Writing python scripts to change fonts in FontForge

I assume you have a working knowledge of python (because I do not). Python documentation may be found on the web. FontForge implements two python modules -- one great huge one called fontforge which provides access to as much of FontForge's functionality as I've had time to write, and one tiny one called psMat which provides quick access to some useful transformations expressed as PostScript matrices.

In python terms fontforge embeds python. It is possible to build fontforge so that it is also a python extension (configure --enable-pyextension)

The fontforge module also defines several types

psMat
Methods
method args comments
identity () returns an identity matrix as a 6 element tuple
compose (mat1,mat2) returns a matrix which is the composition of the two input transformations
inverse (mat) returns a matrix which is the inverse of the input transformation. (Note: There will not always be an inverse)
rotate (theta) returns a matrix which will rotate by theta. Theta is expressed in radians
scale (x[,y]) returns a matrix which will scale by x in the horizontal direction and y in the vertical. If y is omitted, it will scale by the same amount (x) in both directions
skew (theta) returns a matrix which will skew by theta (to produce a oblique font). Theta is expressed in radians
translate (x,y) returns a matrix which will translate by x in the horizontal direction and y in the vertical
Types
None

fontforge
Global Variables
variable comments
hooks A dictionary which the user may fill to associate certain fontforge events with a python function to be run when those events happen. The function will be passed the font (or possibly glyph) for which the relevant event occurred.
Hook names
newFontHook This function will be called when a new font has been created.
loadFontHook This function will be called when a font is loaded from disk.
(if a font has an "initScriptString" entry in its persistent dictionary, that script will be invoked before this function).

Other hooks are defined in a font's own temporary and persistent dictionaries.

Methods
method args comments
getPrefs (pref-name) returns the value of the named preference item
setPrefs (pref-name,value) sets the value of the named preference item
hasSpiro () Returns a boolean, True if Raph Levien's spiro package is available for use in FontForge.
savePrefs () Saves the current preference settings
loadPrefs () Loads the user's default preference settings. Not done automatically in a script.
defaultOtherSubrs () Sets the type1 PostScript OtherSubrs to the default value
readOtherSubrsFile (filename) Sets the type1 PostScript OtherSubrs to the stuff found in the file.
loadEncodingFile (filename) Loads an encoding file, returns the name of the encoding or None.
loadNamelist (filename) Loads a namelist
loadNamelistDir (dirname) Loads all namelist files in the directory
loadPlugin (filename) Loads a fontforge plugin
loadPluginDir (dirname) Loads all fontforge plugins in the directory
preloadCidmap (filename,registry,order,supplement) Loads a fontforge cidmap file (first three args are strings, last is an integer)
printSetup (type[,printer|cmd|file,
width,height])
Prepare to print a font sample. The first argument may be one of:
lp
Queues postscript output to the printer using lp. You may use the optional second argument to specify the printer name.
lpr
Queues postscript output to the printer using lpr. You may use the optional second argument to specify the printer name.
ghostview
Displays the output using ghostview (or gv). The second argument is ignored.
command
Use a custom shell command to print the output. The second argument should contain the command and its arguments.
ps-file
Dump the postscript output to a file. The second argument specifies the filename.
pdf-file
Dump the output as pdf to a file. The second argument specifies the filename.

The third and fourth arguments are optional and specify the page size (in points) for the output. The third argument is the page width and the fourth is the height.

These setting remain until changed

nameFromUnicode (uni[,namelist]) Finds the glyph name associated with a given unicode codepoint. If a namelist is specified the name will be taken from that.
unicodeFromName (glyphname) Looks up glyph name in its dictionary and if it is associated with a unicode code point returns that number. Otherwise it returns -1
version () Returns fontforge's version number. This will be a large number like 20070406.
fonts () Returns a tuple of all fonts currently loaded into fontforge for editing
activeFont () If the script were invoked from the File->Execute Script... dialog, or invoked by a menu item in the font view, this returns the font that was active at the time. Otherwise it returns None.
activeGlyph () If the script were invoked from the File->Execute Script... dialog or a menu item from an outline glyph window or a glyph import/export command this returns the glyph that was active at the time. Otherwise it returns None.
activeLayer () This returns the currently active layer as an integer between 0 (inclusive) and the font/glyph's layer count (exclusive). It may also be set to -1 if the current glyph window is displaying the font's guidline layer.
fontsInFile (filename) Returns a tuple of all fontnames found in the specified file. The tuple may be empty if fontforge couldn't find any.
open (filename[,flags]) Opens a filename and returns the font it contains. If it does.

If the flags argument is 4, then ff will load all glyphs in the 'glyf' table of a ttc file (rather than just the glyphs used in the font picked). This will not load all 'glyf' tables though.

parseTTInstrs (string) Returns a binary string each byte of which corresponds to a truetype instruction. The input string should contain a set of instruction names as "SRP0\nMIRP[min,rnd,black]"
unParseTTInstrs (sequence) Reverse of the above. Converts a binary string into a human (sort of) readable string
unitShape (n) Returns a closed contour which is a regular n-gon. The contour will be inscribed in the unit circle. If n is negative, then the contour will be circumscribed around the unit circle. A value of 0 will produce a unit circle. If n==1 it is treated as if n were -4 -- a circumscribed square where each side is 2 units long (this is for historical reasons). Behavior is undefined for n=2,-1,-2.
registerGlyphSeparationHook (hook) The GlpyphSeparationHook is a python routine which FontForge will call when it wants to figure out the optical separation between two glyphs. If you neve call this, or if you call it with a value of None FontForge will use a built-in default. This routine gets called during AutoWidth, AutoKern, and computing the optical left and right side bearings (for 'lfbd' and 'rtbd' features). For more infomation see its own section.
User Interface Methods
Users may define scripts to be run when menu items are invoked. Some of these scripts will want to ask users questions, so this section provides routines to determine if fontforge has a user interface, a command to add menu items, and various small standard dialogs to interact with the user. I do not currently provide a mechanism for allowing people to define special purpose dialogs (for example they might want to ask more than one question in a dialog, and I don't support that).

When FontForge starts (if it's a fontforge with python) it will look at the directories $(PREFIX)/share/fontforge/python and ~/.FontForge/python and attempt to run all files in those directories which end in ".py". Presumably these files will allow people to customize the user interface to suit their needs.

Currently it reads the files in directory order (which is generally somewhere between creation order and totally random). It will read the system directory before the user directory.

Example

import fontforge;

def nameGlyph(junk,glyph):
  print glyph.glyphname;

fontforge.registerMenuItem(nameGlyph,None,None,"Glyph",None,"Print Glyph Name");

def neverEnableMe(junk,glyph):
  return False;

fontforge.registerMenuItem(nameGlyph,neverEnableMe,None,"Glyph",None,"SubMenu","Print Glyph Name");

def importGlyph(junk,glyph,filename,toback):
  print "Import";
  print glyph.glyphname;
  print filename;
def exportGlyph(junk,glyph,filename):
  print "Import";
  print glyph.glyphname;
  print filename;

fontforge.registerImportExport(importGlyph,exportGlyph,None,"foosball","foo","foo,foobar")

The first call will define a menu item in the Tools menu of the Glyph window. The menu will be called "Print Glyph Name". It has no shortcut to invoke it. It needs no external data. It is always enabled. And when activated it will invoke the function "nameGlyph" whch prints the name of the glyph in the window from which the command is invoked.

The second call defines a menu item in a submenu of the Tools menu. This submenu is called "SubMenu". This item will never be enabled -- but if it were enabled it would again call "nameGlyph" to print the name of the current glyph.

The last provides a way to import and export files of type "foosball" (or it would if the routines did anything).

Not a very useful example

hasUserInterface () Returns True if this session of FontForge has a user interface
registerMenuItem (menu-function,
enable_function,
data,
which_window,
shortcut_string,
{submenu-names,}
menu-name-string)
If fontforge has a user interface this will add this menu item to FontForge's Tool menu, either in the font or the outline glyph view (or both).
menu-function
This is the function that will be called when the menu item is activated. It will be passed two arguments, the first is the data value specified here (which may be None, indeed will probably usually be None), and the second is the glyph or font (depending on the window type) from which the menu item was activated. Its return value is ignored.
enable_function
This may be None -- in which case the menu item will always be enabled. Otherwise it will be called before the menu pops up on the screen to determine whether this item should be enabled. It will be passed the same arguments as above. It should return True if the item should be enabled and False otherwise.
data
This can be whatever you want (including None). FontForge keeps track of it and passes it to both of the above functions. Use it if you need to provide some context for the menu item.
which_window
May be either of the strings "Font" or "Glyph (or the tuple ("Font","Glyph")) and it determines which type of window will have this menu item in its "Tools" menu.
shortcut-string
May be None if you do not wish to supply a shortcut. Otherwise should be a string like "Menu Name|Cntl-H" (the syntax is defined in the translation section).
submenu-names
You may specify as many of these as you wish (including leaving them out altogether), this allows you to organize the Tools menu into submenus. (If a submenu of this name does not currently exist, fontforge will create it).
menu-name
The name that will appear in the menu for this item.

