Skip to content

new option in cfg file: ltcube and minos#92

Open
dcfidalgo wants to merge 3 commits into
gammapy:masterfrom
dcfidalgo:ltcube
Open

new option in cfg file: ltcube and minos#92
dcfidalgo wants to merge 3 commits into
gammapy:masterfrom
dcfidalgo:ltcube

Conversation

@dcfidalgo

Copy link
Copy Markdown

I know that there is a very similar request, but i think it is worth reconsidering.

  • New option 'ltcube' introduced in the cfg file. Now you can specify a live time cube that will be used for the analysis.
  • For the analysis in the different energy bins, enrico now uses the ltcube from the global fit instead of producing the same file x times. This was very easy to implement with the new ltcube option.
  • New option 'minos' introduced in the cfg file. Now you can choose if you want to calculate MINOS errors or not. I found that depending on your analysis it can take a very long time, so i thought it would be nice to have the possibility to skip it.
    Cheers,
    David

dfidalgo added 2 commits May 10, 2016 19:19
…ault.conf:

 - Now you can specify a ltcube in the config file. Enrico will skip the
   calculation of the ltcube and use the given file instead.
 - When doing the analysis for the spectral points, Enrico will now use the
   ltcube from the global fit, or the ltcube specified in the config file.
 - Now you can choose if you want to calculate MINOS errors or not through the
   config file.
@davidsanchez

davidsanchez commented May 19, 2016

Copy link
Copy Markdown
Member

Hi thanks for the contribution.
If you give the option for the user to use an ltcube, you should give him the possibility to either compute it or download it

In the config file generation, this has to be also transparent for the user.
cheers
david

@dcfidalgo

Copy link
Copy Markdown
Author

Hi David,
I guess the idea was that if someone already computed the ltcube by themselves or wants to use the same ltcube for several analyses, they can do so and save a significant amount of time (especially when using a full, now almost 8 year data set).
But since both options (ltcube and minos) are both not vital for the analysis in the sense that they only speed it up, i would not include them in the std config file and just mention them in the docs. In this way the config file stays manageable, but users who bump into the same "issues" that i had can search the docs.
Cheers,
david

@davidsanchez

Copy link
Copy Markdown
Member

Hi

in the std config file you can have them with default value as you defined in the default config file.

It would be nice to have some tool to compute the ltcube before the analysis (as enrico_download is use to download weekly data), maybe put it at some place where enrico can get it easly when creating a config file

cheers
david

   * For the spec model PLSuperExpCutoff the cutoff_min was to high for some
     sources in the 3FGL. Reduced it to 100.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants