/processCcdDecam.py): ** Uses the catalog star selector, which isn't wise given the current heterogeneity of reference catalogs. ** Sets the background {{undersampleStype}} to {{REDUCE_INTERP_ORDER}}, which is the default.
| 1 |
837 |
DM-4387
|
11/19/2015 12:08:06
|
Skymap fails tests on testFindTractPatchList
|
When skymap is built and healpy is loaded, {{testFindTractPatchList}} fails with: {quote}====================================================================== FAIL: Test findTractPatchList ---------------------------------------------------------------------- Traceback (most recent call last): File "/Users/ctslater/lsstsw/build/skymap/tests/SkyMapTestCase.py", line 245, in testFindTractPatchList self.assertClosestTractPatchList(skyMap, [tractInfo.getCtrCoord()], tractId) File "/Users/ctslater/lsstsw/build/skymap/tests/SkyMapTestCase.py", line 284, in assertClosestTractPatchList tractPatchList = skyMap.findClosestTractPatchList(coordList) File "/Users/ctslater/lsstsw/build/skymap/python/lsst/skymap/baseSkyMap.py", line 146, in findClosestTractPatchList tractInfo = self.findTract(coord) File "/Users/ctslater/lsstsw/build/skymap/python/lsst/skymap/healpixSkyMap.py", line 97, in findTract index = healpy.ang2pix(self._nside, theta, phi, nest=self.config.nest) File "/Users/ctslater/lsstsw/stack/DarwinX86/healpy/1.8.1+12/lib/python/healpy-1.8.1-py2.7-macosx-10.5-x86_64.egg/healpy/pixelfunc.py", line 367, in ang2pix check_theta_valid(theta) File "/Users/ctslater/lsstsw/stack/DarwinX86/healpy/1.8.1+12/lib/python/healpy-1.8.1-py2.7-macosx-10.5-x86_64.egg/healpy/pixelfunc.py", line 110, in check_theta_valid assert (np.asarray(theta) >= 0).all() & (np.asarray(theta) <= np.pi + 1e-5).all(), "Theta is defined between 0 and pi" AssertionError: Theta is defined between 0 and pi{quote} This was missed during regular CI testing since healpy is not normally setup.
| 1 |
838 |
DM-4391
|
11/20/2015 10:32:09
|
Update testCoadds.py to accommodate changes in DM-2915
|
As of DM-2915, the config setting: {code}self.measurement.plugins['base_PixelFlags'].masksFpAnywhere = ['CLIPPED']{code} is set as a default for {{MeasureMergedCoaddSourcesTask}}. However, this *CLIPPED* mask plane only exists if a given coadd was created using the newly implemented {{SafeClipAssembleCoaddTask}}. If a coadd was built using {{AssembleCoaddTask}}, the *CLIPPED* mask plane is not present, so the above default must be overridden to exclude it when using {{MeasureMergedCoaddSourcesTask}}. This is the case for the mock coadd that is assembled in the unittest code in {{testCoadds.py}}, so the config needs to be set for the test to run properly. Note that the associated tests for {{SafeClipAssembleCoaddTask}} will be added as part of DM-4209.
| 0.5 |
839 |
DM-4395
|
11/20/2015 12:50:19
|
Update cmsd configuration for multi-node tests
|
A particular cmsd configuration parameter prefixes a hardcoded path for QueryResource, which needs to be removed. This seems to appear only during multi-node tests.
| 1 |
840 |
DM-4396
|
11/20/2015 12:55:17
|
ctrl_execute test fails to find test binary
|
There's a test in ctrl_execute that exercises the bin/dagIdInfo.py test program. Since the rewrite_shebang rewrites happen after the tests are executed, the test that looks for the bin/dagIdInfo.py binary fails, since it's not there before the tests execute.
| 0.5 |
841 |
DM-4398
|
11/20/2015 14:44:48
|
Fix regexp for gcc48
|
DM-2622 inttoduced some regexes which raise exceptions when built with gcc48 (e.g. on centos7). gcc48 support for regexes is generally broken, so it's better to replace that with boost regexes.
| 1 |
842 |
DM-4399
|
11/20/2015 15:15:58
|
ctrl_execute test fails under El Capitan
|
The test/testDagIdInfo.py because it runs a script from bin.src, rather than bin. This test needs to be rewritten.
| 1 |
843 |
DM-4402
|
11/23/2015 12:06:44
|
Experiment with light-weight SQL databases for secondary index
|
Evaluate the use of light-weight SQL, such as InnoDB, TokuDB (now Kyoto Cabinet), or RocksDB to create and manage the secondary index.
