1. 22 Jan, 2021 1 commit
  2. 10 Jan, 2021 1 commit
  3. 03 Dec, 2020 1 commit
  4. 16 Nov, 2020 1 commit
  5. 13 Nov, 2020 1 commit
  6. 04 Nov, 2020 2 commits
  7. 02 Nov, 2020 1 commit
  8. 01 Nov, 2020 1 commit
  9. 25 Oct, 2020 2 commits
    • Jasem Mutlaq's avatar
    • Jasem Mutlaq's avatar
      Major refactoring for FITSViewer and how FITS and other data is handled: · e1289951
      Jasem Mutlaq authored
      1. Eliminate any unncessary "temporary" files. If we have a buffer, use it without creating any files on disk unless strictly required.
      2. Use a shared pointer for data so that FITSView instances are implicitly sharing the same pointer. This finally allowed for Summary View to load the same view as FITS Viewer without incurring any memory penalty.
      3. Add metadata to the FITSImage object to pass information that were passed before using AUX members of the blob which was not safe.
      4. Add support to load RAW camera files directly in the FITS Viewer along with loading JPEG/PNGs that were added by Robert earliar.
      
      CCMAIL:kstars-devel@kde.org
      e1289951
  10. 20 Oct, 2020 1 commit
  11. 09 Oct, 2020 2 commits
  12. 08 Oct, 2020 2 commits
    • Robert Lancaster's avatar
      This is a pretty big change. It is the start of the integration of the... · e7815484
      Robert Lancaster authored
      This is a pretty big change.  It is the start of the integration of the StellarSolver Library that I created this spring.   Here are some of the changes that it includes:
      
      Utilizes StellarSolver for both SEP and for Plate Solving.
      Adds an Options Profile Editor for creating profiles for use with StellarSolver
      Consolidates all the multitudes of options for plate solving and SEP in the Options Profiles.   These profiles are Parameters files that are shared between KStars and StellarSolver and can be used for both SEP and Plate Solving.
      Usage of Stellarsolver allows users of all operating systems to plate solve without astrometry.net or an internet connection.  But you do need index files.
      Usage of Stellarsolver allows Windows users to now use the index file downloader, directly plate solve using astrometry.net in either Cygwin or ANSVR, and have full control over plate solving.  ANSVR does not need to be "running" and the "online" option is no longer required for using ANSVR.  You can still use that if you like however.
      All users can still make use of ASTAP, online Astrometry.net, local astrometry.net, sextractor, and other plate solving features.  I have added more options as well, such as using the internal sextractor with external ASTAP.  These all now share the same options profiles.
      Usage of StellarSolver eliminates the need for any external configuration files for Sextractor or Astrometry.net.  You can use internal or external ones without the need for configuration files.
      Using the parallel processor options in StellarSolver will allow blind plate solving for many images in just a few seconds instead of minutes.
      Several structures are now shared between StellarSolver and KStars so that information can be sent back and forth between them.  These include the Options Profiles (Parameters), Statistic, WCS_Point, Background, and Star (for found sources).
      The information contained within the above structures can be retained to add all sorts of new features to KStars for use in Photometry, image analysis, object analysis etc.  In the past, this information was lost as soon as the plate solve or SEP was done.
      On the recommendation of Hy, dataType, which was formerly a property of the FITS Data object, is now incorporated into Statistic and I made changes throughout KStars to support this.
      The Astrometry Parsers, the Options Maps, some of the settings, and some of the options editor windows have been eliminated because they were redundant, but a lot of the code is still there commented out in case some is still needed.  This needs more work.
      For now, the "Remote" option was eliminated because I can't support that with StellarSolver because it depends on INDI to function and that would mean making StellarSolver INDI dependent.  The Remote option could be re-enabled if needed but I think that the fact that KStars can now plate solve on its own without the need for any astrometry.net installation makes it unnecessary.
      I have not yet implemented StellarSolver for SEP in Guiding.  I think it will help there too, but I want to make sure that everything else works and check with Hy before I touch his great work there.  He already helped with Focus, but I haven't consulted him yet on Guiding, so it is not implemented there yet.
      Load and Slew now loads the image it is trying to solve in the AlignView, and then blind solves with StellarSolver using the options that were selected.  Note that it ignores information in the file such as position or scale.  Load and slew is capable of solving many types of images
      "Open FITS Image" in the file menu has been replaced with "Open Image" and is capable of loading any type of image QT can open as well as the FITS types it could before.  This was possible because of the function I had to add to get Load and Slew to work with StellarSolver.
      
      What you can do:
      
      Please check to make sure that this massive update did not break anything important.
      Specifically, please verify that none of the code that I commented out or removed when I put StellarSolver in Align or SEP was actually needed after this update.  There was A LOT of code that I had to change and I could have easily messed something up.
      Please test StellarSolver to make sure that it plate solves fine in the Align module, and that it works well in the Focus module for SEP in real life situations.  If there are any issues with it, please let me know.
      Please help me create good Options Profiles.  StellarSolver comes with a default set of profiles, so there is no initial installation required.  If you change any of them or add any of your own, then it saves a file with the profiles in it. I believe that my profiles for plate solving are pretty good, but that doesn't mean they can't be improved upon.  I think that my profiles meant for SEP need some work.  I haven't gotten a lot of feedback on them so far.  You can easily change them and you can create your own options profiles.  I would like to provide a good default set for users.  If you find anything in the default set that you think would be better, or if you come up with a profile that you think works better for certain situations, I would like to include that, please send it to me.
      
      Testing instructions:
      
      First, please download and install StellarSolver.  Use the latest Master because I have been making some changes recently based on things that I encountered while integrating into KStars and I am still updating based on that.  Those changes are not in the latest release yet.  https://github.com/rlancaste/stellarsolver. You can use the install script if you like or you can build and install it manually.  For Windows and Mac testing, the repository includes a craft recipe for installing in Craft on Windows and Mac OS.
      Then you can test the code in my merge request.  One option to test it would be to download the code in my fork and then build and install KStars normally.  Another would be to use the features in GitLab to download a patch and then apply it to a local copy of the latest KStars you have.  Since this is my first merge request since we moved to Gitlab, I don't know all the details of how to do the second option.
      e7815484
    • Jasem Mutlaq's avatar
      Revert "This is a pretty big change. It is the start of the integration of... · 0ed5f933
      Jasem Mutlaq authored
      Revert "This is a pretty big change.  It is the start of the integration of the StellarSolver Library that I created this spring.   Here are some of the changes that it includes:"
      
      This reverts commit 1dd9baf7, reversing
      changes made to 7120f95a.
      0ed5f933
  13. 29 Sep, 2020 1 commit
  14. 22 Sep, 2020 1 commit
  15. 20 Sep, 2020 1 commit
  16. 14 Sep, 2020 1 commit
  17. 13 Sep, 2020 1 commit
  18. 03 Jul, 2020 1 commit
  19. 12 Jun, 2020 1 commit
  20. 27 May, 2020 2 commits
  21. 04 May, 2020 1 commit
    • Jasem Mutlaq's avatar
      Implemented Bahtinov mask focus assistant, fix compiler warnings (in... · 77c2bbe3
      Jasem Mutlaq authored
      Implemented Bahtinov mask focus assistant, fix compiler warnings (in fpackutil.c and modcalcvizequinox.cpp)
      
      Summary:
      Implemented Bahtinov mask focus assistant, fix compiler warnings (in fpackutil.c and modcalcvizequinox.cpp)
      Added unit test for bahtinov focus assistent
      Fix compiler error (comparison between int and double)
      Fix compiler error after merge
      
      # Conflicts:
      #	kstars/fitsviewer/fitsview.cpp
      
      Test Plan: Tested manually by installing software on Raspberry Pi and connect it to my telescope, pointing to a bright star and check if bahtinov mask is overlaying the diffraction pattern correctly.
      
      Reviewers: #kstars, TallFurryMan, mutlaqja
      
      Reviewed By: #kstars, TallFurryMan, mutlaqja
      
      Subscribers: lancaster, astrorunner, mutlaqja, TallFurryMan, kde-edu
      
      Tags: #kde_edu
      
      Maniphest Tasks: T13035
      
      Differential Revision: https://phabricator.kde.org/D29325
      77c2bbe3
  22. 26 Apr, 2020 2 commits
    • Yuri Chornoivan's avatar
      Fix minor typos · ec82fc3b
      Yuri Chornoivan authored
      ec82fc3b
    • Hy Murveit's avatar
      Improve Overlay Rendering · 68fda247
      Hy Murveit authored
      Summary:
      Previous PR to reduce fitsview zoom memory use created pixelated rendering for some overlays
      with large amount of zoom--e.g. when using a small view (e.g. around a focus view with subframe enabled.
      Here we employ a different rendering strategy when the view has a smaller number of pixels (and thus
      memory use shouldn't be as much of an issue) so that the rendering is smoother.
      
