Skip to main content

New Light Curve Generator(LCGv2): Updates, Questions, & Suggestions

60 posts / 0 new
Last post
fhemsher
New Light Curve Generator(LCGv2): Updates, Questions, & Suggestions

This topic will report updates/improvements to the LCG. Also, you can post questions you may have, and offer any suggestions that you believe would improve it for your use.

Regards,

Francis Hemsher

 

fhemsher
New Feature: Plot Previous Days

Hi All,

A new feature has been added to the LCG. This provides the ability to plot a star back a desired number of days. See the image below. Provide the number of days, then select the checkbox to set the time span.

File upload: 
lmk
lmk's picture
"No data found" error for a star with tons of data!

I just ran for TT ARI going back 15 years, and it comes back no data found! When I run it for just a couple years back (2013), it works fine. What is the issue here? It appears there is a limit to the number of observations displayed? LCGv1 did not have any limit (apparently), so this needs to be fixed. Thanks!

 

fhemsher
No Data Found

It seems that a request to the AAVSO database, where the time span request that preceeds any data being available causes a 'No Data Found' response (Nothing is returned to the LCG). I'll confirm this soonest. Also, we'll remove the 30,000 observation limitation early next week.

Thanks for your feedback,

Francis

David Benn
David Benn's picture
No Data Found

Hi Francis

Let me know if you want me to try a particular AID request via VStar to sanity check anything you are seeing, since we're both making use of the web service.

David

fhemsher
Sanity Check ;)

Hi David,

Ain't that the truth...

I guess it's time to document data access methods for those of the future that are to use it. George Silvis recently mentioned he's addressing this. I volunteered to participate to provide a web developer's expectations. 

Francis

David Benn
David Benn's picture
Sanity Check, web service API documentation

Hi Francis

Okay, cool. Having made heavy use of the web service for VStar, I'd be happy to help provide input/feedback. George, if you're reading this, let me know if you want input/review.

I've tested, implemented and documented a RESTful web service or two in my work life as well.

David

pukemaru
pukemaru's picture
TT ARI

Mike, I just selected VISUAL for TT ARI going back to 1975 and the graph displayed all the information. I had to wait awhile. I wonder if this has been fixed.

Cheers

Stephen [HSP]

lmk
lmk's picture
"fainter than" stays active in band select

Another ssue, even I select just one band and nothing else, the fainter thans still show up in the plot, and skews the y-axis to an inappropriate scale. Please fix! Thanks.

 

lmk
lmk's picture
Short timespans come back "no data" too!

If I use a short time span for calendar dates, say under 24 hours (?) It comes back with that quirky "no data" error, even though there is definitely data within that interval as can be verified by expanding the time interval. Please fix. Thanks....

fhemsher
Single observation: cannot plot

If only one observtions happens to be in the selected time span, then the LCG does not plot.

I've added a message to that effect to be included next week.

Francis

HQA
HQA's picture
data limit

When I try to plot a light curve (such as KIC 8462852) from the home page "Pick a Star", it fails, with the error message  "maximum 30000 records exceeded".  This is a VERY small limit, as there are many stars with far more than 30K observations.  LCG1 does not have such a limit.  If this is a true limit, then you might put up an error message, but plot the most recent 30K datapoints.  Also, "Plot a Light Curve" box has three options for date: From Date, To Date, and Plot Previous.  It is unclear how these interact with one another, and an option to set the "To Date" to the current JD is not available (say that you erased it for some reason, and now want to set it to current).  Plot Previous also only has space for 999 days, and quite often for LPVs you want to look at a longer time span.  I suggest enhancing this aspect and describing somewhere how to use the entire data entry box.

Arne

PYG
PYG's picture
Default

I don't wish to spoil anybody's party, but can we have a fully functional LCG as the default please (i.e. version 1) and leave version 2 to the background for those who wish to play.

