Comments (11)
Hi. I've already made an implementation for this (based on the exocomets proposal) in my TLS fork. Do you want to check it? Maybe if you found it as a good solution (which might be extensible to any other transit templates), I could create a Pull Request to merge it into the official TLS repository. If you also think that the solutions is not good enough yet, please let me know what additional changes could I do to step forward towards a Pull Request.
Kind regards.
from tls.
Cool! I'm very interested in adding this to the main branch of TLS.
Can you elaborate on the features you added, the API, etc? Perhaps even write some documentation on how it works and what it can do? An example to showcase the feature would be cool.
from tls.
Great. I'll come back with a good explanation, a notebook containing examples and some documentation about the code as soon as I can. Do you have any prefered format for the documentation?
from tls.
Wonderful! You could build on the existing documentation (source, ReadTheDocs) and link an iPython notebook?
from tls.
Great. I will do so. I'll let you know when I'm finished.
from tls.
I think that I have added a good start to the documentation of the feature in both places. Please have a look at the notebooks that I added and to the light api changes and let me know whether you think that I should add more info or modify the one I'm providing.
In general terms, I have moved the template related logic to a class named TransitTemplateGenerator, which is implemented by DefaultTransitTemplateGenerator for planetary transits. There is also CometTransitTemplateGenerator now. Finally, the user can decide to provide his custom implementation of TransitTemplateGenerator to search for custom templates. You can see in the last notebook that I tried a template for flares.
I had to modify tls logic which assumed that the templates only caused flux drops, and now flux increments would be supported.
The main issue I'm concerned about right now is the transit depth adjustment, which can be inexact for several reasons. One of them is the overshoot parameter, whose meaning I dont understand. I think that you created an issue to provide depth fits instead of an analytical one and that might be related too. However I think that the solution works properly enough now and the things that I'm commenting here could probably be developed in the future.
Kind regards.
from tls.
As soon as I can I will create some tests and will open a pull request so you can better inspect the changes.
from tls.
Sorry @martindevora again for the delay. Schools are still closed which leaves me hardly any time. I hope the situation will get better in February. Current estimate is either 1 Feb or 15 Feb for re-opening.
from tls.
There is no hurry at all @hippke . I'm adding a few improvements into the new API so we can have a more maintainable code. The most important thing right now is to keep your family and yourself safe and healthy. I'm doing as well. I would probably like to use these changes as a base to develop some kind of publication at some point in the future and I would like to know what you think about it.
As already said, there is no hurry, so don't worry about this until you feel like you can do it happily without pressure.
Kind regards,
Martín.
from tls.
I think it's a great idea worth implementing in code, and worth searching for in real data. I'm not a specialst of flare-finding algorithms, but the few I have seen are rather simple. So it appears useful to try another approach.
I'm a bit sceptical with regards to periodicity. TLS has built-in phase-folding, so that's sort of given. Many flares and comets might be somewhat periodic, e.g. due to stellar rotation, but in many cases not perfectly periodic. As you have shown in your example, TLS can still find such things, but the sensitivity will be reduced. If you want to find stuff with it, it might work! But you should compare sensitivities to other algorithms out there, to check in which use cases TLS+mod is good, and in which it is bad.
from tls.
Yes totally. I'm planning to compare TLS against such other algorithms (with its improvements and counterparts) and with the standard transit models in case of tailed transits.
I was also wondering in using a final_fit method within TLS to add the option to the user of adding a complete custom fit for the given template, which could be a scipy.curve_fit for instance (or whatever the user would like to introduce). What do you think about it? I guess that it is not strictly necessary as it can be done after TLS returns the results, but it would be a way to unify the final fit processing for all the search algorithms.
Kind regards.
from tls.
Related Issues (20)
- catalog_info_ returning the errors for radius/mass_min/max instead of the min/max values HOT 1
- Bug
- TLS failing for Kepler prime mission data HOT 5
- Dependency on batman / numpy makes it impossible to install TLS HOT 3
- MultiProcessing Bug with M1 Pro chipped Mac books HOT 2
- Non-integer oversampling factors cause index error in TLS pipeline
- Why TLS use duration from minimum chi2 period instead of maximum power period? HOT 6
- Setting minimum duration limit HOT 2
- maximum orbital frequency formula derivation HOT 4
- Occasional buggy spectrum if array of equal errors given HOT 3
- Bug: calls numpy.int which is deprecated since numpy version 1.20 HOT 2
- Period Grid error in stellar mass (maybe radius too?) > 1 HOT 13
- Insert Sort on previous trial period
- Bug: Default Limb Darkening Coefficients overriding input HOT 2
- Smooth the periodograms with log-normal biweight
- Bug: Can't install due to batman-package failure HOT 6
- Relax minimum period HOT 4
- Bug: expected dtype object, got 'numpy.dtype[float64]' with numpy 1.21 HOT 4
- Add `verbose` kwarg to transitleastsquares HOT 6
- Bug running multiprocessing. HOT 8
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from tls.