| 8 |
844 |
DM-4408
|
11/24/2015 10:02:54
|
HSC backport: fix memory leak in afw:geom:polygon
|
This is a backport of a bug fix that got included as part of [HSC-1311|https://hsc-jira.astro.princeton.edu/jira/browse/HSC-1311]. It is not related to that issue in particular, so is being ported here as an isolated bug fix. {panel} Original commit message: pprice@tiger-sumire:/tigress/pprice/hsc-1311/afw (tickets/HSC-1311=) $ git sub commit 55ad42d37fd1346f8ebc11e4077366dff4eaa87b Author: Paul Price <[email protected]> Date: Wed Oct 21 10:59:56 2015 -0400 imageLib: import polygonLib to prevent memory leak When doing "exposure.getInfo().getValidPolygon()", was getting: swig/python detected a memory leak of type 'boost::shared_ptr< lsst::afw::geom::polygon::Polygon > *', no destructor found. This was due to the polygonLib not being imported in imageLib. Using polygonLib in imageLib then requires adding polygon.h to all the swig interface files that use imageLib.i. examples/testSpatialCellLib.i | 1 + python/lsst/afw/cameraGeom/cameraGeomLib.i | 1 + python/lsst/afw/detection/detectionLib.i | 1 + python/lsst/afw/display/displayLib.i | 1 + python/lsst/afw/geom/polygon/Polygon.i | 1 + python/lsst/afw/image/imageLib.i | 2 ++ python/lsst/afw/math/detail/detailLib.i | 1 + python/lsst/afw/math/mathLib.i | 1 + 8 files changed, 9 insertions(+) {panel}
| 0.5 |
845 |
DM-4410
|
11/24/2015 12:45:53
|
Port detection task footprint growth changes from HSC
|
In hsc the default behavior for the detection task is to updated footprints with a footprint which has been grown by the psf. This behavior needs to be ported to LSST, as some source records have footprints which are too small. When making this change, the new default needs to be overridden for the calibrateTask, as it needs the original size. The port includes 8e9fb159a3227f848e0db1ecacf7819599f1c03b from meas_algorithms and 8bf0f4a44c924259d9eefbd109aadec7d839e0f2 from pipe_tasks
| 5 |
846 |
DM-4421
|
11/25/2015 10:35:24
|
faulty assumption about order dependency in ctrl_event unit tests
|
A recent change to daf_base uncovered a couple of faulty tests in ctrl_events that incorrectly assumes the order in which assumed the order in which data in a PropertySet would be received. We can't assume which order these values will be put into the property set, and therefore into the list retrieved from the Event object.
| 0.5 |
847 |
DM-4431
|
11/25/2015 16:47:08
|
setup mechanism to measure the portal performance
|
Setup the method to measure the query response time: # query sent to the data provider from client (from portal to FF server, to data provider) # result returns from data provider (FF server gets data, prepare the data for portal) # result displayed in the client Setup the method to measure image preparation in Firefly server: * measure the time needed to prepare the image (generate the image in PNG or other suitable format) for client display. * Track/identify the reasons if performance is not satisfactory * Make plan for design/code change Setup the method to measure image/plot rendering in Firefly client (portal): * Measure the time needed to render the image after the data received by the client for display. * Measure the time needed to generate the 2D plot after the data is received by the client * Track/identify the reasons if the performance is not satisfactory * Make plan for design/code change
| 40 |
848 |
DM-4438
|
11/26/2015 08:32:14
|
Replace sed with stronger template engine in docker scripts
|
Dockerfile are generated using templates and sed, this should be strengthened.
| 2 |
849 |
DM-4440
|
11/26/2015 08:41:45
|
Remove QSW_RESULTPATH and XROOTD_RUN_DIR if useless
|
These parameters may be useless (see DM-4395). If yes they can be removed to simplify configuration procedure.
| 2 |
850 |
DM-4443
|
11/30/2015 06:58:16
|
Please document the --rerun option
|
DM-3371 adds the {{--rerun}} option to command line tasks. The help for this option reads: {quote} rerun name: sets OUTPUT to ROOT/rerun/OUTPUT; optionally sets ROOT to ROOT/rerun/INPUT {quote} While essentially correct, that's not particularly helpful in understanding what's actually going on here. A motivation and description of this functionality is available in RFC-95: please ensure that, or some variation of it, is included in the stack documentation.
| 1 |
851 |
DM-4451
|
11/30/2015 11:42:24
|
F17 Qserv Disconnected Queries
|
* Design and implement *basic* system for determining whether particular query is synchronous or asynchronous. The complete version will come through DM-1490. Note that this work is related to shared scans (e.g., we need to know what scans we have running) * Design SQL API for starting and interacting with async queries. * Modify Qserv to support async queries (starting, getting status, retrieving results) Note, async queries are indirectly related to authentication (users should not see each other' async queries). Deliverable: Qserv that accepts and executes queries asynchronously, and allows users to retrieve results.
