Activity for Cliff Kotnik
-
Cliff Kotnik committed [r1844]
Correct workaround for loading resources running in Eclipse
-
No, but I do not use that feature much. Cliff Kotnik clkotnik@att.net
-
I can reproduce the issue with the period dialog, but not the new star from file. So I think it is something in the difference between these two.
-
I guess we never got Sara to add the doc to the web.
-
I'm afraid I do not understand the request and the forum link above gives me access denied.
-
So are we trying to continue to make webstart work?
-
Cannot repro
-
No luck on repro
-
So make the phase plot dialog behave like the new star from file dialog.
-
I think it makes sense to pause at this point and see if this is a problem for users.
-
Sounds good enough to me
-
I tired to reproduce Brad's issue with variations of full screen or not, JNLP or local, etc. I cannot reproduce this on Windows with 2.21.1. Steps: Open Vstar (local or jnlp) and make it full screen or not. VStar->New star from file and leave dialog open Windows task bar bring Firefox to front and focus Windows task bar bring Vstar to front and focus Result both VStar dialog and main window appear with dialog on top
-
I'm not sure now if this ever was a problem. I did not have the MTBS plugin dialog open while I made the selections. When I do that it works with the current build. Sorry for the confusion.
-
I have pulled the latest update from SF and done a build of the main program and all plugins. I still see no observations in this plugin after selecting. In this case, I selected by double clicking in the plot window to get the details dialog. I tried this build on my ubuntu machine where it failed as well. I tried this on my windows machine with the current 2.20 release and it failed as well.
-
Sorry to spam this ticket, but I notice the Hipparcos plugin generates BJD(TT) on the time axis which still behaves the same as BJD in observations details and JD to HJD plugin.
-
And JD to HJD converter offers to convert to HJD which would not make sense.
-
Additional note - Loaded TESS file and plot time axis shows Time (BJD) which I think we agree goes away. However, observations details for these points show simply "Julian Date" instead of "Barycentric Julian Date".
-
Yes 1 & 2 for this release
-
One way to look at this is to consider any user will be working to get all observations on the same time base. BJD will become more common. So I think the most useful thing is to provide a way for the user to see where things stand. Max's suggestion in Slack seems good. There could also be a column in the observations list. So option 1 for the plot pane is ok and avoids future maintenance. Regardless, the documentation needs a tweak.
-
Plot Time Axis not Changed to HJD After JD 2 HJD
-
Mean Time Between Selections is not registering selections
-
Update to Maksym Pyatnytskyy's version 2.2
-
Replace the Kepler/TESS plugin documentation with Maksym Pyatnytskyy's version.
-
This behavior exists in 2.20. So if it has not surfaced before it may not be worth worrying about. I only exercised it because it was described in the manual.
-
I do not understand why there is an exception. What is it?
-
It seems to me that both URLs are valid. If I use nslookup to find the IP or aavso.org and www.aavso.org they resolve to the same thing. If I use curl to fetch the file they both work. I tried to see a redirect in the fetch for the URL without www that might through the Java class off and did not see one.
-
It appears that my Firefox browser on Windows provides a URL without "www" as https://aavso.org/sites/default/files/tmp/aavsodata_5e9ae4a49963f.txt However, on ubunut my Firefox browser provides the www. https://www.aavso.org/sites/default/files/tmp/aavsodata_5e9ae4a49963f.txt The former fails in VStar while the latter works.
-
Both of those seem ok to me. I was surprise how different the layout manager operates for Windows even swapping left and right sometimes.
-
User Manual for 2.21.0 Needs Some Clarification
-
Version of VStar is 2.21.0 snapshot 2020-04-14
-
Load AID download file from URL fails
-
Removed Vela UI directory
-
Create a TESS plug-in
-
I altered the code to pull the FITS header value for the ‘TELESCOP’ keyword and use it for the band. This will give “Kepler” for Kepler and K2 light curves. It will give “TESS” for TESS light curves. If a new mission launches a telescope with a different name, that name should get placed in the band. Should a file without this FITS header be imported, the band will default to “MAST”. I feel this completes the plugin and I am closing this ticket.