 

HQA
HQA's picture
hanging

another, probably related, problem is that the LCGv2 seems to hang sometimes when accessing data.  For example, for omi Cet, I can use the Plot Another Curve box and select the last 2000 days, and it plots fine and quickly.  If I choose 3000 days, it hangs.  Perhaps others can check this so that we can eliminate my computer/FireFox browser as being the culprit.  I'm guessing it is because at 3000 days, more than 30K datapoints are being retrieved from the AID, but no error message appears.

Arne

fhemsher
Hanging fixed

Hi Arne,

This was fixed. Also, Sara & I have been reconsidering the 30,000 data points limit. I've been testing the removal of the limit and it looks good. The LCG responds nicely if certain features are disabled when a ton of points are requested. The List of Contributors, the "Box" , Mean Curve, and Annotate Data are disabled when a large number of points are returned. We'll let you know when this change is in effect.

Francis

lmk
lmk's picture
Should be easy fixes

Arne, the "plot previous" does allow entry >999, it just scrolls out of the box, so the box should be expanded to show all the digits. You have to check the box besides it, and that appears to activate that option and set the actual from/to dates.

The problem I was seeing, is the from/to dates do not match up with the output of the plot! You can get very quirky results, it can skip data rather than plotting it between the from/to dates. And, shorter from/to dates, say a few days or less than a day, seems not output what you expect at all! I get "no data" errors, sometimes, but other times it works... This is clearly a bug in the program.

If there is a limit of 30k data points, that is not acceptable. Long time spans are frequently plotted, and it works with LCGv1, so it appears that the program needs to use dynamic memory allocation, rather than a fixed array size of 30k? That's my guess of what the problem is.

 

TRE
TRE's picture
I like the star request

I like the star request history feature. Continueing to explore what it can do. Some snags:

When going from from the home page with a star request, it does it with some default time span. Then one wishes to expand or limit the time axis. However, the star name needs to be re-entered. It would be nice to have it remembered. It is easy to retype GSC names but things like KIC names get long.

A chart format thing; I asked for the last 40 days of NQ Gem and got 20 days, 2457744 to 2457764, with the first and last date data sitting on the chart edges. It's not ideal charting form. There should be some white space between the edges of charts and it's first or last data points. At a minimum, the requested time base should be provided.

If I then want to look at the last 1 days, It doesn't read the number in the days box unless I check the green days box. If there is no data available, I get a red-on-blue text No Data Found. Meanwhile, my eyes have darted to the plot, which doesn't get updated. Then i go back to the form and spot the No Data Found.

Next, I am trying to figure out what "annotate data in print" will do. Will there be a Help file for some of this?

Ray

fhemsher
Requested time span truncated

Hi Ray,

If there are no observations at either end of the requested time span then the LCG does not show empty space along the x-axis extremities. The reason for this is to assure that the returned observation symbols have as much area as possible for viewing, rather than clumping together. Sorry this is confusing. I'm hoping that once this realized by LCG users it won't be a problem.

Francis

TRE
TRE's picture
Actually, I strongly prefer

Actually, I strongly prefer to see the base line I asked for.

Don't know if you ever used an network analyzer or an oscilloscope.

There, you ask for a baseline and you get it. Makes it much easier to quickly interpret signals without making mistakes.

If you put up a plot of data on an overhead at a meeting and say "this is last month's data", you better have a baseline that is 30 days. Otherwise, besides the usual question of "where are the error bars?" you will get  "You only have ten days  on the plot, what happened douring the other 20 days?" I ask myself that same question every time I use LCG2, before i remind myself that it is an non-standard base line. 

If I want to zoom in on data sections, I make the baseline/timebase smaller. Some programs like LCG2 also have a nice zoom feature for that.

I know cosmoligists and astronomers are famous for poor plotting (like invisible greek captions and axis lables that are unreadable from the front row, sparse data on plot boundries, etc). Do we need to perpetuate that tradition here?

A good guide to good plotting might be the many wonderful popular spreadsheets.