| 20 |
852 |
DM-4454
|
11/30/2015 12:35:47
|
Fix multiple patch catalog sorting for forcedPhotCcd.py
|
{{forcedPhotCcd.py}} is currently broken due to the requirement of the {{lsst.afw.table.getChildren()}} function that the *SourceCatalog* is sorted by the parent key (i.e. {{lsst.afw.table.SourceTable.getParentKey()}}). This occurs naturally in the case of *SourceCatalogs* produced by the detection and deblending tasks, but it may not be true when concatenating multiple such catalogs. This is indeed the case for {{forcedPhotCcd.py}} as a given CCD can be overlapped by multiple patches, thus requiring a concatenation of the reference catalogs of all overlapping patches. There two places in the running of {{forcedPhotCcd.py}} where calls to {{getChildren()}} can cause a failure: one in the {{subset()}} function in {{references.py}}, and the other in the {{run}} function of *SingleFrameMeasurementTask* in {{sfm.py}}.
| 2 |
853 |
DM-4457
|
11/30/2015 15:55:02
|
Investigate MemSQL
|
Take a look at the MemSQL distributed database.
| 8 |
854 |
DM-4503
|
12/01/2015 22:43:18
|
FITS Visualizer porting: selecting points of catalog from image view, showing selected points
|
able to draw a rectangle on the image, and select the catalog entries overlaid on the image
| 5 |
855 |
DM-4504
|
12/01/2015 22:44:20
|
FITS Visualizer porting: Image Select Panel/Dialog
|
Converting the image select dialog/panel is a very big job and should be to be broken up into several tickets: Each ticket should reference this ticket as the base. Panel includes the following: * issa, 2mass, wise, dss, sdss tabs * file upload tab, upload widget might have to be written * url tab * blank image tab * target info reusable widget * 3 color support - any panel should show for 3 times, for read, green, and blue in 3 color mode * must be able to appear in a panel or dialog * must add or modify a plot * Allow to create version with most or less than the standard tabs. example - see existing wise 3 color or finder chart 3 color * A plot might need to be tied to specific type of image select dialog, we need a way to tie a plotId to and non-standard image select panel.
| 1 |
856 |
DM-4510
|
12/02/2015 10:55:36
|
makeDocs uses old style python
|
{{makeDocs}} is written in python 2.4 style. This ticket is for updating it to python 2.7.
| 0.5 |
857 |
DM-4511
|
12/02/2015 11:39:08
|
Improve reStructuredText documentation
|
Enhance docs by covering - Images as links - Table spans - Abbreviations - :file: semantics, etc.
| 2 |
858 |
DM-4515
|
12/02/2015 13:48:28
|
Flag out the glowing edges of DECam CCDs
|
Pixels near the edges of the DECam CCDs are bigger/brighter and correcting them is not trivial. One way to move forward is to mask them out. DESDM and CP mask 15 pixels on each edge. The cut was later raised to 25 pixels, with the inner 10 pixels flagged as SUSPECT.
| 5 |
859 |
DM-4523
|
12/03/2015 03:12:05
|
Fix startup.py inside Docker container
|
qserv tag should be replace with qserv_latest
| 0.5 |
860 |
DM-4529
|
12/03/2015 10:17:24
|
Compilation errors from CLang (Apple LLVM 7.0) in XCode 7 on MacOSX
|
Compiling on MacOSX Yosemite with XCode 7, a number of files fail compilation. ---- {{core/modules/util/EventThread.h,cc}} fails because {{uint}} is used as a data type. This is non-standard (though some compilers support it), and should be replaced with {{unsigned int}}. ---- {{core/modules/wbase/SendChannel.h,cc}} fails because {{#include <functional>}} is missing. ---- {{core/modules/wsched/ChunkState.cc}} fails because {{#include <iostream>}} is missing. ---- {{build/qmeta/qmetaLib_wrap.cc}} (generated by SWIG) fails with many errors because the {{typedef unsigned long int uint64_t}} included in {{qmetaLib.i}} conflicts with MacOSX's typedef of it as {{unsigned long long}}.
| 1 |
861 |
DM-4534
|
12/03/2015 14:56:34
|
shellcheck linting of lsstsw bash scripts
|
This issue is to recover a branch from DM-4113 that was not merged due to issues with installing shellcheck under travis.
| 1 |
862 |
DM-4544
|
12/04/2015 16:27:33
|
Revisit short and long term plans for butler
|
Revisit short and long term requirements and needs and capture it through stories.
| 5 |
863 |
DM-4556
|
12/05/2015 16:54:40
|
Fix docker workflow
|
Some issues where discovered while trying to package DM-2699 in Docker (for IN2P3 cluster deployment), they're fixed here. - apt-get update times out: why? - git clone then pull is too weak (if building the first clone fails, pull never occurs) => step merged - eupspkg -er build creates lib/python in /qserv/stack/.../qserv/... and next install can't remove it for unknow reason => build and install merged.
| 2 |
864 |
DM-4564
|
12/07/2015 09:55:38
|
Convert basic table functionalities to JS.