      Test Plan:
      Load fits images, try all the overlays, zoom in and out.
      Overlays should not change with images greater than a megapixel.
      Try this with various views with subframing (e.g. focus), and the rendering should be less pixelated.
      
      Reviewers: mutlaqja
      
      Reviewed By: mutlaqja
      
      Subscribers: kde-edu
      
      Tags: #kde_edu
      
      Differential Revision: https://phabricator.kde.org/D29192
      68fda247
  23. 22 Apr, 2020 1 commit
    • Hy Murveit's avatar
      Fix stretch control bug created in D28935 · 1133ae45
      Hy Murveit authored
      Summary: When using the stretch controls, bug had image 4X smaller than it should be.
      
      Test Plan:
      With D28935, when using the custom stretch sliders, the image displayed while using the sliders
      was 4X smaller than it should be (just the preview images--once the sliders were released, the image was
      rendered fine). To test, load an image, make a custom stretch change (either changing shadows, midtones or highlights)
      and the image displayed while changing the sliders should not change in size as it would without this fix.
      
      Reviewers: mutlaqja
      
      Reviewed By: mutlaqja
      
      Subscribers: kde-edu
      
      Tags: #kde_edu
      
      Differential Revision: https://phabricator.kde.org/D29085
      1133ae45
  24. 21 Apr, 2020 2 commits
    • Jasem Mutlaq's avatar
      sysconf for Linux and MacOS only · cc087cd9
      Jasem Mutlaq authored
      cc087cd9
    • Hy Murveit's avatar
      Reduce memory use when zooming in fitsviewer · 48491864
      Hy Murveit authored
      Summary:
      Instead of creating a scaled image, let Qt do the work.
      This saves half the memory when zooming 200% or 300%.
      I also limited the zoom to 300% running out of memory has been
      an issue for users. This change was relatively straight-forward
      for the image rendering, but there was complexity with the
      overlays drawn, espcially those rendering text.
      
      Test Plan:
      Load an image, zoom in. You should be limited to 300%. You should notice less than
      1Gb of memory use from a ~20megapixel camera, whereas this would be over 2.5Gb
      previously. Try the various overlays. They should work the same as before (some line
      widths may be slightly different, but I tried to match the previous rendering.
      Similarly, test focus, align, guide views, zooming in and so forth.
      
      Reviewers: mutlaqja
      
      Reviewed By: mutlaqja
      
      Subscribers: kde-edu
      
      Tags: #kde_edu
      
      Differential Revision: https://phabricator.kde.org/D28935
      48491864
  25. 17 Apr, 2020 1 commit
    • Hy Murveit's avatar
      SEP Focus improvements · 515efd42
      Hy Murveit authored
      Summary:
      This PR should improve the HFR calculations with SEP for the focus module and for displaying HFR for those that want to track it in subs.
      It is not the end-all, but should be an improvement. There is still work to do.
      - Updated and added to tests for the fitssepdetector.cpp
      - Removed saturated stars from the HFR calculation, if possible.
      - Sped up the SEP star detector 2X when used with full exposures.
      - Improved SEP star detector quality, especially for focus/hfr, by (1) changing the deblend constant, and (2) removing the largest stars
      - Fixed two bugs in the SEP star detector (1) major issue, ushorts being treated as shorts, and (2) a memory leak for (rare) float images
      - Added logging for HFR when calculated
      - Fixed bug so that Mark Stars now works
      - Improved drawing of circles for Mark Stars, now uses star radius, instead of flakey star width.
      - Updated the fitsviewer status to include the HFR when calculated
      - Increased a threshold in SEP extract.c which previously would fail on images, e.g. like M42, where there was a lot of area over background.
      
      Not done, but related:
      - Should improve the MarkStars UI so that HFR could be calculated without star annotation on display, and could turn on on first image.
      - API to star detection methods should be generalized (should wait on rlancaste's investigations of SEP best practices)
      - Some display of HFR trends.
      
      Test Plan:
      Basic test added to testfitsdata for SEP, run that.
      Load a .fits file into fitsviewer. Select view->MarkStars, for a full-exposure sub, you should see ~100 stars circled with an HFR displayed in the status bar.
      Load another image and the HFR should continue to be displayed.
      Similarly, one could make a capture sequence job with, say 20 5s simulator exposures. After the first exposure completes and the fitsviewer displays
      the image, go to the view menu and turn on Mark Stars, then the rest of the images should display with stars circled and HFR displayed in the status bar.
      This way of turning on the star/HFR detection should be improved in a future PR.
      