-
Updated Kepler/TESS pluging to put telescope name in band
-
Sara, No problem publishing the revised plugin. I will experiment with a general way to handle the band and post my results here. Might be a week. Cliff
-
I completely missed the fact that this is a custom series created on the fly. I looked at the FITS HDU header in the same HDU as the BJDREFI and BJDREFF values are contained, there is a TELESCOP header that contains either Kepler or TESS. Perhaps the series could be created using the value from the header? Or would it be too late in the processing to instantiate the series at that point?
-
The revised Kepler/TESS plugin was deployed. Two observations by a beta tester seem worth discussing. 1) While the lightcurve shows BJD, the date column heading in the observations list reads "Julian Day". This is also true for HJD. Is it worth changing this? 2) The band for both TESS and Kepler reads "Kepler". It appears Kepler is one of the standard VStar bands and has this description. Since neither Kepler nor TESS are a real passband, only a conversion of flux to magnitude with an arbitrary zero...
-
Code and rewritten doc checked in. Will wait a bit for feedback and then close this.
-
Change Kepler plugging to handle dates for Kepler, K2 and TESS; change date to BJD
-
For reference, I find the Kepler LC documentation at http://archive.stsci.edu/kepler/manuals/archive_manual_jul2011.pdf#page=15 For TESS https://archive.stsci.edu/missions/tess/doc/EXP-TESS-ARC-ICD-TM-0014.pdf#page=32 The Kepler plugin pulls columns 0, 7 and 8 from the LC FITS file. According to the above documentation these are the same metrics for Kepler (hence K2) and TESS: TIME, PDCSAP_FLUX and PDCSAP_FLUX_ERR. However, the BJD offset for TIME is the hardcoded value 2454833.0 in the plugin. However,...
-
This document is my attempt to capture some thoughts on this potential plug-in.
-
Makes sense to me Cliff Kotnik clkotnik@att.net On Dec 12, 2019, at 16:10, David Benn david_benn@users.sourceforge.net wrote: Hi Cliff I’ve really been appreciating you handle this. It’s helped a lot because I’ve been busy with various things. Thank you! Joe’s comments are very helpful too. I may make some summary comment on the forum. Thank you also for offering to integrate AIJ’s code and I may well take you up on it. Before we go down that route though, now that things are starting to settle...
-
I think we now have a representative set of comments on the AAVSO forum on this topic. Additionally there is this comment from Joe Patterson: Some comments on timing accuracy... Nearly all CV timing projects are happy with 10-second accuracy. For several reasons: * The fastest timescale you can get from a white dwarf is ~10 s, and we need (many) hundreds of cycles to get a detection. The accuracy of such an ensemble would be much better then the accuracy of each point. Nearly all (maybe all) CVs...
-
The point is valid - how accurate does this need to be. I did not capture in the ticket the discussion David and I had prior to the ticket creation about whether this was in need of work. We could pose that question on the AAVSO forum. Cliff Kotnik clkotnik@att.net
-
AstroImageJ provides the JD to HJD functionality. It seems to be a fairly self contained implementation - all within the source file SkyCalc.java, attached. I believe there would be no problem reusing this, but I think it would be wise to request permission on the project forum:http://astroimagej.1065399.n5.nabble.com/. The comments in the code are: / Calculates HelioCentric Julian Day correction based on Julian Day * Adapted from javascript written by Dan Bruton of Stephen F. Austin State University...
-
JD to HJD converter plugin is consistently different from other tools by a small amount
-
Doc updated needed for Leavitt's Law
-
Completed
-
Rename period luminosity to Leavitt's Law plugin
-
Doc updated needed for Leavitt's Law
-
Updated and committed. Cliff Cliff Kotnik clkotnik@att.net From: David Benn david_benn@users.sourceforge.net Sent: Wednesday, October 30, 2019 4:32 AM To: [vstar:bugs-and-features] 675@bugs-and-features.vstar.p.re.sourceforge.net Subject: [vstar:bugs-and-features] Re: #675 APASS and AAVSOnet Epoch Populate from URL Issues Hi Cliff Thanks very much for this. Looks good! Feedback: vstar_plugins folder (not vstar_plugin) Can be installed via the plugin manager too VStar (vs Vstar) in a few places Other...