|
Task includes server-side json conversion, data modeling, and a simple React table for presentation.
| 20 |
865 |
DM-4572
|
12/07/2015 10:46:58
|
Table (JS): table options
|
This task is composed of: - adding table options panel to TablePanel. - providing features: - show/hide units in header - show/hide columns, reset to defaults, etc - page size
| 5 |
866 |
DM-4574
|
12/07/2015 10:49:37
|
Table (JS): text view
|
This task is composed of: - adding text view option to TablePanel
| 2 |
867 |
DM-4576
|
12/07/2015 11:03:46
|
XY Scatter Plot (JS)
|
Implement basic scatter plot widget using react-highcharts library
| 8 |
868 |
DM-4580
|
12/07/2015 11:18:36
|
XY Plot view of a table (JS) - Toolbar
|
Toolbar, which toggles plot options, selection and filter buttons Extra: - handling zoom from the toolbar rather than using built-in zoom - ability to switch between histogram and scatter plot view
| 8 |
869 |
DM-4582
|
12/07/2015 11:30:51
|
XY Plot View of a table (JS) - selection support
|
Show/change selected/highlighted points. Ideally, this should be done without redrawing the whole plot.
| 8 |
870 |
DM-4583
|
12/07/2015 11:38:15
|
SUIT: search returning images in a directory
|
- Create a sample search processor, which returns images in a given directory. - It should be using an external python task - Update search form configuration to use this search processor to return image metadata
| 2 |
871 |
DM-4591
|
12/07/2015 13:05:45
|
GWT conversion: System notifications
|
This task is composed of: - adding notification panel to the application - convert server-side code to use messaging for notifications - use messaging on client-side to handle notifications - creating action, action creator, and reducing functions - depends on DM-4578 Integrate websocket messaging into flux
| 3 |
872 |
DM-4596
|
12/07/2015 14:15:02
|
Remove deprecated versions of warpExposure and warpImage
|
afw.math supports two templated variants of warpExposure and warpImage, one that takes a warping control and the other which does not. The latter have been deprecated for a long time and are no longer used. I think it is time to remove them.
| 1 |
873 |
DM-4603
|
12/08/2015 10:37:07
|
sconsUtils tests should depend on shebang target
|
Some tests rely on code in the {{bin}} directory. Whilst these tests have been modified to use {{bin.src}} the general feeling is that the test code should be able to rely upon the {{shebang}} target having been executed before they are run.
| 0.5 |
874 |
DM-4609
|
12/09/2015 13:15:06
|
Partition package should use the standard package layout
|
The partition package does not build on OS X El Capitan because the package is not laid out in the standard manner and whilst {{sconsUtils}} is used most of the default behaviors are over-ridden. This means that fixes implemented for DM-3200 do not migrate over to {{partition}}. I think the best approach would be to reorganize the package so that it does build in the normal way.
| 1 |
875 |
DM-4617
|
12/09/2015 19:44:22
|
Send all chunk-queries to primary copy of the chunk
|
We are planning to distribute chunks / replicas across worker nodes such that each node will have a mix of primary copies for some chunks, and backup copies for some chunks. While doing shared scan, we are going to always rely on the primary chunks (e.g., all queries that need a given chunk should be sent to the same machine so that we read that chunks only once on one node). This story involves tweaking xrootd to ensure we don't send chunk-queries to nodes hosting non-primary copies.
| 5 |
876 |
DM-4631
|
12/11/2015 08:58:10
|
Create IDL pipeline workflow for DRP processing - processCcdDecam
|
For the verification datasets work we need the ability to run a dataset all the way through DRP processing that takes advantage of many cores (orchestration), keeps track of successful and failed dataIDs, creates individual log files, and creates helpful QA plots/metrics/webpages. Nidever will use his PHOTRED IDL workflow and rewrite it for the stack. The first step is processCcdDecam.
| 5 |
877 |
DM-4643
|
12/14/2015 11:47:13
|
Add utility function to handle client-side download requests.
|
Create utility function to handle client-side download requests. It needs to be done in a way that does not mess with history and current page state.
