Project

General

Profile

Actions

Feature #4286

closed

Undocumented "multifile 2" mode

Added by Johan Richard over 6 years ago. Updated over 6 years ago.

Status:
Closed
Priority:
Normal
Assignee:
Category:
-
Target version:
-
Start date:
2017-09-22
Due date:
% Done:

100%

Estimated time:
Resolution:
fixed

Description

Regarding the int number next to multfile in "image" subsection of the parameter file.

In the current documentation only options 0 and 1 are documented (true/false flag).

However there is a specific case (in o_global.c) when this int=2 : the required format
for the multfile is different (need the shape error?)

In addition, it seems the variable I.n_mult is later used in lenstool as the number of
multiple systems, correct? It is a bit confusing if we want to use I.n_mult as the format
of the multfile (not the number of multiple systems): should we duplicate the variable?

Actions #1

Updated by Eric Jullo over 6 years ago

  • Status changed from New to In Progress

Thank you for mentioning the issue.

It looks to me that we prefer the integer to specify the format of the input catalog, rather than the number of image systems to consider in it.

Today, the code considers the integer as a specifier of the input catalog format, but the in the bestopt.par, we write the number of image systems. This is inconsistent.

I'd rather change the code that writes the bestopt.par file.

What do you think?
Eric

Actions #2

Updated by Johan Richard over 6 years ago

I agree, but do I understand correctly that there are many loops over I.n_mult (using it as the number of multiple images)?
It seems easier to read this integer as a different variable name in r_image.c

In any case currently the code accepts :
0 = nothing
any number >0 but different from 2 : usual multiple image format
2 : "shape" multiple image format.

We should only accept 0, 1, and 2 and otherwise return a warning (format not supported) just for consistency.

Sorry there are kind of 2 issues mixed together (and also 3 if we count the documentation).

Thanks!

Johan

Actions #3

Updated by Eric Jullo over 6 years ago

The format definition in f_shape is * flag=1 : ID RA DEC A B THETA Z MAG (default) * flag=2 : ID RA DEC A B THETA Z MAG VARE1 VARE2 (for WL or source with Sersic index) * flag=3 : ID RA DEC E1 E2 Z VARE1 VARE2 (for WL)

For multiple images, if n_mult=0, nothing happens, if n_mult=2, then we use flag=2 format, otherwise we use flag=1 format.

In any case, only n_mult systems are used.

It seems natural to me to specify the format of the input catalog in front of the catalog name, as an argument of the multifile keyword. Then, we can add another keyword in the image section with the number of multiple systems to take into account (the actual n_mult variable)

What do you think?

Actions #4

Updated by Johan Richard over 6 years ago

But flag=3 is (I believe) not coded? Only flag=2 is used (in o_global) for the format.

I think we agree that we need two variable names. I don't think specifying the n_mult in the parameter
file is very useful, so just storing the "file format integer" in a new variable is necessary, and also
if we want to have flag=3 used it needs to be coded?

Actions #5

Updated by Eric Jullo over 6 years ago

flag=3 is coded, but not used for SL (it is used for WL).

So, shall we create a new variable in the image section, let's say I.multmode that contains the format of the input catalog?

Already, for WL we have I.statmode that contains the arclet catalog format.

Actions #6

Updated by Johan Richard over 6 years ago

  • % Done changed from 0 to 100
  • Resolution set to fixed

Corrected in revision r1395 and documentation

Actions #7

Updated by Johan Richard over 6 years ago

  • Status changed from In Progress to Resolved
Actions #8

Updated by Johan Richard over 6 years ago

  • Status changed from Resolved to Closed
Actions

Also available in: Atom PDF