-
Add doc for epoch aavsonet and apass
-
Attached is the doc I created for these plug-ins. If this looks ok, I'll add it to SF.
-
I have reproduced this bug in 2 out of 4 attempts with 2.21.0 svn ver 1703. Attacted is my log for one of the failures. I tried to get more exact with reprocing this. Perhaps I am on to something: From a fresh launch I open Max's test file. I do the DC DFT with period as Max describes. I then dismiss the dialog in 2 ways: a) Dismiss button or b) Dialog close X button I then do AoV with period as Max describes. I switch to top hits and pick the 3rd one 63.7. In the a) case I do not get an error. In...
-
I'm not sure there is a separate doc for these plugins. I can take that on. Regarding an alternate interface, I think it is not critical when doc is provided. Cliff Cliff Kotnik clkotnik@att.net From: David Benn david_benn@users.sourceforge.net Sent: Saturday, October 26, 2019 6:33 AM To: [vstar:bugs-and-features] 675@bugs-and-features.vstar.p.re.sourceforge.net Subject: [vstar:bugs-and-features] #675 APASS and AAVSOnet Epoch Populate from URL Issues Comment: Started lookinhb at this. The code looks...
-
APASS and AAVSOnet Epoch Populate from URL Issues
-
Issues Building Release Candidate
-
Index Out of Bounds after marking observation discremant in phase plot
-
Orignal system svn, version 1.9.3 (r1718519) Upgraded system svn, version 1.9.7 (r1800392) That option on the cat call seems to work. Thanks!
-
Sorry to have overlooked the command line test you requested. The output of "svn cat <url repo="" to="" file="" in="">" expands the Id keyword on both the original and updated systems.</url>
-
Just for completeness, the attached file is the output from the test program run with the original version of pysvn. It used the same repo different svn working directory.
-
Compressed tar
-
It does not appear to be a case problem. Attached to this post, please find a simplified example that demonstrates the problem. I simply printed the lines of the file as returned by client.cat to stdout along with the file properties. In the following post, I have attached a compressed tar-ball that contains the results of testing on my repo. In it, svntest.txt is the stdout from running the program. There are three subdirectories that contain the files in question. Any pair of these can be compared...
-
Turn on property Id
-
Adding a test document
-
I upgrade the machine from Ubuntu 16.04.5 LTS to Ubuntu 18.04.2 LTS. The upgrade to python-svn was automatically performed as part of that. Attached is the program we have been running daily for a couple of months. After the upgrade it throws an exception at line 208, "svnID= int(ll2[3]);" because the $Id line has simply that keyword. Clearly, the code should gaurd against this. However, with the change 95% of the files retrieved at line 193 with the client.cat call have none of the details they...
-
I upgrade the machine from Ubuntu 16.04.5 LTS to Ubuntu 18.04.2 LTS. The upgrade to python-svn was automatically performed as part of that. Attached is the program we have been running daily for a couple of months. After the upgrade it throws an exception at line 208, "svnID= int(ll2[3]);" because the $Id line has simply that keyword. Clearly, the code should gaurd against this. However, with the change 95% of the files retrieved at line 193 with the client.cat call have none of the details they...
-
Id lines are blank
-
on behalf of Dave Hinzel:a few changes to improve docs
-
on behalf of Dave Hinzel: documentation for new plugin to convert JD to HJD
-
on behalf of Dave Hinzel: add doc for Hipparcos observation source plugin
-
Allow either old or new format to be loaded
-
Change for different file format and handle exceptions for trailing non-data lines
-
on behalf of Dave Hinzel: adding doc for existing plugin ASAS
-
On behalf Dave Hinzel: Update/create docs for AoV, Loess and DASCH
-
Changed ASAS plugin filter to V and moved dataset info to comment
-
Add a paragraph to describe considerations for reloading saved observations to the observations list section
-
Add Open doc format (ODT) versions for documents without them
-
When observations are saved to a text file do not save excluded observations. Otherwise, these will be reloaded without any indication of their excluded status.