| 1 |
878 |
DM-4648
|
12/14/2015 16:25:39
|
Support sqlalchemy use with qserv
|
When one tries to connect to qserv using sqlalchemy there is an exception generated currently: {noformat} $ python -c 'import sqlalchemy; sqlalchemy.create_engine("mysql+mysqldb://[email protected]:4040/test").connect()' /u2/salnikov/STACK/Linux64/sqlalchemy/2015_10.0/lib/python/SQLAlchemy-1.0.8-py2.7-linux-x86_64.egg/sqlalchemy/engine/default.py:298: SAWarning: Exception attempting to detect unicode returns: InterfaceError("(_mysql_exceptions.InterfaceError) (-1, 'error totally whack')",) "detect unicode returns: %r" % de) Traceback (most recent call last): File "<string>", line 1, in <module> File "/u2/salnikov/STACK/Linux64/sqlalchemy/2015_10.0/lib/python/SQLAlchemy-1.0.8-py2.7-linux-x86_64.egg/sqlalchemy/engine/base.py", line 2018, in connect return self._connection_cls(self, **kwargs) File "/u2/salnikov/STACK/Linux64/sqlalchemy/2015_10.0/lib/python/SQLAlchemy-1.0.8-py2.7-linux-x86_64.egg/sqlalchemy/engine/base.py", line 72, in __init__ if connection is not None else engine.raw_connection() File "/u2/salnikov/STACK/Linux64/sqlalchemy/2015_10.0/lib/python/SQLAlchemy-1.0.8-py2.7-linux-x86_64.egg/sqlalchemy/engine/base.py", line 2104, in raw_connection self.pool.unique_connection, _connection) File "/u2/salnikov/STACK/Linux64/sqlalchemy/2015_10.0/lib/python/SQLAlchemy-1.0.8-py2.7-linux-x86_64.egg/sqlalchemy/engine/base.py", line 2078, in _wrap_pool_connect e, dialect, self) File "/u2/salnikov/STACK/Linux64/sqlalchemy/2015_10.0/lib/python/SQLAlchemy-1.0.8-py2.7-linux-x86_64.egg/sqlalchemy/engine/base.py", line 1405, in _handle_dbapi_exception_noconnection exc_info File "/u2/salnikov/STACK/Linux64/sqlalchemy/2015_10.0/lib/python/SQLAlchemy-1.0.8-py2.7-linux-x86_64.egg/sqlalchemy/util/compat.py", line 199, in raise_from_cause reraise(type(exception), exception, tb=exc_tb) File "/u2/salnikov/STACK/Linux64/sqlalchemy/2015_10.0/lib/python/SQLAlchemy-1.0.8-py2.7-linux-x86_64.egg/sqlalchemy/engine/base.py", line 2074, in _wrap_pool_connect return fn() File "/u2/salnikov/STACK/Linux64/sqlalchemy/2015_10.0/lib/python/SQLAlchemy-1.0.8-py2.7-linux-x86_64.egg/sqlalchemy/pool.py", line 318, in unique_connection return _ConnectionFairy._checkout(self) File "/u2/salnikov/STACK/Linux64/sqlalchemy/2015_10.0/lib/python/SQLAlchemy-1.0.8-py2.7-linux-x86_64.egg/sqlalchemy/pool.py", line 713, in _checkout fairy = _ConnectionRecord.checkout(pool) File "/u2/salnikov/STACK/Linux64/sqlalchemy/2015_10.0/lib/python/SQLAlchemy-1.0.8-py2.7-linux-x86_64.egg/sqlalchemy/pool.py", line 480, in checkout rec = pool._do_get() File "/u2/salnikov/STACK/Linux64/sqlalchemy/2015_10.0/lib/python/SQLAlchemy-1.0.8-py2.7-linux-x86_64.egg/sqlalchemy/pool.py", line 1060, in _do_get self._dec_overflow() File "/u2/salnikov/STACK/Linux64/sqlalchemy/2015_10.0/lib/python/SQLAlchemy-1.0.8-py2.7-linux-x86_64.egg/sqlalchemy/util/langhelpers.py", line 60, in __exit__ compat.reraise(exc_type, exc_value, exc_tb) File "/u2/salnikov/STACK/Linux64/sqlalchemy/2015_10.0/lib/python/SQLAlchemy-1.0.8-py2.7-linux-x86_64.egg/sqlalchemy/pool.py", line 1057, in _do_get return self._create_connection() File "/u2/salnikov/STACK/Linux64/sqlalchemy/2015_10.0/lib/python/SQLAlchemy-1.0.8-py2.7-linux-x86_64.egg/sqlalchemy/pool.py", line 323, in _create_connection return _ConnectionRecord(self) File "/u2/salnikov/STACK/Linux64/sqlalchemy/2015_10.0/lib/python/SQLAlchemy-1.0.8-py2.7-linux-x86_64.egg/sqlalchemy/pool.py", line 454, in __init__ exec_once(self.connection, self) File "/u2/salnikov/STACK/Linux64/sqlalchemy/2015_10.0/lib/python/SQLAlchemy-1.0.8-py2.7-linux-x86_64.egg/sqlalchemy/event/attr.py", line 246, in exec_once self(*args, **kw) File "/u2/salnikov/STACK/Linux64/sqlalchemy/2015_10.0/lib/python/SQLAlchemy-1.0.8-py2.7-linux-x86_64.egg/sqlalchemy/event/attr.py", line 256, in __call__ fn(*args, **kw) File "/u2/salnikov/STACK/Linux64/sqlalchemy/2015_10.0/lib/python/SQLAlchemy-1.0.8-py2.7-linux-x86_64.egg/sqlalchemy/util/langhelpers.py", line 1312, in go return once_fn(*arg, **kw) File "/u2/salnikov/STACK/Linux64/sqlalchemy/2015_10.0/lib/python/SQLAlchemy-1.0.8-py2.7-linux-x86_64.egg/sqlalchemy/engine/strategies.py", line 165, in first_connect dialect.initialize(c) File "/u2/salnikov/STACK/Linux64/sqlalchemy/2015_10.0/lib/python/SQLAlchemy-1.0.8-py2.7-linux-x86_64.egg/sqlalchemy/dialects/mysql/base.py", line 2626, in initialize default.DefaultDialect.initialize(self, connection) File "/u2/salnikov/STACK/Linux64/sqlalchemy/2015_10.0/lib/python/SQLAlchemy-1.0.8-py2.7-linux-x86_64.egg/sqlalchemy/engine/default.py", line 256, in initialize self._check_unicode_description(connection): File "/u2/salnikov/STACK/Linux64/sqlalchemy/2015_10.0/lib/python/SQLAlchemy-1.0.8-py2.7-linux-x86_64.egg/sqlalchemy/engine/default.py", line 343, in _check_unicode_description ]).compile(dialect=self) File "/u2/salnikov/STACK/Linux64/mysqlpython/1.2.3.lsst1/lib/python/MySQL_python-1.2.3-py2.7-linux-x86_64.egg/MySQLdb/cursors.py", line 174, in execute self.errorhandler(self, exc, value) File "/u2/salnikov/STACK/Linux64/mysqlpython/1.2.3.lsst1/lib/python/MySQL_python-1.2.3-py2.7-linux-x86_64.egg/MySQLdb/connections.py", line 36, in defaulterrorhandler raise errorclass, errorvalue sqlalchemy.exc.InterfaceError: (_mysql_exceptions.InterfaceError) (-1, 'error totally whack') {noformat} The reason for that is that sqlalchemy generate few SELECT queries to figure out unicode support by the engine, and those selects are passed to qserv which cannot parse them. Here is the list of SELECTs which appears in proxy log: {code:sql} SELECT CAST('test plain returns' AS CHAR(60)) AS anon_1 SELECT CAST('test unicode returns' AS CHAR(60)) AS anon_1 SELECT CAST('test collated returns' AS CHAR CHARACTER SET utf8) COLLATE utf8_bin AS anon_1 SELECT 'x' AS some_label {code}
| 3 |
879 |
DM-4652
|
12/14/2015 17:10:24
|
CI debugging
|
diagnosing build failures and refreshing build slaves
| 1 |
880 |
DM-4656
|
12/15/2015 00:12:29
|
Port code style guidelines to new DM Developer Guide
|
Verbatim port of DM Coding style guidelines to Sphinx doc platform from Confluence. - https://confluence.lsstcorp.org/display/LDMDG/DM+Coding+Style+Policy - https://confluence.lsstcorp.org/display/LDMDG/Python+Coding+Standard - https://confluence.lsstcorp.org/pages/viewpage.action?pageId=16908666 and contents I’m unclear whether these pages should be included: - https://confluence.lsstcorp.org/pages/viewpage.action?pageId=20283399 (C++ ‘using’) - https://confluence.lsstcorp.org/pages/viewpage.action?pageId=20284190 (how to use C++ templates) - https://confluence.lsstcorp.org/pages/viewpage.action?pageId=20283399 (C++11/14; which should seem to belong in the code style guide) Any temptation to amend and update the style guideline content will be avoided.
| 5 |
881 |
DM-4667
|
12/15/2015 14:25:03
|
Improve sphgeom documentation
|
Per RFC-117, the sphgeom package needs decent overview documentation, linked from the top-level README.md. The doxygen output should also be reviewed.
| 2 |
882 |
DM-4677
|
12/15/2015 18:33:13
|
Design Interfaces for Memory Management for Shared Scans
|
Part of the shared scans involve memory management - a system that will be used by Qserv that will manage memory allocation / pin chunks in memory. This story involves designing the API between Qserv and the memory management system.
| 8 |
883 |
DM-4679
|
12/16/2015 09:56:40
|
work with database team to exercise all the APIs for data access (F16)
|
SUI will continue to work with database team to exercise all the APIs for data access. All known issues should be worked out in S16 cycle.
| 40 |
884 |
DM-4688
|
12/16/2015 19:25:12
|
Changed the implementation of HistogramProcessorTest due to the minor change about the algorithm in the HistogramProcessor
|
In Histogram, when the data points fall on the bin edges, the following rules are used: # For each bin, it contains the data points fall inside the bin and the data point fall on the left edge. For example, if binSize=2, the bin[0] is in the range of [0,2]. The data value 0 is in bin[0] . # For each bin, the data point falls on the right edge is not included in the number point count. For example if binSize=2, the bin[0] is having the range of [0,2]. The data value 0 is in bin[0] but the data value 2 is not in the bin[0]. # For the last bin, the data points fall inside the bin or fall on the left or right bin are counted as the number of bin points. The last rule is newly introduced.