This will only affect windows created after this command is executed. Normally the command will be executed at startup and so it will affect all windows.

registerImportExport (import-function,
export_function,
data,
name,
extension,
[extension-list])
This will add the capability to import or export files of a given type, presumably a way of specifying the splines in a given glyph.
import-function
The function to call to import a file into a glyph. It will be called with: The data argument (specified below), A pointer to the glyph into which the import is to happen, A filename, A flag indicating whether the import should go to the background layer or foreground.

This function may be None. In which case there is no import method for this file type.

export-function
The function to call to export a glyph into a file. It will be called with: The data argument (see below), a pointer to the glyph, and a filename.

This function may be None, in which case there is no export method for this file type.

data
Anything you like (including None). It will be passed to the import/export routines and can provide them with context if they need that.
name
The name to be displayed in the user interface for this file type. This may just be the extension, or it might be something more informative.
extension
This is the default extension for this file type. It is used by the export dialog to pick an extension for the generated filename.
extension-list
Some file types have more than one common extension (eps files are usually named "eps", but I have also seen "ps" and "art" used). The import dialog needs to filter all possible filenames of this file type. This argument should be a comma separated list of extensions. It may be omitted, in which case it defaults to being the same as the "extension" argument above.
logWarning (msg) Adds the message (a string) to FontForge's Warnings window. (if you wish to display a % character you must represent it as two percents). If there is no user interface the output will go to stderr.
postError (win-title,msg) Creates a popup dialog to display the message (a string) in that dlg. (if you wish to display a % character you must represent it as two percents). If there is no user interface the output will go to stderr.
postNotice (win-title,msg) Creates a little window which will silently vanish after a minute or two and displays the message (a string) in that window. (if you wish to display a % character you must represent it as two percents). If there is no user interface the output will go to stderr.
openFilename (question,[def-name,[filter]]) All arguments are strings. The first is a question asked to the user (for which a filename to open is presumed to be the answer). The second is optional and provides a default filename. The third is optional and provides a filter (like "*.sfd")

The result is either a filename or None if the user canceled the dialog.

Throws an exception if there is no user interface.

saveFilename (question,[def-name,[filter]]) All arguments are strings. The first is a question asked to the user (for which a filename to save something to is presumed to be the answer). The second is optional and provides a default filename. The third is optional and provides a filter (like "*.sfd")

The result is either a filename or None if the user canceled the dialog.

Throws an exception if there is no user interface.

ask (title,question,answers,
[def,cancel])
Allows you to ask the user a multiple choice question. It popups up a dialog posing the question with a list of buttons ranged underneath it -- one for each answer.

The first argument is the dialog's title, the second is the question to be asked, the third is a tuple of strings -- each string will become a button, the fourth and fifth arguments are option, the fourth is the index in the answer array that will be the default answer (the one invoked if the user presses the [Return] key), and the fifth is the answer invoked if the user presses the [Escape] key. If omitted the default answer will be the first, and the cancel answer will be the last.

The function returns the index in the answer array of the answer choosen by the user.

Throws an exception if there is no user interface.

askChoices (title,question,answers,
[def])
Similar to the above allows you to ask the user a multiple choice question. It popups up a dialog posing the question with a scrollable list of choices -- one for each answer.

The first argument is the dialog's title, the second is the question to be asked, the third is a tuple of strings -- each string will become a button, the fourth and fifth arguments are option, the fourth is the index in the answer array that will be the default answer (the one invoked if the user presses the [Return] key). If omitted the default answer will be the first.

The function returns the index in the answer array of the answer choosen by the user. If the user cancels the dialog, a -1 is returned.

Throws an exception if there is no user interface.

askString (title,question,[def-string]) Allows you to as the user a question for which a string is the answer.

The first argument is the dialog's title, the second is the question to be asked, the third is optional and specified a default answer.

The function returns the string the user typed or None if they cancelled the dialog.

Throws an exception if there is no user interface.

Types
point
contour
layer
glyphPen
glyph
selection
font

point
Creation
fontforge.point ([x,y,on-curve]) Creates a new point. Optionally specifying its location
Members
member comments
x The x location of the point
y The y location of the point
on_curve Whether this is an on curve point or an off curve point (a control point)
selected Whether this point is selected in the UI. If an off-curve point is selected in means the preceding (interpolated) on-curve point is selected.
Methods
method args comments
dup () Returns a copy of the current point.
transform (tuple) Transforms the point by the transformation matrix
Pickling Methods
__reduce__ () This function allows the pickler to work on this type. I don't think it is useful for anything else.

contour
Description
A contour is a collection of points. A contour may be either based on cubic or quadratic splines.

If based on cubic splines there should be either 0 or 2 off-curve points between every two on-curve points. If there are no off-curve points then we have a line between those two points. If there are 2 off-curve points we have a cubic bezier curve between the two end points.

If based on quadratic splines things are more complex. Again, two adjacent on-curve points yield a line between those points. Two on-curve points with an off-curve point between them yields a quadratic bezier curve. However if there are two adjacent off-curve points then an on-curve point will be interpolated between them. (This should be familiar to anyone who has read the truetype 'glyf' table docs).

For examples of what these splines can look like see the section on bezier curves.

A contour may be open in which case it is just a long wiggly line, or closed when it is more like a circle with an inside and an outside. Unless you are making stroked fonts all your contours should eventually be closed.

Contours may also be expressed in terms of Raph Levien's spiro points. This is an alternate representation for the contour, and is not always available (Only if fontforge.hasSpiro() is True. If available the spiro member will return a tuple of spiro control points, while assigning to this member will change the shape of the contour to match the new spiros.

Two contours may be compared to see if they describe similar paths.

Creation
fontforge.contour () Creates a new contour
Members
member comments
is_quadratic Whether the contour should be interpretted as a set of quadratic or cubic splines. Setting this value has the side effect of converting the point list to the appropriate format
closed Whether the contour is open or closed
name The contour name (generally there is no name)
spiros This is an alternate representation of a curve. This member is only available if fontforge.hasSpiro() is True. Returns a tuple of spiro control points. Each of these is itself a tuple of four elements, an x,y location, a type field, and a set of flags. The type field takes on values (which are predefined constants in the fontforge module):
fontforge.spiroG4 1 A Spiro G4 curve point
fontforge.spiroG2 2 A Spiro G2 curve point
fontforge.spiroCorner 3 A Spiro corner point
fontforge.spiroLeft 4 A Spiro left "tangent" point
fontforge.spiroRight 5 A Spiro right "tangent" point
fontforge.spiroOpen 6 This may only be used on the first point in a spiro tuple. It indicates that the tuple describes an open contour.

For more information on what these point types mean see Raph Levien's work.

The flags argument is treated as a bitmap of which currently on one bit (0x1) is defined. This indicates that this point is selected in the UI.

When you assign a tuple of spiro control points to this member, the point list for the Bezier interpretation of the contour will change. And when you change the Bezier interpretation the set of spiro points will change.

Sequence Protocol
len(c) The number of points in the contour
c[i] The ith point on the contour. You may assign to this
c[i:j] The contour containing points between i and j. You may assign to this
c + d A contour concatenating c and d. D may be either another contour or a point.
c += d Appends d to c. D may be either another contour or a point.
p in c Returns whether the point p is in the contour c. p may be either a point or a tuple of two numbers
Does not support the repeat concept
Iterator Protocol
__iter__ () Returns an iterator for the contour which will return the points in order.
Methods
method args comments
dup () Returns a deep copy of the contour. That is, it copies the points that make up the contour.
isEmpty () Returns whether the contour is empty (contains no points)
  Contour construction
moveTo (x,y) Adds an initial, on-curve point at (x,y) to the contour
lineTo (x,y[,pos]) Adds an line to the contour. If the optional third argument is give, the line will be added after the pos'th point, otherwise it will be at the end of the contour.
cubicTo ((cp1x,cp1y)(cp2x,cp2y)(x,y)[,pos]) Adds a cubic curve to the contour. If the optional third argument is give, the line will be added after the pos'th point, otherwise it will be at the end of the contour.
quadraticTo ((cpx,cpy)(x,y)[,pos]) Adds a quadratic curve to the contour. If the optional third argument is give, the line will be added after the pos'th point, otherwise it will be at the end of the contour.
insertPoint (point[,pos]) Adds point to the contour. If the optional third argument is give, the line will be added after the pos'th point, otherwise it will be at the end of the contour. The point may be either a point or a tuple with three members (x,y,on_curve)
makeFirst (pos) Rotate the point list so that the pos'th point becomes the first point
isClockwise () Returns whether the contour is drawn in a clockwise direction. A return value of -1 indicates that no consistant direction could be found (the contour self-intersects).
reverseDirection () Reverse the order in which the contour is drawn (turns a clockwise contour into a counter-clockwise one). See also layer.correctDirection.
similar (other-contour[,error]) Checks whether this contour is similar to the other one where error is the maximum distance (in em-units) allowed for the two contours to diverge.