-
Add star name to all observations sources so that if observations are subsequently saved, they will pass the edit to be reloaded
-
I completed a test of all the ob src plugins except Hipparcos and DASCH. I could not locate any data for those two. I was successful doing a round trip with only AID and Gaia. Oddly, the upload file format would not round trip in spite of having a star name in the original file. Cliff
-
David, I looked at this a bit. Not unlike DIY home and auto repairs, it seems more complex after getting into it. So I thought I should understand the edits within the code that reads the download file format/”Now star from file…”. There is enough indirection involved in the implementation that I am unsure I can easily determine this. So I thought I’d look at the APASS plugin since it was like the Gaia plugin. I now realize that the APASS and AAVSONet plugins are implemented mostly in the main VStar...
-
Ok. I'll start with examining the code for loading the download format.
-
Consider Y Oph which is Gaia ID 4175017625462647168 and is located at RA/DEC 268.16125 -6.14358. So do a load from AID from 2457000 to 2458500. I get 297 observations. No errors. Save to file AID.tab. Loaded Observations: 297 Johnson B: 5 Johnson V: 11 Tri-Color Blue: 6 Visual: 275 Note the AID.tab file contains only the Visual and V observations??? Now load AID.tab with New Star from file. It loads 286 observations. Now do a New Star from APASS, NOT ADDITIVE. All filters. Loaded Observations: 9...
-
David, I tried this "round trip" approach. While it initially failed, I analyzed the saved file format. It appears to be tab delimited and line up with the download file format . After some experimentation, it appears that often the star name is blank, as when loading from a source like Gaia, Kepler or ASASSN. When I simply added a star name by hand, the file loaded with the "New Star from File..." option. Does this make sense?
-
Update the Kepler plugin documentation and place the Word document in the repository
-
Plugin created and release for Gaia data release 2.
-
GAIA observation source plugin
-
And the R Lyr CSV file used for testing
-
I have attempted to summarize the results in the attached XLSX file. I should say that for these large lists the elapsed time varies by 10-20% from one run to the next of the same test. To summarize, the new version is a significant improvement but still takes 3 times as long as the sum of downloading and importing a CSV file with the same data. And counterintuitively, loading a subset of fields and filters takes 2 or 3 times longer in the new version than loading everything. The tests with ASASSN-18ey...
-
I saw the chatter on the AAVSO forums regarding LCG and initially just let it go by since I use VStar to look at light curves. When I had some spare time, I read the comments (https://www.aavso.org/lcg-v1-issues-clarification-understanding-and-solutions). If you put aside the "somebody moved my cheese" issue, it seems that there are two groups of issues. One big group seems to be performance and the new LCG as well as VStar for medium and large observation sets. I spent a bit of time looking at Eric's...
-
Add to current? warning message
-
Closed since not needed Cliff
-
I believe this enhancement request resulted from my misinterpretation of a privat email exchange with Dave. The "Add to current" feature does work based on the testing we have done and reported above. The fact that enabling "Add to current" when there is nothing currently in the observation list causes no problem and does not seem worth the time to do anything about. I recommend this enhancement request be rejected. Cliff
-
Dave, David, I fear I have created a problem where none existed. I apologize for that. Dave, in your email to me summarizing your test results, you said "The passband transformations work as well as the Add to current? (nothing happened since there was nothing to add). I didn't get any error messages." I interpreted this to be case 1 in the attached document. Above in this discussion you said "When testing the Add to current? function, I just simply found that nothing happened since I didn't add...
-
Dave, I initiated this based on your feedback while testing the Gaia plugin. I it thought this was straight forward but perhaps not exactly. Let us use the example of V0548 Oph = Gaia source_id 4068192477406359808. Note that if you load from AAVSO DB by name using the default dates, there will be no observations - you must specify all data or a wider range. Similarly, if you mistype the Gaia source_id there will be no observations (unless you are VERY lucky). Can you specify a sequence of steps starting...
-
Add Know Issues section to Gaia Plugin doc for clipboard problem with Webstart.
-
Copy/Paste from System Clipboard Silently Fails for Plugins With Webstart Launch
-
Change build for Gaia plugin, add doc and cosmetic changes to source
1 >