| 2 |
885 |
DM-4704
|
12/17/2015 17:28:46
|
Qserv integration tests fail on CentOS7 with gcc 4.8.5
|
The version of gcc that ships with CentOS7, {{gcc (GCC) 4.8.5 20150623 (Red Hat 4.8.5-4)}}, appears to miscompile the qserv worker source in a way that makes it impossible to actually run queries. Installing {{devtoolset-3-toolchain}} and {{devtoolset-3-perftools}} via {{yum}} to get gcc 4.9 resolves the issue.
| 2 |
886 |
DM-4705
|
12/17/2015 19:54:09
|
qdisp/testQDisp fails with mariadb
|
Fabrice fried to build qserv with mariadb and it caused failure in one of the unit test: qdisp/testQDisp with the message: {noformat} pure virtual method called terminate called without an active exception {noformat} Runnig it with GDB it' obvious that there is a problem with resource lifetime management in qdisp/testQDisp.cc. The problem is that XrdSsiSessionMock is destroyed sooner than other objects that use it. One way to resolve this problem is to instantiate XrdSsiSessionMock earlier than other objects that use it (to reverse the order of destructors), possibly make it a global instance. Big mystery here is how mariadb could trigger this interesting behavior and why did not we see this earlier.
| 1 |
887 |
DM-4706
|
12/17/2015 22:28:39
|
Rerun and create a repository for CFHT astrometry test.
|
Understand, re-rerun, and recreate clean version of [~boutigny] 's CFHT astrometry test for the astrometry RMS for two sample CFHT observations. This test is on the NCSA machines in /lsst8/boutigny/valid_cfht Create a repository for this test with an eye toward it becoming integrated in a validation suite for the stack.
| 1 |
888 |
DM-4708
|
12/17/2015 22:37:20
|
Integrate astrometry test into SDSS demo lsst_dm_stack_demo
|
Incorporate the astrometry test as an optional component in lsst_dm_stack_demo. This is chosen because lsst_dm_stack_demo currently serves as the very loose stack validation and understanding how to do astrometric repeatibility testing in this demo will help explore how it would make sense to put in a fuller CFHT validation test of the DM stack.
| 0.5 |
889 |
DM-4710
|
12/18/2015 08:22:48
|
host identification info needs to be part of log message
|
The EventAppender needs to add host identification (host/process/id) information to the log message it transmits. This was inadvertently left out.
| 3 |
890 |
DM-4711
|
12/18/2015 09:51:04
|
Edit testdata_cfht to pass obs_cfht unit tests
|
This ticket covers the first half of the issues in DM-2917. {{testdata_cfht}} was left unedited while some past changes in {{obs_cfht}} {{MegacamMapper}} required coordinated changes. The goal of this ticket is to simply pass the unit tests currently in {{obs_cfht}}.
| 1 |
891 |
DM-4716
|
12/20/2015 12:18:20
|
Track down reason for slow performance when running many jobs of processCcdDEcam on bambam
|
During the processing of the COSMOS data for the verification dataset work I ran many jobs of processCcdDecam.py on the new linux server, bambam. The performance was very slow, 4x longer than running a single job at a time. Figure out what is going on.
| 2 |
892 |
DM-4722
|
12/21/2015 13:43:34
|
File tickets for list of stack deficiencies and suggested upgrades
|
K-T suggested that I take my list of "stack deficiencies and suggested improvements" [https://confluence.lsstcorp.org/display/SQRE/Stack+Deficiencies+and+Suggested+Upgrades] on confluence and (with Tim J.'s help) create tickets for each item (as much as possible) so that the work could be scheduled.
| 3 |
893 |
DM-4728
|
12/22/2015 06:22:03
|
Doxygen package fails to build with flex 2.6
|
To wit: {code} $ flex --version flex 2.6.0 $ bash newinstall.sh LSST Software Stack Builder [...stuff...] eups distrib: Failed to build doxygen-1.8.5.eupspkg: Command: source /Users/jds/Projects/Astronomy/LSST/stack/eups/bin/setups.sh; export EUPS_PATH=/Users/jds/Projects/Astronomy/LSST/stack; (/Users/jds/Projects/Astronomy/LSST/stack/EupsBuildDir/DarwinX86/doxygen-1.8.5/build.sh) >> /Users/jds/Projects/Astronomy/LSST/stack/EupsBuildDir/DarwinX86/doxygen-1.8.5/build.log 2>&1 4>/Users/jds/Projects/Astronomy/LSST/stack/EupsBuildDir/DarwinX86/doxygen-1.8.5/build.msg exited with code 252 $ grep error /Users/jds/Projects/Astronomy/LSST/stack/EupsBuildDir/DarwinX86/doxygen-1.8.5/build.log commentscan.l:1064:55: error: use of undeclared identifier 'yy_current_buffer' commentscan.l:1126:58: error: use of undeclared identifier 'yy_current_buffer' {code} Builds fine using {{flex 2.5.35 Apple(flex-31)}}.