      Reviewers: mutlaqja, lancaster, TallFurryMan
      
      Reviewed By: mutlaqja, lancaster, TallFurryMan
      
      Subscribers: kde-edu
      
      Tags: #kde_edu
      
      Differential Revision: https://phabricator.kde.org/D28767
      515efd42
  26. 27 Jan, 2020 1 commit
  27. 25 Jan, 2020 1 commit
  28. 20 Jan, 2020 1 commit
    • Hy Murveit's avatar
      Re-allocate the rawImage buffer if the number of channels has changed. · 64306685
      Hy Murveit authored
      Summary:
      If the number of channels changes during a session (rare, but can happen now that the
      simulator can be changed from mono to rgb and back) then the rawImage buffer needs to be
      re-allocated.
      
      Test Plan:
      Using the new Indi ccd simulator capability to switch between rgb and mono,
      if you do this in one session without this fix, you'll see the fitsviewer
      not display the image properly after the change (though it would it you saved configuration
      and restarted kstars). With this fix, that bug should go away--e.g. change the
      ccd simulator to rgb, capture an image, note the display is reasonable, switch again, should
      be reasonable again.
      
      Reviewers: mutlaqja
      
      Reviewed By: mutlaqja
      
      Subscribers: kde-edu
      
      Tags: #kde_edu
      
      Differential Revision: https://phabricator.kde.org/D26778
      64306685
  29. 12 Jan, 2020 1 commit
  30. 09 Jan, 2020 1 commit
  31. 08 Dec, 2019 1 commit
    • Hy Murveit's avatar
      New UI to allow the user to modify the stretch being used. · 08fd2a3b
      Hy Murveit authored
      Summary:
      In fitsviewer, the toggle stretch button is removed and replaced with a
      stretch bar at the bottom of the window. It is always there and allows
      the user to turn on/off stretching, use the automatically generated stretch
      or modify the stretch parameters. As before, this is a display-only change
      to the image, it does not modify the image itself.
      
      Test Plan:
      In kstars, open a new fits file with the auto-stretch option either set
      or not, and see if it displays accordingly. Then switch the other way
      using the UI and play with the sliders to adjust the stretch. Try another
      image also. Go back to the first image, it should remember the parameters
      that were last used. Should work similarly with images from capture.
      Evaluate the UI.
      
      Reviewers: mutlaqja, lancaster, wreissenberger
      
      Reviewed By: mutlaqja, lancaster
      
      Subscribers: kde-edu
      
      Tags: #kde_edu
      
      Differential Revision: https://phabricator.kde.org/D25645
      08fd2a3b
  32. 19 Nov, 2019 1 commit
    • Hy Murveit's avatar
      Update the auto-stretch scheme and UI · 3eece768
      Hy Murveit authored
      Summary:
      Modifies the auto-stretch algorithm.
      Also, the UI is modified. When images are viewed, they can be viewed stretched or unstretched
      and this can be toggled via a button in fitsviewer or on the floating tool bar in a fitsview.
      This stretching does not modify the contents of the image.
      
      Test Plan:
      Open a variety of fits images in fitsviewer and from captures in the capture, focus, align and guide tabs.
      See if they can be stretched and unstretched.
      
      Reviewers: mutlaqja, wreissenberger
      
      Reviewed By: mutlaqja, wreissenberger
      
      Subscribers: lancaster, TallFurryMan, wreissenberger, kde-edu
      
      Tags: #kde_edu
      
      Differential Revision: https://phabricator.kde.org/D25318
      3eece768
  33. 27 Oct, 2019 1 commit
    • Hy Murveit's avatar
      Write FITS file on separate thread · c72edccb
      Hy Murveit authored
      Summary:
      When capturing images, previously once the capture was complete, images were written
      to disk, then read back, then displayed in the fits viewer (if enabled), and finally the
      next image could be captured. This delayed the next capture by a few seconds.
      With this change, the FITS data is displayed using (a copy of) the INDI blob
      sent to indiccd, a memory buffer, and the FITS data is written to disk on a separate thread,
      which does not block the next capture nor the display of the indi data.
      This was only implemented for FITS data. This change also involved some refactoring
      in the affected files.
      
      Reviewers: mutlaqja
      
      Reviewed By: mutlaqja
      
      Subscribers: kde-edu
      
      Tags: #kde_edu
      
      Differential Revision: https://phabricator.kde.org/D24976
      c72edccb