Ray

spp
spp's picture
Observation Data box; 3rd decimal place 0's,

In the Observation Data box the font is too large for the box on my Mac OS X 10.9.5 /Firefox 50.1.0.  Could the font be made smaller or be made adjustable in Preferences?

Also in the Observation Data box ,magnitude estimates and uncertainties which were reported with a "0" in the third decimal place appear with only 2 decimal places.  In the observation shown in the attached screen shot, magnitude 12.140 shows as 12.14 and Err of 0.010 shows as 0.01.

Phil 

fhemsher
Screen Resolution

Hi Phil,

I looked at Screen Shot. The ideal screen resolution for the LCG is 1280 x 768. Perhaps you could check your current resolution and see if changing it would correct the problem. Also, Firefox has the option to change font size for the web page. See Preferences>Content>Fonts & Colors. The default font size is typcially 12. Possibly that could modified.

Hopes this helps,

Francis

spp
spp's picture
screen resolution

Thanks, Francis

In Firefox, I reset the font size to 12.  That fixed the problem, and the entire Data Box looks completely different (much better).

Phil

 

 

pukemaru
pukemaru's picture
Default

I like Gary Poyner's post above. Keep v1 as the default. I use LCG a lot and prefer that.

Stephen (HSP)
New Zealand

cutterboyus
cutterboyus's picture
New LCG

Although I was a little bit doubtful in the beginning about the new light curve generator I am beginning to come around and like it quite a bit more.  I do hope you keep both versions as I find the LCG v1 to be best for some of my stuff.

The graphic overlay in the new version that shows a particular observer's observations is a bit heavy handed!

Good work overall.

MJHA

 

HQA
HQA's picture
minor changes

I'm using LCGv2 a bit more this week, as people ask questions.  I'd make a couple of additional changes.

First, in the Preferences box, you can set whether to use JD or calendar days - great!  However, the change is not "sticky".  If you do another plot, even of the same star, it reverts back to JD.

Second, if you click on a point to get the Observation Data window, the fields "Comp" and "Check" from the Extended Format are renamed "Comp Star 1" and "Comp Star 2".  I think you should keep the nomenclature the same - call these Comp and Check.  Or better yet, use the Extended Format labels of Cname and Kname.  Comp Star 2 is not correct, as this star is not being used in the determination of the standardized magnitude.  Renaming these entries will also make it easier to relate to the subsequent CMag and KMag fields.

Arne

fhemsher
Comp & Check

Hi Arne,

The following is a comment from Sara Beck to me:

The field you have been calling "Comp Star 2" actually has dual purposes. For CCD, DSLR and PEP type observations it should be called "Check Star" or simply "Check". For Visual, PTG or VISDIG observations it should be "Comp Star 2" or simply "Comp2". If you don't want to change the label for the different obstypes, then I think it would be best just to it "Comp2/Check". The field you are calling "Comp Star 1" could also be shortened to simply "Comp".

I changed Comp Star 1 to 'Comp', and Comp Star 2 to 'Comp/2/Check'. This will show up in the next update. Thanks for your comments.

Francis

HQA
HQA's picture
comp2/check

Hi Francis,

That should work.  My preference would be a dynamic adjustment of the labels depending on the obstype, but I also know that such changes cost programming time (but might be fun to do, if you have the right mindset).

Arne

WI
Here's a suggestion. Three

Here's a suggestion. Three times I have gone to the CreateSpace service to order back issues of the JAAVSO, and each time, I have spent a long, frustrating time trying to get there from the AAVSO site. I click on Publications, then on Journal, and I get the front cover of the latest issue and nothing else. There is no line for "back issues" or "CreateSpace" on either the Publications or the Journal pages. It seems a little malicious to tell us how we can purchase back issues and then not provide any way to accomplish that task.

weo
weo's picture
Print issues of JAAVSO answer for WI

Hi David,

I'm sorry you are having trouble accessing the CreateSpace for ordering a print copy of JAAVSO, but I'm not sure why. Are you logged in when you try to click through to the CreateSpace site? Here's what I just did that worked fine: 1) While you are logged in, on the home page, click on the JAAVSO icon (located beside the Stellar News Feed); 2) click on JAAVSO Volumes; 3) click on the picture of the issue you want; 4) on the page that comes up for that issue, click on Order a printed copy; 5) from the list of issues available, click on the issue you want; 6) you should be taken to the CreateSpace page for that issue and given the opportunity to add it to your shopping cart.