| 1 |
894 |
DM-4729
|
12/23/2015 10:42:51
|
HSC backport: Add functions to generate 'unpacked matches' in a Catalog
|
The qa analysis script under development (see DM-4393) calls to HSC {{hscPipeBase}}'s [matches.py|https://github.com/HyperSuprime-Cam/hscPipeBase/blob/master/python/hsc/pipe/base/matches.py] which adds functions to generate "unpacked matches" in a Catalog (and vice versa). It will be ported into {{lsst.afw.table}}. The port includes following HSC commits: *Add functions to generate 'unpacked matches' in a Catalog.* https://github.com/HyperSuprime-Cam/hscPipeBase/commit/210fcdc6e1d19219e2d9365adeefd9289b2e1186 *Adding check to prevent more obscure error.* https://github.com/HyperSuprime-Cam/hscPipeBase/commit/344a96de741cd5aafb5e368f7fa59fa248305af5 *Some little error handling helps.* https://github.com/HyperSuprime-Cam/hscPipeBase/commit/61cc053b873d42802581adff8cbbdb52a348879e (from branch: {{stage-ncsa-3}}) *matches: add ArrayI to list of field types that require a size* https://github.com/HyperSuprime-Cam/hscPipeBase/commit/d4ccd11d8afbcdd9cf0b35eba948cca4b5d09ba5 (from branch {{tickets/HSC-1228}}) Please also include a unittest.
| 3 |
895 |
DM-4731
|
12/23/2015 15:56:58
|
Add labels to qa analysis plots for better interpretation
|
The plots output by the qa analysis script (see DM-4393) currently do not display any information regarding the selection/rejection criteria used in making the figures and computing the basic statistics. This includes magnitude and clipping thresholds. This information should be added to each plot such that the figures can be interpreted properly.
| 2 |
896 |
DM-4732
|
12/23/2015 21:52:31
|
butler parentSearch incorrectly returns '[]' instead of 'None' in one case.
|
The contract for `daf.butlerUtils.CameraMapping.parentSearch` states that `None` will be returned if no matches are found. But in one of the options ` if os.path.realpath(pathPrefix) != os.path.realpath(root):` the code says `return []` instead of `return None`. This is then not caught properly by `daf.butlerUtils.mapping.map`, which was checking just for the return value `is not None`. The behavior of parentSearch is clearly an error (it's againt the documentation for the function). I would also suggest that `daf.butlerUtils.mapping.map` should check more robustly for `if newPath:` rather than `if newPath is not None:`. Thus we explicitly set up this if/else to return `None` if `not paths` instead of `[]`.
| 0 |
897 |
DM-4734
|
12/30/2015 08:10:24
|
afw fails to build on a machine with many cores
|
The afw package does not build reliably (if at all) on a linux box at UW ("magneto", which has 32 cores and 128 Gb of RAM). The failure is that some unit tests fail with the following error: {code} OpenBLAS: pthread_creat error in blas_thread_init function. Error code:11 {code} For the record, /usr/include/bits/local_lim.h contains this: {code} /* The number of threads per process. */ #define _POSIX_THREAD_THREADS_MAX 64 /* We have no predefined limit on the number of threads. */ #undef PTHREAD_THREADS_MAX {code} It appears that the build system is trying to use too many threads when building afw, which presumably means it is trying to use too many cores. According to [~mjuric] the package responsible for this is {{eupspkg}}, and it tries to use all available cores. A workaround suggested by [~mjuric] is to set environment variable {{EUPSPKG_NJOBS}} to the max number of cores wanted. However, I suggest we fix our build system so that setting this variable is unnecessary. I suggest we hard-code an upper limit for now, though fancier logic is certainly possible. A related request is to document the environment variables that control our build system. I searched for {{NJOBS}} on confluence and found nothing.
| 0.5 |
898 |
DM-4752
|
01/06/2016 15:47:54
|
Build on Mac very slow due to running fc-list
|
Builds on MacOS 10.11 have been painfully slow (for instance 23 minutes to build {{afw}} instead of the more typical 12 minutes, 30 minutes to rebuild {{ip_diffim}}, 20 minutes to rebuild {{meas_astrom}}) and much of this time is spent running fc-list in 8 cores. I suspect {{matplotlib}} is triggering this process, but I have not verified it. I see this using lsstsw to build a fresh stack and with manual builds. A workaround is to repeatedly kill {{fc-list}}, e.g. with this bash script: {code} while sleep 1; do pkill fc-list; done {code} I checked my fonts with Apple's Font Book and found a few dozen with "minor errors" that I deleted, but nothing serious. I'm still seeing the problem.
| 0.5 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.