This is like the comparison operator, but that doesn't allow you to specify an error bound.

xBoundsAtY (ybottom[,ytop]) Finds the minimum and maximum x positions attained by the contour when y is between ybottom and ytop (if ytop is not specified it is assumed the same as ybottom). If the contour does not have any y values in the specified range then ff will return None.
yBoundsAtX (xleft[,xright]) Finds the minimum and maximum y positions attained by the contour when x is between xleft and xright (if xright is not specified it is assumed the same as xleft). If the contour does not have any x values in the specified range then ff will return None.
  Contour manipulation
addExtrema ([flags,emsize]) Extrema should be marked by on-curve points. If a curve lacks a point at an extrema this command will add one. Flags may be one of the following strings
all
Add all missing extrema
only_good
Only add extrema on longer splines (with respect to the em-size)
only_good_rm
As above but also merge away on-curve points which are very close to, but not on, an added extremum
cluster ([within,max]) Moves clustered coordinates to a standard central value. See Also round
merge (pos) Removes the on-curve point a the given position and rearranges the other points to make the curve as similar to the original as possible. (pos may also be a tuple of positions, all of which will be removed) See Also simplify
round ([factor]) Rounds the x and y coordinates. If factor is specified then new-coord = round(factor*old-coord)/factor. See Also cluster
selfIntersects () Returns whether this contour intersects itself.
simplify ([error-bound,flags,tan_bounds,
linefixup,linelenmax])
Tries to remove excess points on the contour if doing so will not perturb the curve by more than error-bound. Flags is a tuple of the following strings
ignoreslopes
Allow slopes to change
ignoreextrema
Allow removal of extrema
smoothcurves
Allow curve smoothing
choosehv
Snap to horizontal or vertical
forcelines
flatten bumps on lines
nearlyhvlines
Make nearly horizontal/vertical lines be so
mergelines
Merge adjacent lines into one
setstarttoextremum
Rotate the point list so that the start point is on an extremum
removesingletonpoints
If the contour contains just one point then remove it

See Also merge

transform (matrix) Transforms the contour by the matrix
random stuff about contours
boundingBox () Returns a tuple representing a rectangle (xmin,ymin, xmax,ymax) into which the contour fits. It is not guaranteed to be the smallest such rectangle, but it will often be.
getSplineAfterPoint (pos) Returns a tuple of two four-element tuples. These tuples are x and y splines for the curve after the specified point.
draw (pen) Draw the contour to the pen argument.
Pickling Methods
__reduce__ () This function allows the pickler to work on this type. I don't think it is useful for anything else.

layer
Description
A layer is a collection of contours. All the contours must be the same order (all quadratic or all cubic). Currently layers do not contain references.

Layers may be compared to see if their contours are similar.

Creation
fontforge.layer () Creates a new layer
Members
member comments
is_quadratic Whether the contours should be interpretted as a set of quadratic or cubic splines. Setting this value has the side effect of converting the contour list to the appropriate format
Sequence Protocol
len(c) The number of contours in the layer
c[i] The ith contour on the layer. You may assign to this
c + d A layer concatenating c and d. D may be either another layer or a contour.
c += d Appends d to c. D may be either another layer or a contour.
Does not support the repeat, slice and contains concepts
Iterator Protocol
__iter__ () Returns an iterator for the layer which will return the contours in order.
Methods
method args comments
dup () Returns a deep copy of the layer. That is, it will copy all the contours and all the points as well as copying the layer object itself.
isEmpty () Returns whether the layer is empty (contains no contour)
addExtrema ([flags,emsize]) Extrema should be marked by on-curve points. If a curve lacks a point at an extrema this command will add one. Flags may be one of the following strings
all
Add all missing extrema
only_good
Only add extrema on longer splines (with respect to the em-size)
only_good_rm
As above but also merge away on-curve points which are very close to, but not on, an added extremum
cluster ([within,max]) Moves clustered coordinates to a standard central value. See also round.
correctDirection () Orients all contours so that external ones are clockwise and internal counter-clockwise. See also contour.isClockwise.
export (filename) Exports the current layer (in outline format) to a file. The type of file is determined by the extension.
exclude (excluded-layer) Removes the excluded area from the current contours. See also removeOverlap and intersect.
intersect () Leaves only areas in the intersection of contours. See also removeOverlap and exclude.
removeOverlap () Removes overlapping areas. See also intersect and exclude.
interpolateNewLayer (other-layer,amount) Creates (and returns) a new layer which contains splines interpolated from the current layer and the first argument. If amount is 0 the result will look like the current layer, if 1 then like the first argument.
round ([factor]) Rounds the x and y coordinates. If factor is specified then new-coord = round(factor*old-coord)/factor. See also cluster.
selfIntersects () Returns whether any of the contours on this layer intersects any other contour (including itself).
similar (other-layer[,error]) Checks whether this layer is similar to the other one where error is the maximum distance (in em-units) allowed for any two corresponding contours in the layers to diverge.

This is like the comparison operator, but that doesn't allow you to specify an error bound.

simplify ([error-bound,flags,tan_bounds,
linefixup,linelenmax])
Tries to remove excess points on the layer if doing so will not perturb the curve by more than error-bound. Flags is a tuple of the following strings
ignoreslopes
Allow slopes to change
ignoreextrema
Allow removal of extrema
smoothcurves
Allow curve smoothing
choosehv
Snap to horizontal or vertical
forcelines
flatten bumps on lines
nearlyhvlines
Make nearly horizontal/vertical lines be so
mergelines
Merge adjacent lines into one
setstarttoextremum
Rotate the point list so that the start point is on an extremum
removesingletonpoints
If the contour contains just one point then remove it
stemControl (stem_width_scale, [hscale, stem_height_scale, vscale, xheight]) Allows you to scale counters and stems independently of each other. Stem_width_scale specifies by how much the widths of stems should be scaled (this should be a number around 1). If omitted hscale defaults to 1, otherwise it will indicate the horizontal scaling factor for the glyph as a whole. If omitted stem_height_scale defaults to stem_width_scale, otherwise it specifies the scaling for stem heights. If omitted vscale defaults to hscale, otherwise it specifies the vertical scale factor for the glyph as a whole.

Xheight is optional, if specified it will fix the points at that height so that they will be at the same level across glyphs.

stroke ("circular",width[,linecap,linejoin,flags])
("eliptical",width,minor-width,angle
 [,linecap,linejoin,flags])
("caligraphic",width,height,angle[,flags])
("polygon",contour[,flags])
Strokes the current line using one of the indicated pens. Line cap may be one of
  • butt
  • round
  • square

line join may be

  • miter
  • round
  • bevel

flags is a tuple containing some of the following strings

  • removeinternal
  • removeexternal
  • cleanup

If a polygonal pen is specified the contour must be a closed, convex polygon (no curved edges) with fewer than 100 vertices.

transform (matrix) Transforms the layer by the matrix
nltransform (xexpr,yexpr) xexpr and yexpr are strings specifying non-linear transformations that will be applied to all points in the layer (with xexpr being applied to x values, and yexpr to y values, of course). The syntax for the expressions is explained in the non-linear transform dialog.
boundingBox () Returns a tuple representing a rectangle (xmin,ymin, xmax,ymax) into which the layer fits. It is not guaranteed to be the smallest such rectangle, but it will often be.
xBoundsAtY (ybottom[,ytop]) Finds the minimum and maximum x positions attained by the contour when y is between ybottom and ytop (if ytop is not specified it is assumed the same as ybottom). If the layer does not have any y values in the specified range then ff will return None.
yBoundsAtX (xleft[,xright]) Finds the minimum and maximum y positions attained by the contour when x is between xleft and xright (if xright is not specified it is assumed the same as xleft). If the layer does not have any x values in the specified range then ff will return None.
draw (pen) Draw the layer to the pen argument.
Pickling Methods
__reduce__ () This function allows the pickler to work on this type. I don't think it is useful for anything else.

glyphPen
This implements the Pen Protocol to draw into a FontForge glyph. You create a glyphPen with the glyphPen attribute of a glyph. You then draw into it with the operators below.
import fontforge;
font = fontforge.open("Ambrosia.sfd");	#Open a font
pen = font["B"].glyphPen();		# Create a pen to draw into glyph "B"
pen.moveTo((100,100));			# draw a square
pen.lineTo((100,200));
pen.lineTo((200,200));
pen.lineTo((200,100));
pen.closePath();			# end the contour

font["A"].draw(pen);			# or you can copy from one glyph to another
					# by having a glyph draw itself into the pen
pen = None;				# Finalize the pen. This tells FontForge
					# that the drawing is done and causes
			                # it to refresh the display (if a UI is active).

This type may not be pickled.

Members
None
Methods
method args comments
moveTo ((x,y)) With one exception this call begins every contor and creates an on curve point at (x,y) as the start point of that contour. This should be the first call after a pen has been created and the call that follows a closePath, endPath.
lineTo ((x,y)) Draws a line from the last point to (x,y) and adds that to the contour.
curveTo ((cp1.x,cp1.y),(cp2.x,cp2.y),(x,y))
((cp.x,cp.y),(x,y))
This routine has slightly different arguments depending on the type of the font. When drawing into a cubic font (PostScript) use the first set of arguments (with two control points -- off curve points -- between each on curve point). When drawing into a quadratic font (TrueType) use the second format with one control point between adjacent on-curve points.

The standard appears to support super-bezier curves with more than two control points between on-curve points. FontForge does not. Nor does FontForge allow you to draw a quadratic spline into a cubic font, nor vice versa.

qCurveTo ([(cp.x,cp.y)]*,(x,y))
([(cp.x,cp.y)]*,None))
This routine may only be used in quadratic (TrueType) fonts and has two different formats. It is used to express the TrueType idiom where an on-curve point mid-way between its control points may be omitted, leading to a run of off-curve points (with implied but unspecified on-curve points between them).

The first format allows an arbetary number of off-curve points followed by one on-curve point.

It is possible to have a contour which consists solely of off-curve points. When this happens the contour is NOT started with a moveTo, instead the entire contour, all the off curve points, are listed in one call, and the argument list is terminated by a None to indicate there are no on-curve points.

closePath () Closes the contour (connects the last point to the first point to make a loop) and ends it.
endPath () Ends the contour without closing it. This is only relevant if you are stroking contours.
addComponent (glypn-name,transform) Adds a reference (a component) to the glyph. The PostScript transformation matrix is a 6 element tuple.

glyph
Description
The glyph type refers to a fontforge Glyph object. It has no independent life of its own, it always lives within a font. It has all the things you expect to be associated with a glyph: a glyph name, a unicode encoding, a drawing layer, GPOS/GSUB features...

This type may not be pickled.

This type may not be created directly -- all glyphs are bound to a font and must be created through the font.

Members
member comments
activeLayer Returns currently active layer in the glyph (as an integer). May be set to an integer or a layer name to change the active layer.
altuni Returns additional unicode code points for this glyph. For a primary code point, see unicode.

Returns either None or a tuple of alternate encodings. Each alternate encoding is a tuple of
(unicode-value, variation-selector, reserved-field)
The first is an unicode value of this alternate code point. The second is an integer for variation selector and can be set to -1 if not used. The third is an empty field reserved for future use and currently must be set to zero.

glyph.altuni can be set to None to clear all alternates, or to a tuple. The elements of the tuple may be either integers (an alternate unicode value with no variation selector) or a tuple with up to 3 values in it as explained above.

anchorPoints Returns the list of anchor points in the glyph. Each anchor point is a tuple of
(anchor-class-name, type, x,y [,ligature-index])
The first two are strings, the next two doubles, and the last (which is only present if type=="ligature") is an integer. Type may be
  • mark
  • base
  • ligature
  • basemark
  • entry
  • exit
anchorPointsWithSel Same as the above, except also includes whether the anchor point is selected in the UI. Returns a tuple of all anchor points in the glyph. Each anchor point is a tuple of
(anchor-class-name, type, x,y, selected [,ligature-index])
The first two are strings, the next two doubles, then a boolean, and the last (which is only present if type=="ligature") is an integer. Type may be
  • mark
  • base
  • ligature
  • basemark
  • entry
  • exit
background The glyph's background layer. This is a copy of the glyph's data. See also foreground, and layers.
changed Whether this glyph has been modified. This is (should be) maintained automatically, but you may set it if you wish.
color The color of the glyph in the fontview. A 6 hex-digit RGB number or -1 for default. 0xffffff is white, 0x0000ff is blue, etc.
comment Any comment you wish to associate with the glyph. UTF-8
dhints A tuple with one entry for each diagonal stem hint. Each stem hint is itself represented by a tuple of three coordinate pairs (themselves tuples of two numbers), these three are: a point on one side of the stem, a point on the other side, and a unit vector pointing in the stem's direction.
encoding Returns the glyph's encoding in the font's encoding. (readonly)

If the glyph has multiple encodings, one will be picked at random.
If the glyph is not in the font's encoding then a number will be returned beyond the encoding size (or in some cases -1 will be returned).

font The font containing this glyph. (readonly)
foreground The glyph's foreground layer. This is a copy of the glyph's data. See also background, layers and references.
glyphclass An opentype glyphclass, one of automatic, noclass, baseglyph, baseligature, mark, component
glyphname The name of the glyph
hhints A tuple of all horizontal postscript hints. Each hint is itself a tuple of starting locations and widths.
horizontalComponents A tuple of tuples.

This allows constructing very large versions of the glyph by stacking the componants together. Some components may be repeated so there is no bound on the size.

This is different from horizontalVariants which expects prebuilt glyphs of various fixed sizes.

The components are stacked in the order they appear in the (top-level) tuple. Each sub-tuple represents information on one component. The subtuple should contain: (String glyph-name, Boolean is-extender, Int startConnectorLength, Int endConnectorLength, Int fullAdvance). Any of these may be omitted (except the glyph name) and will be assumed to be 0 if so.

horizontalComponentItalicCorrection The italic correction for any composite glyph made with the horizontalComponents.
horizontalVariants A string containing a list of glyph names. These are alternate forms of the current glyph for use in typesetting math. Presumably the variants are of different sizes.

Although ff will always return a string of glyph names, you may assign to it with a tuple of glyphs and ff will convert that to corresponding names.

