×

INDI Library v2.0.6 is Released (02 Feb 2024)

Bi-monthly release with minor bug fixes and improvements

New astro imaging catalog software. Help testing needed

  • Posts: 249
  • Thank you received: 62
thank you Doug! your feedback has been really valuable.

As you say, there are many other software that are already good at showing images, I will not try to replicate them. Launching an external viewer seems the best option, even if I had issues when I tried to open external softwares from AD. For example it was a good idea to open a file browse pointing at the directory containing the file when clicking on its name. But on Linux I struggled to find how to do it.

Reducing the interline in the details window will result to be less readable. What if I change layout to landscape orientation? I mean image preview on the left and fields description on a right pane.

cheers,
ferrante
3 years 10 months ago #54501

Please Log in or Create an account to join the conversation.

  • Posts: 398
  • Thank you received: 117
No need to reduce the spacing if it detracts. I suspect changing orientation would be arbitrary too. If you figure out how to launch the Kstars FITS viewer on the selected filename (and probably want to limit this to some number 1-5 to avoid user select mis-steps), I'd say maybe the next thing to consider might be a CSV data export (for user selected lines & columns). This enables other tools (Python, Excel, etc) to take advantage of the growing info content of DB fields.

If you do decide to add CSV export, I would recommend you add both row index selector and column selector. There's no current row index (but users can mouse select/drag). Selecting a column is currently reserved for sort order (asc/decend). A minimum column rectangle above each column heading could serve to allow users to select the entire column contents. On row index, folks are so used to seeing line number, I'd just include it and use it as the entire row selector.

BTW, I couldn't figure out how to see hidden fields after hiding, then unchecking. How is that accomplished? If no auto-refresh on the GUI, then should a refresh button be added to see the fields again?
3 years 10 months ago #54517

Please Log in or Create an account to join the conversation.

  • Posts: 249
  • Thank you received: 62
I succeeded in opening the default external viewer associated with fits files; on my system it launches Pixinsight. Again, as this is OS dependent hope it works with Win too.
But for each file it opens a new instance of the external software. Hoped that files would be opened in the same instance but I don't know how to control this.

Basic CSV export is easy, will add in next release; it will not export selection of cells but the whole data set (as selected with filters applied) .
As the export is based on the model of data and not the view (what you see in the table) it will export also hidden columns and coordinates columns without proper formatting (ad decimals).
So part of the work still have to be made using Excel: delete unwanted columns, format data.

About hidden columns: after applying hide/show to any column AD has to be restarted and columns width preferences will be reset. I had to do this because when column width preferences were reloaded they were overriding the hide/show settings. Not optimal behavior I know.

ferrante
Last edit: 3 years 10 months ago by Ferrante Enriques.
3 years 10 months ago #54539

Please Log in or Create an account to join the conversation.

  • Posts: 398
  • Thank you received: 117
On update to 0.2.12, I was able to have the Kstars FITS viewer launch (via a full Kstars launch). Nice add, but with an unintended kstars launch (likely caused by no separate executable for the viewer). Maybe this is something the Kstars senior developers might want to comment on. Having the Kstars viewer for AD is definitely an improvement, but having a full kstars launch (one per selected image) is not as nice. Ideally, each file select would be passed to the viewer without a new kstars launch (similar to what happens during normal image capture sequences).

I think because of the above, and due to single mouse click function overload (delete rows), I suggest the FITS viewer launch be invoked only by double click mouse operation.

Finally, I'm noticing a bit of odd behavior of delete rows (Image and FITS Import tabs). Deletes definitely work, but on import tab, only one row deletes at a time (no matter how many rows you've selected). On the Image tab, multi-selected rows can be deleted, but sometimes less than all rows selected are deleted. It seems inconsistent....not quite sure what's going on there.
3 years 9 months ago #54816

Please Log in or Create an account to join the conversation.

  • Posts: 249
  • Thank you received: 62
On my pc .fits files are associated to Pixinsight and fits are opened in the same instance of PI.

About clicking rows, as of now behavior is:
- single click selects a row. Used when selecting a single night or deleting row(s).
- double click launches the Detail window.
- double click on the file name only , opens the default fits software.
If double click would open kstars, there should another option to launch detail windows. Any suggestion?

You're right: the behavior when deleting rows is not consistent. I'll fix it.

thanks again for your feedback
ferrante
3 years 9 months ago #54826

Please Log in or Create an account to join the conversation.

  • Posts: 398
  • Thank you received: 117
Hi Ferrante, on my system, Import fits tab has a left index for selecting entire rows. I don't have a similar selector index on the Images tab (but would like it). I hadn't realized that a double click on other fields would bring up the old Detail window. That's fine. On my system's Images tab, I don't get the opportunity to double click on filename. As soon as a single click is sensed, I get a Kstars and FITS viewer launch. So, I think my recommendation would change in favor of adding an index for row selection consistent with the import tab, and do nothing about the single vs double click.

We need help from someone who understands FITS viewer better. We'd prefer to avoid multiple copies of kstars. Not sure if that means a stand-alone copy of FITS viewer, or a different interface/setting. One copy of Kstars is ok (can always minimize). Several copies are undesirable. Cheers, Doug
3 years 9 months ago #54835

Please Log in or Create an account to join the conversation.

