Re: [ferret_users] Pyferret OS X

[Torben Mueller <torbenmllr@xxxxxxx>]

Hi Karl,

I am def. interested in trying out the PyFerret beta once you have a OS x build.


On 19 July 2013 13:01, Karl Smith - NOAA Affiliate <karl.smith@xxxxxxxx> wrote:
> Hi,
>
> I just finished (yesterday) putting together the beta version of the latest
> PyFerret (1.0).  If anybody wants to try it out and give me some feedback,
> let me know.  So far, I have just built and tested on RedHat and Ubuntu
> Linux systems, but I do want to get an OS X version put together.
>
> Karl
>
>
> Brief (?) release notes for 1.0:
> (1) Added -nodisplay command-line option (since you don't need a display
> when run in this mode); deprecated -gif, -unmapped, -batch, -transparent
> (2) SET GRAPHICS is gone, replaced by SET WINDOW options; in particular
> /QUALITY={HIGH, DRAFT}, default HIGH
> (3) SET WINDOW: added /QUALITY= /ANTIALIAS /NOANTIALIAS /COLOR= /OPACITY=
> /THICKEN= /XPIXELS= /YPIXELS= /XINCHES= /YINCHES= /TEXTPROM= to the other
> SET WINDOW options
>     (3a) /COLOR= /OPACITY= sets window background color and opacity
>     (3b) /XINCHES=,/YINCHES= gives full plot size in inches.  Default size
> of vector (PS, PDF) plots.
>     (3c) /XPIXELS=,/YPIXELS= gives initial scaling factor to viewer so
> displayed image is this size in pixels.  Default size of raster (PNG) plots.
>     (3d) Default /XIN=10.5, /YIN=8.5 to fit letter-sized page; all margins
> 1.375 inches to keep (close to) /asp=0.75:ax;
>     (3e) /SIZE=0.7, as in Ferret.  Display's DPI (dots per inch), or 120 DPI
> for -nodisplay, used if /XPIXELS and /YPIXELS are not given
>     (3f) over-specification of sizes (e.g., /XIN=10.5, /YIN=8.5,
> /ASP=0.75:axis) gives an error
> (4) SET WINDOW /THICKEN= scales all line thicknesses (which are now in
> points = 1/72 inches); default 0.75 to match ferret displayed thicknesses.
> (5) All heights of automatic labels scaled according window size using
> /XINCHES= and /YINCHES= values.  Thus SET WINDOW /TEXTPROM= brought back,
> but now actually does scale the height of all automatic text.
> (6) automatic scaling of the image to fit the displayed window frame
>     (6a) forces window frame to be correct aspect ratio, unless window is
> maximized
>     (6b) can be turned off and back on from the Image->Scale menu item
> (7) SHOW WINDOW shows more info (quality, displayed, antialias, thicken
> factor, text prominence factor)
> (8) FRAME: added /XINCHES= /YINCHES= /XPIXELS= /YPIXELS= to override default
> image sizes from SET WINDOW
>     (8a) enforces same aspect ratio, so only one of /XINCHES and /YINCHES,
> and only one of /XPIXELS and /YPIXELS, can be given
>     (8b) PS and PDF output no longer scaled (or centered) to a letter-sized
> page
> (9) Image->Save As... no longer brings up the print dialog after selecting
> the PS or PDF format in the filename box.
>     (9a) When high-quality displayed (Qt), PS and PDF always saves to a
> letter-sized page, so a large image could get cut off
>     (9b) With -nodisplay (Cairo), PS and PDF saved are the size of the
> image, so never cut off and no extra whitespace
> (10) FRAME /FILE=... /FORMAT=... can now also be issued any time, using any
> supported format, when running with -nodisplay.  Eliminates the need for
> -batch and set mode meta; eliminates most differences between high-quality
> displayed and -nodisplay sessions.
> (11) SET WINDOW /THICK= scales symbol sizes as well as line thickness.  Only
> future drawing in a window affected by changes in the value.
> (12) CONTOUR (including FILL), PLOT, VECTOR, SHADE, POLYGON: added /OPACITY=
> (13) /OPACITY= overrides color opacity, in fill only if simultaneous filling
> and line drawing (i.e., contour /fill /line /opac=10 /l=1 sst has faint
> color but normal contour lines)
> (14) CONTOUR, PLOT, VECTOR, POLYGON: enhanced /COLOR=, /THICK=
> (15) palettes and /COLOR= options now accept RGB 3-tuple (gives opaque
> color) or RGBA 4-tuple
> (16) /THICK= can (in most places) be any positive value, and is in points
> (1/72 inches); will override thickness from line numbers
> (17) PPL COLOR can include the option fifth number (fourth color value) for
> opacity; if not given, A=100, which is opaque
> (18) PPL COLOR 0, R, G, B [, A]  can be set to any color and will change the
> background color; same as SET WINDOW /COLOR=(R,G,B,A)
> (19) Change in behavior in PyFerret from Ferret (was in earlier PyFerret as
> well): except for color 0 (background color), changing the color values for
> a color number have no effect of past drawings in a window with that color
> number; only future drawings in the window with that color number will use
> the new color (i.e., PPL COLOR 1, 100, 0, 100 will not change black lines
> already plotted, but future drawings will use magenta instead of black.)  No
> need to protect colors.
> (20) PPL LABSET HLAB1, HXLAB, HYLAB values now used for automatic labels, so
> user can customize heights of automatic labels
> (21) Fixed issue with viewports not clearing.  Also works when using
> -nodisplay.
> (22) SET WINDOW /QUALITY= ignored if started with -nodisplay option.  So the
> same script can be used for a high-quality displayed session and a
> -nodisplay session.
> (23) removed deferral of graphics updates, which cause incomplete images to
> be saved or displayed
> (24) If ESMP (ESMF Python) package present, adds curv2rect and curv3srect
> Python external functions to standard set of external functions.
> (25) Added symbols PROGRAM_NAME ("Ferret" or "PyFerret") and PPL$SCALE
> (scaling factor for automatic lengths and text heights).
> (26) Includes latest fixes and enhancements in Ferret code (6.8.5 beta)
>
> Some known issues:
> (A) control-C handling problems: (1) if done while reading from a data set,
> can leave Ferret in strange state with regards to that data set; (2)
> sometimes slips through to the viewers and kills the viewers.
> (B) The white lines in shades and fills is seen again in PS and PDF.  My
> previous patch, which was removed, instead showed darker color lines in
> translucent fills and shades from overlap.  Can do a fill then shade, or
> shade then fill, to work around the problem.  The worst offender is "go
> fland", but multiple calls reduces or removes the problem.
> (C) Need to SET MODE LINECOLORS 7 (or larger) in order for the color 'white'
> (now color 7) to be interpreted correctly; otherwise gives purple.
> (D) update online documentation !!!
>
>
>
> On Fri, Jul 19, 2013 at 1:21 PM, Torben Mueller <torbenmllr@xxxxxxx> wrote:
>>
>> Hi,
>>
>> I wanted to ask if there are any updates on a successful Pyferret port to
>> OS X?
>>
>> Thanks!
>
>
>
>
> --
> Karl M. Smith, Ph.D.
> JISAO Univ. Wash. and TMAP/PMEL NOAA
> "The contents of this message are mine personally and do
> not necessarily reflect any position of the Government
> or the National Oceanic and Atmospheric Administration."