isExtendedShape A boolean containing the MATH "is extended shape" field.
italicCorrection The glyph's italic correction field. Used by both TeX and MATH. The special value of -32768 (0x8000) means the value is unspecified (An unspecified value will not go into the output tables, a value of 0 will)
layer_cnt The number of layers in this glyph. (Cannot be set)
layers A dictionary like object containing the layers of the glyph. It may be indexed by either a layer name, or an integer between 0 and glyph.layer_cnt-1 to produce a Layer object. Layer 0 is the background layer. Layer 1 is the foreground layer.
layerrefs A dictionary like object containing the references in the layers of the glyph. It may be indexed by either a layer name, or an integer between 0 and glyph.layer_cnt-1 to produce a reference tuple object. Layer 0 is the background layer. Layer 1 is the foreground layer.
lcarets A tuple containing the glyph's ligature caret locations. Setting this will also either enable or disable the "Default Ligature Caret Count" flag depending from the number of elements in the tuple.
left_side_bearing The left side bearing of the glyph
manualHints The glyph's hints have been set by hand, and the glyph should not be autohinted without a specific request from the user. The "Don't AutoHint" flag.
mathKern.bottomLeft The glyph's math kerning data associated with the bottom left vertex. This returns a tuple of two element tuples, each of which contains a kerning offset and an associated height (in the last entry the height term is meaningless, but present).
mathKern.bottomRight The glyph's math kerning data associated with the bottom right vertex. This returns a tuple of two element tuples, each of which contains a kerning offset and an associated height (in the last entry the height term is meaningless, but present).
mathKern.topLeft The glyph's math kerning data associated with the top left vertex. This returns a tuple of two element tuples, each of which contains a kerning offset and an associated height (in the last entry the height term is meaningless, but present).
mathKern.topRight The glyph's math kerning data associated with the top right vertex. This returns a tuple of two element tuples, each of which contains a kerning offset and an associated height (in the last entry the height term is meaningless, but present).
originalgid The GID of this glyph in the font it was read from. (readonly)
persistent Whatever you want (these data will be saved as a pickled object in the sfd file. It is your job to insure that whatever you put here can be pickled). See also the temporary field.
references A tuple of tuples containing glyph-name and a transformation matrix for each reference in the foreground. See also foreground and layerrefs.
right_side_bearing The right side bearing of the glyph
script A string containing the OpenType 4 letter tag for the script associated with this glyph (readonly)
temporary Whatever you want (these data will be lost once the font is closed) See also the persistent field.
texheight The Tex height. The special value of -32768 (0x8000) means the field is unspecified (An unspecified value will not go into the output tables, a value of 0 will)
texdepth The Tex depth. The special value of -32768 (0x8000) means the field is unspecified (An unspecified value will not go into the output tables, a value of 0 will)
topaccent The glyph's top accent position field. Used by MATH. The special value of -32768 (0x8000) means the field is unspecified (An unspecified value will not go into the output tables, a value of 0 will)
ttinstrs Any truetype instructions, returned as a binary string
unicode The glyph's unicode code point, or -1. In addition to this primary mapping, a glyph can have multiple secondary mappings - see altuni.
unlinkRmOvrlpSave A flag that indicates the glyph's references should be unlinked and remove overlap run on it before the font is saved (and then the original references replaced after the save finishes)
userdata Deprecated name for temporary field above
vhints A tuple of all vertical postscript hints. Each hint is itself a tuple of starting locations and widths.
validation_state A bit mask indicating some problems this glyph might have. (readonly)
0x1 If set then this glyph has been validated.
If unset then other bits are meaningless.
0x2 Glyph has an open contour.
0x4 Glyph intersects itself somewhere.
0x8 At least one contour is drawn in the wrong direction
0x10 At least one reference in the glyph has been flipped
(and so is drawn in the wrong direction)
0x20 Missing extrema
0x40 A glyph name referred to from this glyph, in an opentype table, is not present in the font.
0x40000 Points (or control points) are too far apart. (Coordinates must be within 32767)
PostScript only
0x80 PostScript has a limit of 1500 points in a glyph.
0x100 PostScript has a limit of 96 hints in a glyph.
0x200 Invalid glyph name.
TrueType only, errors in original file
0x400 More points in a glyph than allowed in 'maxp'
0x800 More paths in a glyph than allowed in 'maxp'
0x1000 More points in a composite glyph than allowed in 'maxp'
0x2000 More paths in a composite glyph than allowed in 'maxp'
0x4000 Instructions longer than allowed in 'maxp'
0x8000 More references in a glyph than allowed in 'maxp'
0x10000 References nested more deeply than allowed in 'maxp'
0x40000 Points too far apart. TrueType and Type2 fonts are limited to 16 bit numbers, and so adjacent points must be within 32767 em-units of each other.
0x80000 Points non-integral. TrueType points and control points must be integer aligned. (FontForge will round them if they aren't)
0x100000 Missing anchor. According to the opentype spec, if a glyph contains an anchor point for one anchor class in a subtable, it must contain anchor points for all anchor classes in the subtable. Even it, logically, they do not apply and are unnecessary.
0x200000 Duplicate glyph name. Two (or more) glyphs in this font have the same name. When outputting a PostScript font only one of them will ever be seen.

It's a little hard to detect this in normal use, but if you change the encoding to "Glyph Order", and then use Edit->Select->Wildcard and enter the glyph name, both of them should be selected.

0x400000 Duplicate unicode code point. Two (or more) glyphs in this font have the code point. When outputting an sfnt (TrueType/OpenType) font only one of them will ever be seen.

It's a little hard to detect this in normal use, but if you change the encoding to "Glyph Order", and then use Edit->Select->Wildcard and enter the code point, both of them should be selected.

0x800000 Overlapped hints. Either the glyph has no hint masks and there are overlapped hints, or a hint mask specifies two overlapping hints.
verticalComponents A tuple of tuples.

This allows constructing very large versions of the glyph by stacking the componants together. Some components may be repeated so there is no bound on the size.

This is different from verticalVariants which expects prebuilt glyphs of various fixed sizes.

The components are stacked in the order they appear in the (top-level) tuple. Each sub-tuple represents information on one component. The subtuple should contain: (String glyph-name, Boolean is-extender, Int startConnectorLength, Int endConnectorLength, Int fullAdvance). Any of these may be omitted (except the glyph name) and will be assumed to be 0 if so.

verticalComponentItalicCorrection The italic correction for any composite glyph made with the verticalComponents.
verticalVariants A string containing a list of glyph names. These are alternate forms of the current glyph for use in typesetting math. Presumably the variants are of different sizes.
width The advance width of the glyph. See also vwidth.
vwidth The vertical advance width of the glyph. See also width.
Methods
method args comments
addAnchorPoint (anchor-class-name,
anchor-type,
x,y
[,ligature-index])
Adds an anchor point. anchor-type may be one of the strings
  • "mark"
  • "base"
  • "ligature"
  • "basemark"
  • "entry"
  • "exit"

If there is an anchor point with the same anchor-class-name and:

  • lookup type is "gpos_mark2base" or
  • lookup type is "gpos_mark2ligature" and ligature-index is the same or
  • anchor-type is the same

then the existing anchor will be overwritten.

addExtrema ([flags,emsize]) Extrema should be marked by on-curve points. If a curve lacks a point at an extrema this command will add one. Flags may be one of the following strings
all
Add all missing extrema
only_good
Only add extrema on longer splines (with respect to the em-size)
only_good_rm
As above but also merge away on-curve points which are very close to, but not on, an added extremum
addReference (glyph-name[,transform]) Adds a reference to the specified glyph into the current glyph. Optionally specifying a transformation matrix
addHint (is-vertical,start,width) Adds a postscript hint. Takes a boolean flag indicating whether the hint is horizontal or vertical, a start location and the hint's width.
addPosSub (subtable-name,variant)
(subtable-name,variants)
(subtable-name,ligature-components)
(subtable-name,xoff,yoff,xadv,yadv)
(subtable-name,other-glyph-name,kerning)
(subtable-name,other-glyph-name,  xoff1,yoff1,xadv1,yadv1,
 xoff2,yoff2,xadv2,yadv2)
Adds position/substitution data to the glyph. The number and type of the arguments vary acording to the type of the lookup containing the subtable. The first argument should always be a lookup subtable name. If the lookup is for single substitutions then the second argument should be a string containing a single glyph name. For multiple and alternated substitutions a tuple of glyph names. For ligatures, a tuple of the ligature components (glyph names). For single positionings the second through fifth arguments should be small integers representing the adjustment along the appropriate axis. For pairwise positionings (kerning) the second argument should be the name of the other glyph being kerned with, and the third through tenth should be small integers -- or, if there are exactly three arguments then the third specifies traditional, one-axis, kerning

If there is a previously existing entry, this will replace it (except for ligatures).

appendAccent (name="glyph-name")
(unicode=<codepoint>)
Makes a reference to the specified glyph, adds that reference to the current layer of this glyph, and positions it to make a reasonable accent.
autoHint () Generates PostScript hints for this glyph.
autoInstr () Generates TrueType instructions for this glyph.
autoTrace () Auto traces any background images
boundingBox () Returns a tuple representing a rectangle (xmin,ymin, xmax,ymax) which is the minimum bounding box of the glyph.
build () If the character is a composite character, then clears it and inserts references to its components
canonicalContours () Orders the contours in the current glyph by the x coordinate of their leftmost point. (This can reduce the size of the charstring needed to describe the glyph(s).
canonicalStart () Sets the start point of all the contours of the current glyph to be the leftmost point on the contour. (If there are several points with that value then use the one which is closest to the baseline). This can reduce the size of the charstring needed to describe the glyph(s). By regularizing things it can also make more things available to be put in subroutines.
changeWeight (stroke_width[,type,
serif_height,serif_fuzz,
counter_type,custom_zones])
See the Element->Style->Change Width command for a more complete description of these arguments.

Stroke_width is the amount by which all stems are expanded.

Type is one of "LCG", "CJK", "auto", "custom".

Serif_height tells ff not to expand serifs which are that much off the baseline, while serif_fuzz specifies the amount of fuzziness allowed in the match. If you don't want special serif behavior set this to 0.

Counter_type is one of "squish", "retain", "auto".

Custom_zones is only meaningful if the type argument were "custom". It may be either a number, which specifies the "top hint" value (bottom hint is assumed to be 0, others are between), or a tuple of 4 numbers (top hint, top zone, bottom zone, bottom hint).

condenseExtend (c_factor,c_add[,sb_factor,sb_add,correct]) Condenses or extends the size of the counters and side-bearings of the glyph. The first two arguments provide information on shrinking/growing the counters, the second two the sidebearings. If the last two are omitted they default to the same values as the first two.