  • Posts: 249
  • Thank you received: 62
hi Doug,
some fixes in v 0.2.13:
- Delete multiple rows in Fits import and CSV imports.
- Added index column in image list. There's still an issue when sorting: the index keep always the same ordering.
- In the table, single click only selects an item. Double click opens the external viewer (if clicked on the file name) or the detail windows if clicked elsewhere.

ferrante
3 years 9 months ago #54908

Please Log in or Create an account to join the conversation.

  • Posts: 398
  • Thank you received: 117
yes, I confirm 0.2.13 works as advertised/expected. I don't have a problem with the index sort order remaining as is; it's just a useful handle for selecting whole rows. Cheers...
The following user(s) said Thank You: Ferrante Enriques
3 years 9 months ago #54910

Please Log in or Create an account to join the conversation.

  • Posts: 398
  • Thank you received: 117
Hi Ferrante, I was wondering your thoughts about a generalized CSV import capability for AD. There are lots of ways folks might want to associate/import CSV data, and it's likely impossible to think of them all specifically. Having a general CSV import interface might be ideal. Needed would be a way to associate the primary field (e.g. filename?), plus field name definition(s). Against those, import of values could proceed. For example, CCD_Inspector can generate useful CSV information (indirectly, through clipboard, but then via Excel -> CSV). Here's a sample of some fields of interest (there are other fields....this is just an example) that would be nice to associate:

Date & Time, Image File, FWHM, Aspect (%), Background (adu), Stars Used, Contrast Ratio, Curvature, Temperature
6/11/2020 1:16, M_27_Light_023.fits, 3.24", 14, 792, 2940, 31.59, 21.6, -5
6/11/2020 0:47, M_27_Light_011.fits, 3.29", 19, 712, 2949, 32.65, 33.1, -5
6/11/2020 1:16, M_27_Light_024.fits, 3.31", 25, 792, 2911, 31.91, 18.8, -5

Clearly, some pre-massage on data might be required of users (e.g. removal of " for arcsecs), but what do you think?
Could a mechanism for new CSV field(s) definition, plus an association key, allow import of more generic CSV data?
Just wondering.... Doug
3 years 9 months ago #55381

Please Log in or Create an account to join the conversation.

  • Posts: 249
  • Thank you received: 62
hi Doug,
the most flexible approach would be to allow the user to add custom fields to AD database: name of the field, type, format could be a minimum set. And then choose the source for these data, like CSV file, json, xml or even direct connection to a server. Filtering, sorting and plotting these data would be then straightforward.
It requires some restructuring of the code but it will be a nice improvement that will allow AD to be a single point of information for image metadata.

As of now I'm following another similar request by a github user (Rick) that suggested to import PHD2 guiding data through its log. It will allow to correlate guiding performance vs eccentricity or Alt/Az etc.
I'm not a PHD2 user so it's quite difficult to test. For example I'm not able to recreate the RMS calculation to have the same result as PHDLogViewer. Seemed to be easy but it's not.

A note about the filename as identifier key: names could change, files could be moved and different images could have the same file name. I chose a different approach that prevents these issues: hashing the file itself (a part of it, for performance reason). But for our understanding saying that file name is a key it is ok, I was just more specific because you're quite into the software details.

ferrante
3 years 9 months ago #55413

Please Log in or Create an account to join the conversation.

  • Posts: 398
  • Thank you received: 117
Not sure what you're chasing in the PHD log, but I would be interested in the details if you care to share them (as I too use PHD and the logViewer). I can't imagine why AD would want to read the PHD log directly as opposed to having a user prepare a CSV file for import. For Guide RMS (example), are you thinking just to average reported values over the exposure interval? Just thinking out loud here, but it seems that PI or CCD_Inspector (eccentricity, Aspect %) directly imply guiding performance (only entangled with coma for worst case collimation issues). I think the fundamental issue for AD is to work at the same time scale as the primary key (Fits file, thus exposure time). Working at smaller time scales seems fraught with integration problems (as you seem to be discovering). The nice thing about PI statistics and CCD_Inspector stats is that they're at the same time scale, so you don't have the issues that come from working on smaller time scales.
3 years 9 months ago #55458

Please Log in or Create an account to join the conversation.

  • Posts: 249
  • Thank you received: 62
hi Doug,
there's a thread on Github where there are additional details:
github.com/fenriques/AstroDom/issues/1

I preferred reading the log directly because it is more user friendly than manipulating a csv file. Anyway the difficult part was the logic processing the data not reading the file.
I do not guide but I guess that knowing the rms correlation with not only eccentricity but also with alt/az, filter or event different set ups would be helpful to improve guiding performance.

AD time scale will not be affected: as you correctly guessed, phd log data are integrated over the exposure time of the fits file and RA / DEC rms is calculated for that time interval.

Attached a screenshot of an early version.
3 years 9 months ago #55469
Attachments:

Please Log in or Create an account to join the conversation.

Time to create page: 0.685 seconds