It should be as simple as that. If it doesn't work for you, could you please email me off-list (eowaagen@aavso.org) to tell me where it is breaking?  Thanks!

best wishes,

Elizabeth

TRE
TRE's picture
comment limit is 100 chars

I just noticed that LCG2 doesn't like to display the long comments from TA. It says it's limit is 100 charactors.

fhemsher
LCG Updated

Today various changes have been incorporated.

Removed the 30,000 data points limit.

We've added "All Days" feature - you may request all observations for any star.  You will not receive a 'No Data Found' error message anymore. (This is OK for all stars except SS CYG which exceeds the AAVSO data request size limit.)

Also we've considered all of your previous comments and suggestions. Hope the results are satisfactory.

Please continue to make further suggestions.

Thankyou,

Francis

lmk
lmk's picture
Scaling on the axes needs manual control

I still find the auto-scaling features here to be fairly annoying. Along the time axis, it always adjusts the data to fill the full plot width, there is no way to override it manually. This makes it difficult to compare standard time spans of other stars, etc. A star with a lot of data will auto-adjust pretty closely to your desired range, but one with sparse data will auto-adjust to a totally different time scale, and I cannot change it. This is not acceptable for a general plotting tool. You can't generate sensible scaled plots for publication!

Secondly, the "fainter than" symbols always appear when first generating the plot, even if you uncheck them, yes they go away, but the original y-axis scaling stays the same. So, you end up with a huge amount of empty space in the plot, and all the data compressed together. It would be nice to add the "fainter than" as the first option in the band selects option, so they can be eliminated from the plot right off the bat.

I do see the preferences now allows manual adjusting the magnitude axis, can you allow this for time axis too, please? Thank you!

 

fhemsher
Manual Selection - Requested Time Span

I'm considering this: Under 'Preferences', including a checkbox titled "Requested Time Span"

If the checkbox is selected, then the x-axis is expanded to the original to/from values if not already there.

This feature could only be selected if the original chosen date format is Julian. Also, the Box feature is disabled. You may uncheck the checkbox to return to the truncated time span, and enable to Box feature.

Is the above suitable for your needs?

lmk
lmk's picture
Not exactly

Thanks for the quick response! Well, I don't see why it simply cannot be made either a user selection of the time span, or the default, the program-derived values time span? It seems it would just be the xmin-xmax coordinates of the time axis passed to the plotting routine. Why must you restrct to JD and remove other options necessitate changing other things?

fhemsher
Manual feature for publications/comparisons

You stated that you needed a manual feature (similar to Magnitude) primarily for publications and comparing various stars during specific time spans. I believe that can be acheived. We're posting this to site today, so please it a try.

Thanks,

Francis

fhemsher
LCG Update

Hi All,

Today we've included a new feature. It can override the default truncated time span where it does not include empty space along the x-axis extremities when no observations are present. This a available  via a check box at: Preferences > Requested Time Span. 

Also, the 'My Star History' has been improved. It will now remember if you have chosen a specific time span where the 'to' time is not the current time. When the star is called again it will show that specific to/from time. Before, it was calculating only the span and subtracting from the current time.

Regards,

Francis

lmk
lmk's picture
I see that new feature, but

I see that new feature, but it is "greyed out", I cannot try it!