A counter's width will become:
   new_width = c_factor * old_width + c_add

If present the correct argument allows you to specify whether you want to correct for the italic angle before condensing the glyph. (it defaults to True)

clear () Clears the contents of the glyph (and marks it as not worth outputting).
cluster ([within,max]) Moves clustered coordinates to a standard central value. See also round.
correctDirection () Orients all contours so that external ones are clockwise and internal counter-clockwise.
exclude (excluded-layer) Removes the excluded area from the current glyph. Takes an argument which is a layer. See also removeOverlap and intersect.
export
(filename[,pixelsize,bitdepth]) bitmap images
(filename[,layer]) vector outlines
Uses the file's extension to determine output file type. Exports outline formats to the file. For bitmap formats it will rasterize the glyph and output that. There are different optional arguments for rasterizing images and for direct outline output. bitdepth must be 1 or 8.
getPosSub (lookup-subtable-name) Returns any positioning/substitution data attached to the glyph controlled by the lookup-subtable. If the name is "*" then returns data from all subtables.

The data are returned as a tuple of tuples. The first element of the subtuples is the name of the lookup-subtable. The second element will be one of the strings: "Position", "Pair", "Substitution", "AltSubs", "MultSubs","Ligature".

Positioning data will be followed by four small integers representing adjustments to the: x position of the glyph, the y position, the horizontal advance, and the vertical advance.

Pair data will be followed by the name of the other glyph in the pair and then eight small integers representing adjustments to the: x position of the first glyph, the y position, the horizontal advance, and the vertical advance, and then a similar foursome for the second glyph.

Substitution data will be followed by a string containing the name of the glyph to replace the current one.

Multiple and Alternate data will be followed by several strings each containing the name of a replacement glyph.

Ligature data will be followed by several strings each containing the name of a ligature component glyph.

importOutlines (filename,[flags]) Uses the file's extension to determine behavior. Imports outline descriptions (eps, svg, glif files) into the forground layer. Imports image descriptions (bmp, png, xbm, etc.) into the background layer. Optionally, flags can be used to control PostScript import, it'll be ignored for other file types. Flags is a tuple of the following strings
toobigwarn
Supress warning window about too big stroke width
removeoverlap
When FontForge detects that an expanded stroke will self-intersect, then setting this flag will cause it to try to make things nice by removing the intersections
handle_eraser
Certain programs use pens with white ink as erasers. When this flag is set, FontForge will attempt to simulate that.
correctdir
intersect () Leaves only areas in the intersection of contours. See also removeOverlap and exclude.
isWorthOutputting () Returns whether the glyph is worth outputting into a font file. Basically a glyph is worth outputting if it contains any contours, or references or has had its width set.
preserveLayerAsUndo ([layer,dohints]) Normally undo handling is turned off during python scripting. If you wish you may tell fontforge to preserve the current state of a layer so that whatever you do later can be undone by the user. You may omit the layer parameter (in which case the currently active layer will be used). You may also request that hints be preserved (they are not, by default).
removeOverlap () Removes overlapping areas. See also intersect and exclude.
removePosSub (lookup-subtable-name) Removes all data from the glyph corresponding to the given lookup-subtable. If the name is "*" then all data will be removed.
round ([factor]) Rounds the x and y coordinates of each point in the glyph. If factor is specified then new-coord = round(factor*old-coord)/factor. See also cluster.
selfIntersects () Returns whether any of the contours in this glyph intersects any other contour in the glyph (including itself).
simplify ([error-bound,flags,tan_bounds,
linefixup,linelenmax])
Tries to remove excess points in the glyph if doing so will not perturb the curve by more than error-bound. Flags is a tuple of the following strings
ignoreslopes
Allow slopes to change
ignoreextrema
Allow removal of extrema
smoothcurves
Allow curve smoothing
choosehv
Snap to horizontal or vertical
forcelines
flatten bumps on lines
nearlyhvlines
Make nearly horizontal/vertical lines be so
mergelines
Merge adjacent lines into one
setstarttoextremum
Rotate the point list so that the start point is on an extremum
removesingletonpoints
If the contour contains just one point then remove it
stroke ("circular",width[,linecap,linejoin,flags])
("eliptical",width,minor-width,angle
 [,linecap,linejoin,flags])
("caligraphic",width,height,angle[,flags])
("polygonal",contour[,flags])
Strokes the contours of the glyph using one of the indicated pens. Line cap may be one of
  • butt
  • round
  • square

line join may be

  • miter
  • round
  • bevel

flags is a tuple containing some of the following strings

  • removeinternal
  • removeexternal
  • cleanup

If a polygonal pen is specified, the contour must be a closed convex polygon (no curved edges) with fewer than 100 vertices.

transform (matrix[,flags]) Transforms the glyph by the matrix. The optional flags argument should be a tuple containing any of the following strings:
  • partialRefs -- Don't transform any references in the glyph, but do transform their offsets. This is useful if the refered glyph will be (or has been) transformed.
  • round -- Round to int after the transformation is done.
nltransform (xexpr,yexpr) xexpr and yexpr are strings specifying non-linear transformations that will be applied to all points in the current layer (with xexpr being applied to x values, and yexpr to y values, of course). The syntax for the expressions is explained in the non-linear transform dialog.
unlinkRef ([ref-name]) Unlinks the reference to the glyph named ref-name. If ref-name is omitted, unlinks all references.
unlinkThisGlyph () Unlinks all the references to the current glyph within any other glyph in the font.
useRefsMetrics (ref-name[,flag]) Finds a reference with the given name and sets the "use_my_metrics" flag on it (so this glyph will have the same advance width as the glyph the reference points to).

If the optional flag argument is False, then the glyph will no longer have its metrics bound to the reference.

validate ([force]) Validates the glyph and returns the validation_state of the glyph (except bit 0x1 will always be clear). If the glyph passed the validation then the return value will be 0 (not 0x1). Otherwise the return value will be the set of errors found. If force is specified true this will always be validated, if force is unspecified (or specified as false) then it will return the cached value if it is known, otherwise will validate it.
draw (pen) Draw the glyph's outline to the pen argument.
glyphPen ([replace=False]) Creates a new glyphPen which will draw into the current glyph. By default the pen will replace any existing contours and references, but setting the optional keyword argument, replace to false will retain the old contents.
Note: Glyphs do not have an independent existence. They live in fonts. You may not create them with stand-alone, only in the context of a font. See font.createChar

selection
This represents a font's selection. You may index it with an encoding value (in the encoding ISO-646-US (ASCII) the character "A" has encoding index 65), or with a glyph's name, or with a string like "uXXXXX" where XXXXX represent the glyph's unicode codepoint in hex, or with a fontforge glyph object. The value of indexing into a selection will be either True or False.
>>> print fontforge.activeFont().selection[65]
True

This type may not be pickled.

Members
member comments
byGlyphs Returns another selection, just the same as this one except that its iterator function will return glyphs (rather than encoding slots) and will only return those entries for which glyphs exist. (This is read only)
Iterator Protocol
__iter__ () Returns an iterator for the selection which will return all selected encoding slots in encoding order.
Methods
method args comments
all () Select everything.
none () Deselect everything.
changed () Select all glyphs which have changed.
invert () Invert the selection.
select (args) There may be an arbetrary number of arguments. Each argument may be either:
  • A glyph name
    Note: There need not be a glyph with this name in the font yet, but if you use a standard name (like "A") fontforge will still know where that glyph should be.
  • An integer (this will be interpreted as either an encoding index or (default) a unicode code point depending on the flags).
  • A fontforge glyph.
  • A tuple of flags.
    (If you wish to specify a single flag it must still be in a tuple, and you must append a trailing comma to the flag (so ("more",) rather than just ("more")). FF needs the flags to be in a tuple otherwise it can't distinguish them from glyph names)
    unicode
    Interpret integer arguments as unicode code points
    encoding
    Interpret integer arguments as encoding indeces.
    more
    Specified items should be selected
    less
    Specified items should be deselected.
    singletons
    Specified items should be interpreted individually and mean the obvious.
    ranges
    Specified items should be interpreted in pairs and represent all encoding slots between the start and end points specified by the pair. So .select(("ranges",None),"A","Z") would select all the upper case (latin) letters.

If the first argument is not a flag argument (or if it doesn't specify either "more" or "less") then the selection will be cleared. So .select("A") would produce a selection with only "A" selected, .select(("more",None),"A") would add "A" to the current selection, while .select(("less",None),"A") would remove "A" from the current selection.

private
This represents a font's postscript private dictionary. You may index it with one of the standard names of things that live in the private dictionary.

This type may not be pickled.

