idl-coyote / coyote Goto Github PK

In cgimgscl.pro, at line 179: CONSTANT=contant

What steps will reproduce the problem?

1. Obtain an image with a horizontal 'blob' (i.e., single row)
2. Run fit_ellipse on said image
3. Fail

Can be reproduced with the following code:
IDL> x = bytarr(64, 64)
IDL> x[10:24, 32] = 1
IDL> res = fit_ellipse(x)
% Compiled module: FIT_ELLIPSE.
% Compiled module: ARRAY_INDICES.
% TOTAL: For input argument ARRAY, Dimension must be 1.
% Execution halted at: FIT_ELLIPSE 157
....\coyote\fit_ellipse.pro
% $MAIN$

NetCDF attribute names are case-sensitive, IDL tag names are case-insensitive and this may lead to duplicate name errors. In NCDF_data__define.pro, the check at line 1947 should be repeated at line 1974.

In cgimagegroup__define.pro the call to ClipScl(original) (line 63) should be replace with cgClipScl(original)

Hi all,
I have a netCDF file with a variable which has both attributes _FillValue and missing_value.
In the GetInfo() function for ncdf_variables, lines 622-634, the info is constructed in such a way, that the _FillValue tag is added to the info structure whenever _FillValue or missing_value is found as a variable attribute. This is what the code looks like:
attrNames = self -> GetAttrNames()
attrIndex = Where(attrNames EQ 'scale_factor', count)
IF count GT 0 THEN info = Create_Struct(info, 'scale_factor', $
self -> GetAttrValue('scale_factor'))
attrIndex = Where(attrNames EQ 'add_offset', count)
IF count GT 0 THEN info = Create_Struct(info, 'add_offset', $
self -> GetAttrValue('add_offset'))
attrIndex = Where(attrNames EQ 'missing_value', count)
IF count GT 0 THEN info = Create_Struct(info, '_FillValue', $
self -> GetAttrValue('missing_value'))
attrIndex = Where(attrNames EQ '_FillValue', count)
IF count GT 0 THEN info = Create_Struct(info, '_FillValue', $
self -> GetAttrValue('_FillValue'))

However, that will create a "Conflicting or duplicate structure tag definition" error if both attributes are there.
I think, one could either put both attributes into the info structure, or avoid the error by filling the _FillValue tag in the info structure with the _FillValue attribute value as the priority, doing something like this:
attrNames = self -> GetAttrNames()
attrIndex = Where(attrNames EQ 'scale_factor', count)
IF count GT 0 THEN info = Create_Struct(info, 'scale_factor', $
self -> GetAttrValue('scale_factor'))
attrIndex = Where(attrNames EQ 'add_offset', count)
IF count GT 0 THEN info = Create_Struct(info, 'add_offset', $
self -> GetAttrValue('add_offset'))

attrIndex = Where(attrNames EQ 'missing_value', count_fill)
attrIndex = Where(attrNames EQ '_FillValue', count_miss)
IF count_fill GT 0 THEN info = Create_Struct(info, '_FillValue', $
self -> GetAttrValue('_FillValue'))
IF count_miss GT 0 AND count_fill EQ 0 THEN info = Create_Struct(info, '_FillValue', $
self -> GetAttrValue('missing_value'))

Thanks for the support, cheers
Achim

Hi all,
since commit "Small changes in a number of files." (af66b7c), graphics are saved upside down when using this code:

data = cgDemoData(2)
LoadCT, 33, NCOLORS=12, BOTTOM=1
cgContour, data, NLEVELS=12, C_COLORS=Indgen(12)+1, /FILL, /WINDOW
cgControl, CREATE_PNG='some file'

I had come across a similar problem earlier in a different context. There, the issue was related to the !ORDER system variable. In this case, changing it, doesn't seem to make a difference.

Thanks for the support!
Cheers, Achim

Each time, we launch the cgPS_open, we rise an error message (line 227)
% Program caused arithmetic error: Floating illegal operand
the concerned line is
IF N_Elements(struct) EQ 0 THEN ps_struct = {cgPS_SETUP}

I tried with ISA but it doesn't work...
IF ISA(struct,'CGPS_SETUP') EQ 0 THEN ps_struct = {cgPS_SETUP}

In my code, I call at least 20 times this routine, it is a bit tiresome to get the same error messages 20 times but it doesn't have a impact on the results.

Thanks very much if somebody could improve this line or suggest a code modification or a missing (perhaps?) module in my directory to avoid it.
Best regards
A.B.

In cgimgscl.pro, lines 340 - 344:

          4: BEGIN ; Log scale stretch.
             tempImage =  cgLogScl(tempImage, Max=maxvalue, Min=minvalue, $
                       Mean=mean, Exponent=exponent, Negative=negative, $
                       OMIN=bottom, OMAX=top)
             END

should be

          4: BEGIN ; Log scale stretch.
             tempImage =  cgLogScl(tempImage, CONSTANT=constant, NEGATIVE=negative, $
                        MAX=maxValue, MIN=minValue, OMAX=top, OMIN=bottom)
             END

With the last merge, the file cgfixps.pro is broken due to an unresolved merge conflict:

<<<<<<< HEAD
  Get_Lun, in_lun
  OpenR, in_lun, in_filename
=======
>>>>>>> 8ba3a7b34691137b0d0b3d3f508b434ded51fa12
  IF (FStat(in_lun)).size EQ 0 THEN BEGIN
    Free_lun, in_lun
    Free_lun, out_lun
    File_Delete, out_filename
    ;Print, 'Zero Length File Encountered...'
    RETURN
  ENDIF
  
<<<<<<< HEAD
  ; Move along in the file until the end of the Prolog.
  line = ""
  count = 0
  target = "void"
  buffer = StrArr(100)
  
=======
>>>>>>> 8ba3a7b34691137b0d0b3d3f508b434ded51fa12

What steps will reproduce the problem?

1. detects = obj_new('blob_analyzer', large_image, /all_neighbors)
2. stats = detects->getStats(index_of_blob)
3. stats.center[0 or 1] < 0

What is the expected output? What do you see instead?
The index of the blob center is negative if a large image is used (pixels/columns > 32K) and the blob is resides in a region where one of the indices is > 32k.

What version of the product are you using? On what operating system?

Please provide any additional information below.
Problem is due to indgen() use in find_boundary().

xcm = Total( Total(mask, 2) * Indgen(xsize) ) / totalMass
ycm = Total( Total(mask, 1) * Indgen(ysize) ) / totalMass

suggest using Lindgen() instead.

9fc0811 introduced a bug that breaks some existing code. IDL can't distinguish between LABEL and LABEL_WITH_DEGREE because the former may be a shortened version of the latter. Since there already exists a LABEL keyword there's now a conflict. The new keyword needs a different start to the name (maybe DEGREE_LABEL?).

Reproduce with:

IDL> cgmap_set, label=1, /grid

Traceback Report from CGMAP_SET:

     % Ambiguous keyword abbreviation: LABEL.
     % Execution halted at:  CGMAP_SET         617

In cgdisplay.pro at line 287 ps_struct.landscape is used on the LHS, but ps_struct is undefined here.

In line 137, the branch for the Mac uses the nonexistent variable "decomposedState" rather than "decomposed" as in the other branches.

Corrected version available at https://idlastro.gsfc.nasa.gov/ftp/pro/coyote/cggetcolorstate.pro

Wayne Landsman/Gilles Duvert