 

fhemsher
Will Change Grayed Out

Yes, I can see why you believed you could set that value. It is shown as a text input element, which can be misleading. The values From/To are those which you had originally requested for star before truncation, and cannot be changed. They let you know the values when x-axis will be expanded, after you click the checkbox.

During the next update, I'll replace the text element with just the values. See attached image.

File upload: 
nmi
nmi's picture
Continued LCG improvements - Thanks

I've been using the tool for weeks now and really like the ease of use -Thanks

Mike

TRE
TRE's picture
Got another small snag.

Got another small snag.

The requested time span doesn't calculate right for periods less than an integer day.

Ray

fhemsher
LCGv2 - Julian Date Decimal Places

Hi Ray,

Your are correct. The requested time span for LCGv2 does not handle Julian Days with a time interval less than 1.

Francis

TRE
TRE's picture
Thats OK. LCGv1 handles it

Thats OK. LCGv1 handles it nicely. I use it often.

Ray

 

lmk
lmk's picture
Sorry, it still isn't working

Sorry, it still isn't working right. If I specify a time interval, it still fills the entire plot width with the actual max/min dates of the stars found within the specified time span, NOT the actual time span I specify!

Also I notice in the preferences, if I check the JD radiobutton, the scale changes, but if I then check the calendar radiobutton it won't change back!

 

fhemsher
Requested Time Span - Active after requested plot.

Hi Mike,

Truncated time span is always initially displayed when a star request is plotted. Expanding the x-axis, via the 'Requested Time Span' checkbox must be selected after the plot is loaded.

Thanks for pointing out the disabled Calendar Date radio button (Sara Beck also pointed that out). I'll send a fix early next week.

Francis

lmk
lmk's picture
The requested time span

The requested time span checkbox is still disabled!?

 

David Benn
David Benn's picture
Multiple clicks of Plot Another Curve button

If I select Plot Another Curve, then again (and again ...) after the dialog is open, the dialog box grows downward.

I see this in Safari and Chrome.

David

David Benn
David Benn's picture
Dialog box opening order

If I open the Plot Another Curve dialog box, followed by the Preferences dialog box, the latter is opened behind the former and can't be repositioned. It can be closed however. The opening order seems to be fixed.

This and the last comment are minor UI issues vs functionality problems.

I see this in Safari and Chrome.

David

fhemsher
Minor UI Issues

Hi David,

Fixing these issues is important. Addressing the many permutations of user selections is always a challenge. I encourage comments that continue to point out these issues. That being said...

1) The growing 'Plot Another Curve' pane will be fixed in the next update.

2) The 'Preferences' pane is intended to be placed behind the 'Plot Another Curve' pane. This implies that Preferences relate to the current star plot, not the next curve to be plotted.

Thanks,

Francis

David Benn
David Benn's picture
Challenge

Hi Francis

Thanks for that; I appreciate the challenge.

David

fhemsher
Known Issues

Currently there are five(5) issues that are being addressed and will eventually fixed:

1) Last Observation Received - At the AAVSO main page: Selecting the 'Plot <Star Name> light curve' cannot be clicked. This happens periodically with users who have  Windows7/8/10 running on their computers.

2) Plot APASS Data - This checkbox remains disabled under 'Preferences'. The apass database is not yet usable in the LCG .

3) Report Discrepancy - This feature remains disabled. It is intended to allow users, who login, to report any observation to Zapper....FIXED (2/4/2017)

4) AAVSO main page: Last Observation & Plot a light curve - Users with Windows operating system will see an empty plot grid below the requested curve....FIXED (2/4/17)

5) Your Star History - There is a limit on the number of 'history' cookies at the server. If the limit is exceeded, the LCG will receive an error message "Bad Request"....FIXED (2/4/2017)

HTY
HTY's picture
API Interface to LGCv2?

Can we access LGCv2 via HTTP GET?  If so, is there a spec online somewhere?

Tim

Pages

Log in to post comments
AAVSO 49 Bay State Rd. Cambridge, MA 02138 aavso@aavso.org 617-354-0484