Iterator Protocol
__iter__ () Returns an iterator for the dictionary which will return all entres.
Methods
method args comments
guess (name) Guess a value for this this entry in the private dictionary. If fontforge can't make a guess it will simply ignore the request.

math
This represents a font's math constant table. Not all fonts have math tables, and checking this field will not create the underlying object, but examining or assigning to its members will create it..

This type may not be pickled.

Members
Any of the math constant names may be used as member names.
The list is long, and I shall not copy them all. Here is a subset
ScriptPercentScaleDown
ScriptScriptPercentScaleDown
DelimitedSubFormulaMinHeight
...

(These names begin with capital letters, not python's conventions but MicroSoft's)
These all take (16 bit) integer values.
I do not currently provide python access to any associated device tables.
Methods
method args comments
exists () Returns whether the font currently has an underlying math table associated with it. Note that examining or assigning to one of the members will create such a table.
clear () Removes any underlying math table from the font.

font
Description
The font type refers to a fontforge Font object. It generally contains a list of glyphs, an encoding to order those glyphs, a fontname, a list of GPOS/GSUB lookups and many other things.

This type may not be pickled.

Creation
fontforge.font () Creates a new font
Members
member comments
activeLayer Returns currently active layer in the font (as an integer). May be set to an integer or a layer name to change the active layer.
ascent The font's ascent
bitmapSizes A tuple with an entry for each bitmap strike attached to the font. Each strike is identified by pixelsize (if the strike is a grey scale font it will be indicated by (bitmap-depth<<16)|pixelsize.

When setting this value pass in a tuple of the same format. Any existing strike not specified in the tuple will be removed. Any new sizes will be created (but not rasterized -- use regenBitmaps for that).

capHeight (readonly) Computes the Cap Height (the height of capital letters such as "E"). A negative number indicates the value could not be computed (the font might have no capital letters because it was lower case only, or didn't include glyphs for a script with capital letters).
changed Bit indicating whether the font has been modified. This is (should be) maintained automatically, but you may set it if you wish.
cidcopyright Copyright message of the cid-keyed font as a whole (ie. not the current subfont).
cidfamilyname Family name of the cid-keyed font as a whole (ie. not the current subfont).
cidfontname Font name of the cid-keyed font as a whole (ie. not the current subfont).
cidfullname Full name of the cid-keyed font as a whole (ie. not the current subfont).
cidordering  
cidregistry  
cidsubfont Returns the number index of the current subfont in the cid-keyed font (or -1 if this is not a cid-keyed font).

May be set to an index (an integer) or a subfont fontname (a string) to change the current subfont. (To find the name of the current subfont, simply use .fontname).

cidsubfontcnt Returns the number of subfonts in this cid-keyed font (or 0 if it is not a cid-keyed font)
cidsubfontnames Returns a tuple of the subfont names in this cid-keyed font (or None if it is not a cid-keyed font)
cidsupplement  
cidversion  
cidweight Weight of the cid-keyed font as a whole
comment A comment associated with the font. Can be anything.
copyright PostScript copyright notice
cvt Returns a sequence object containing the font's cvt table. Changes made to this object will be made to the font (this is a reference not a copy).
The object has one additional method cvt.find(value[,low,high]) which finds the index of value in the cvt table (or -1 if not found). If low and high are specified then the index will be between [low,high).
default_base_filename The default base for the filename when generating a font
descent The font's descent
design_size Size (in pica points) for which this font was designed.
em The em size of the font. Setting this will scale the entire font to the new size.
encoding The name of the current encoding. Setting it will change the encoding used for indexing
familyname PostScript font family name
fondname Mac fond name
fontlog A comment associated with the font. Can be anything.
fontname PostScript font name
Note that in a CID keyed font this will be the name of the current subfont. Use cidfontname for the name of the font as a whole.
fullname PostScript font name
gasp Returns a tuple of all gasp table entries. Each item in the tuple is itself a tuple composed of a ppem (an integer) and a tuple of flags. The flags are a chosen from:
  • gridfit
  • antialias
  • symmetric-smoothing
  • gridfit+smoothing
gasp_version The version of the 'gasp' table. Currently this may be 0 or 1.
gpos_lookups Returns a tuple of all positioning lookup names in the font. This member cannot be set.
gsub_lookups Returns a tuple of all substitution lookup names in the font. This member cannot be set.
guide A copy of the font's guide layer
hasvmetrics
head_optimized_for_cleartype
hhea_ascent
hhea_ascent_add
hhea_descent
hhea_descent_add
hhea_linegap
horizontalBaseline Returns a tuple of tuples containing the horizontal baseline information in the font (the 'BASE' table). If there is no information NONE will be returned, otherwise the format of the tuple is:
((tuple of baseline tags used), (tuple of script information))
The (tuple of baseline tags used) is simply a tuple of 4 letter strings as ("hang", "ideo", "romn") these are standard baseline tag names as defined in the opentype spec. The number of entries here, and their order is important as there will be subsequent tuples (in the script tuple) which use the same ordering.

The (tuple of script information) is again a tuple of script information tuples.

A script information tuple looks like
(script-tag,default-baseline-tag, (tuple of baseline positions), (tuple of language extents))
If there are no baseline tags defined (an empty tuple), then the default-baseline-tag and the (tuple of baseline positions) will be NONE. Otherwise both tags will be 4 character strings, and the (tuple of baseline positions) will be a tuple of numbers (in the same order as the (tuple of baseline tags used) above) specifying the relative positions of each baseline for this script.

A (tuple of language extents) is a tuple of language extent tuples.

A language extent tuple is
(language-tag,min-extent,max-extent, (tuple of feature extents))
language-tag is a 4 letter string specifying an opentype language,min/max-extent are numbers specifying how far above and below the baseline characters go in this script/language.

A (tuple of feature extents> is a tuple of feature extent tuples.

A feature extent tuple is
(feature-tag,min-extent,max-extent, (tuple of feature extents))
feature-tag is a 4 letter string specifying an opentype (GPOS or GSUB) feature tag,min/max-extent are numbers specifying how far above and below the baseline characters go in this script/language with the feature applied.

Example:

(("hang","ideo","romn"),
  (("cyrl","romn",(1405,-288,0),()),
   ("grek","romn",(1405,-288,0),()),
   ("latn","romn",(1405,-288,0),
    ("dflt",-576,1913,
     ("NoAc",-576,1482))
    ("ENG ",-576,1482,())
   )
  )
)
is_cid Indicates whether the font is a cid-keyed font or not. (Read-only)
is_quadratic Deprecated. Whether the contours should be interpretted as a set of quadratic or cubic splines. Setting this value has the side effect of converting the entire font into the other format

Now each layer may have its own setting for this value, which should be set on the font's layers object.

isnew A flag indicating that this is a new font
italicangle
macstyle
layer_cnt The number of layers in the font. (Read only. Can change using add and del operations on the layers array)
layers Returns a dictionary like object with information on the layers of the font -- a name and a boolean indicating whether the layer is quadratic or not.

You may remove a layer with del font.layers["unneeded layer"];
You may add a new layer with font.layers.add("layer-name",is_quadratic[, is_background]);
You may change a layer's name with font.layers["layer"].name = "new-name";
You may change the type of splines in a layer with font.layers["layer"].is_quadratic = True;
You may change whether it is a background layer by font.layers["layer"].is_background = True;

Note: The layers that live in the font are different from layers that live in a glyph. These objects do not have the Layer type documented earlier.

loadState A bitmask indicating non-fatal errors found when loading the font. (readonly)
0x01 Bad PostScript entry in 'name' table
0x02 Bad 'glyf' or 'loca' table
0x04 Bad 'CFF ' table
0x08 Bad 'hhea', 'hmtx', 'vhea' or 'vmtx' table
0x10 Bad 'cmap' table
0x20 Bad 'EBLC', 'bloc', 'EBDT' or 'bdat' (embedded bitmap) table
0x40 Bad Apple GX advanced typography table
0x80 Bad OpenType advanced typography table (GPOS, GSUB, GDEF, BASE)
0x100 Bad OS/2 version number
Windows will reject all fonts with a OS/2 version number of 0 and will reject OT-CFF fonts with a version number of 1
maxp_FDEFs The number of function definitions used by the tt program
maxp_IDEFs The number of instruction definitions used by the tt program
maxp_maxStackDepth The maximum stack depth used by the tt program
maxp_storageCnt The number of storage locations used by the tt program
maxp_twilightPtCnt The number of points in the twilight zone of the tt program
maxp_zones The number of zones used in the tt program
multilayer
onlybitmaps A flag indicating that this font only contains bitmaps. No outlines.
os2_codepages A 2 element tuple containing the OS/2 Codepages field
os2_family_class
os2_fstype
os2_panose
os2_strikeypos
os2_strikeysize
os2_subxoff
os2_subxsize
os2_subyoff
os2_subysize
os2_supxoff
os2_supxsize
os2_supyoff
os2_supysize
os2_typoascent
os2_typoascent_add
os2_typodescent
os2_typodescent_add
os2_typolinegap
os2_use_typo_metrics
os2_unicoderanges A 4 element tuple containing the OS/2 Unicode Ranges field
os2_vendor
os2_version
os2_weight
os2_weight_width_slope_only
os2_width
os2_winascent
os2_winascent_add
os2_windescent
os2_windescent_add
path (readonly) Returns a string containing the name of the file from which the font was originally read (in this session), or if this is a new font, returns a made up filename in the current directory named something like "Untitled1.sfd". See also sfd_path.
persistent Whatever you want -- though I recommend you store a dict here (these data will be saved as a pickled object in the sfd file. It is your job to insure that whatever you put here can be pickled)

If you do store a dict then the following entries will be treated specially:
initScriptString If present, and if this is a string, then each time the font is loaded from an sfd file, this string will be passed to the python interpretter. Note: This is a string, not a function. Function code cannot be pickled. Since it is a string it will receive no arguments, but the current font will be available in the activeFont method of the fontforge module.

This string will be interpretted before the loadFontHook of the module hooks dictionary.

One possible behavior for this string is to define function hooks to be stored in the temporary dict described below.

math Returns an object which provides information on the font's underlying math constant table. There is only one of these per font.
private Returns a dictionary like object representing the PostScript private dictionary for the font. Changing entries in this object will change them in the font. (It's a reference, not a copy).

There is an iterator associated with this entry.

privateState Checks the (PostScript) Private dictionary and returns a bitmask of some common errors.
0x000001 Odd number of elements in either the BlueValues or OtherBlues array.
0x000002 Elements in either the BlueValues or OtherBlues are disordered.
0x000004 Too many elements in either the BlueValues or OtherBlues array.
0x000008 Elements in either the BlueValues or OtherBlues array are too close (must be at least 2*BlueFuzz +1 appart.
0x000010 Elements in either the BlueValues or OtherBlues array are not integers.
0x000020 Alignment zone height in either the BlueValues or OtherBlues array is too big for the value of BlueScale.
0x000100 Odd number of elements in either the FamilyBlues or FamilyOtherBlues array.
0x000200 Elements in either the FamilyBlues or FamilyOtherBlues are disordered.
0x000400 Too many elements in either the FamilyBlues or FamilyOtherBlues array.
0x000800 Elements in either the FamilyBlues or FamilyOtherBlues array are too close (must be at least 2*BlueFuzz +1 appart.
0x001000 Elements in either the FamilyBlues or FamilyOtherBlues array are not integers.
0x002000 Alignment zone height in either the FamilyBlues or FamilyOtherBlues array is too big for the value of BlueScale.
0x010000 Missing BlueValues entry.
0x020000 Bad BlueFuzz entry.
0x040000 Bad BlueScale entry.
0x080000 Bad StdHW entry.
0x100000 Bad StdVW entry.
0x200000 Bad StemSnapH entry.
0x400000 Bad StemSnapV entry.
0x800000 StemSnapH does not include StdHW.
0x1000000 StemSnapV does not include StdVW.
0x2000000 Bad BlueShift entry.
selection Returns a reference to an array-like object representing the font's selection. There is one entry for each encoding slot (there may not be a glyph attached to every encoding slot). You may set this with a tuple of integers (or boolean values). There should not be more entries in the tuple than there are encoding slots in the current encoding. A True or non-0 value means the slot is selected.
sfd_path (readonly) Returns a string (or None) containing the name of the sfd file associated with this font. Sometimes this will be the same as path (above).
sfnt_names The strings in the sfnt 'name' table. A tuple of all ms names. Each name is itself a tuple of strings (language,strid,string). Language may be either the (english) name of the language/locale, or the number representing that language in MicroSoft's specification. Strid may be one of the (english) string names (Copyright, Family, SubFamily, etc.) or the numeric value of that item. The string itself is in UTF-8.

Mac names will be automagically created from ms names

sfntRevision The font revision field stored in the 'head' table of an sfnt. This is documented to be a fixed 16.16 number (that is a 32 bit number with the binary point between bits 15 and 16).

The field may be unset (in which case when the font is generated, FontForge will guess a default value from one of the version strings).

The value returned with be None if the field is unset or a double.

You may set it to None which "unsets" it, or to a double value, or to an integer. The integer will be treated as a 32 bit integer and right shifted by 16 to get a 16.16 value).

size_feature The OpenType 'size' feature has two formats. It may either represent the design size of the font (and nothing else) or the design size, and range (top and bottom point sizes for which this design works), a style id (used to represent this design size throughout the font family) and a set of language/string pairs used to represent this design size in the menu.

If no size information is specified in the font FontForge will return None.

If only the design size is specified, FontForge will return a tuple containing a single element: the point size for which the font was designed. (This is returned as a real number -- the field can represent tenths of a point).

Otherwise FontForge returns a tuple containing five elements, the design size, the bottom of the design range, the top, the style id and a tuple of tuples. Each sub-tuple is a language/string pair. Language may be either the (english) name of the language/locale, or The string itself is in UTF-8.

strokedfont is this a stroked font?
strokewidth the stroke width of a stroked font
temporary Whatever you want -- though I recommend you store a dict here (these data will be lost once the font is closed)

If you do store a dict then the following entries will be treated specially:
generateFontPreHook If present, and if this is a function it will be called just before a font is generated. It will be called with the font and the filename to which the font will be written.
generateFontPostHook If present, and if this is a function it will be called just after a font is generated. It will be called with the font and the filename to which the font will be written.

texparameters Returns a tuple of TeX font parameters. TeX font type followed by 22 parameters. Font type is one of:
  • text
  • mathsym
  • mathext
  • unset

In case of "unset" default values for font parameters will be returned.

uniqueid
upos underline position
userdata Deprecated name for temporary above
uwidth underline width
version PostScript font version string
verticalBaseline Same format as horizontal_baseline, which see.
vertical_origin deprecated
vhea_linegap
weight PostScript font weight string
woffMajor The major version number of a woff file, an integer.

The field may be unset (in which case when the font is generated, FontForge will guess a default value from one of the version strings).

The value returned with be None if the field is unset or an integer.

You may set it to None which "unsets" it, or to an integer.

woffMinor The minor version number of a woff file, an integer.

The field may be unset (in which case when the font is generated, FontForge will guess a default value from one of the version strings).

The value returned with be None if the field is unset or an integer.

You may set it to None which "unsets" it, or to an integer.

woffMetadata Any metadata associated with a woff file. This is a utf8 string containing unparsed xml.
xHeight (readonly) Computes the X Height (the height of lower case letters such as "x"). A negative number indicates the value could not be computed (the font might have no lower case letters because it was upper case only, or didn't include glyphs for a script with lower case letters).
Iterator Protocol
__iter__ () Returns an iterator for the font which will run through the font, in gid order, returning glyph names
<name> in f Returns whether the font contains a glyph with the given name.
Other iterators over the font: selection, font.find(), font.glyphs()
Mapping Protocol
len(f) The number of glyph slots in the current encoding
f[i] If i is an integer, then returns the glyph at that encoding. If a string then returns the glyph with that name. May not be assigned to.
Methods
method args comments
addAnchorClass (lookup-subtable-name,
 new-anchor-class-name)
Adds an anchor class to the specified (anchor) subtable.
addKerningClass (lookup-name,new-subtable-name,
 first-classes,
 second-classes,
 offsets
 [,after])

or

(lookup-name,new-subtable-name,
 separation,
 first-classes,
 second-classes
 [,onlyCloser,autokern,after])

or

(lookup-name,new-subtable-name,
 separation,class-distance,
 ,first-glyph-list,
 second-glyph-list,
 [,onlyCloser,autokern,after])

or

(lookup-name,new-subtable-name,
 separation,class-distance,
 [,onlyCloser,autokern,after])

Creates a new subtable and a new kerning class in the named lookup. The classes arguments are tuples of tuples of glyph names (each sub-tuble of glyph names is a kerning class). The offsets argument is a tuple of kerning offsets. There must be as many entries as len(first-class)*len(second-class). The optional after argument is used to specify the order of the subtable within the lookup.

The second format will cause FontForge to auto kern the subtable. The separation argument specifies the desired optical distance between any two glyphs (if this is specified as 0 then the kerning class will be designed so glyphs just touch each other). Again the user specifies two sets of predefined classes. If the (optional) onlyCloser flag is set true then only negative kerning values will be inserted into the table.

In the third format the user merely specifies two lists of glyphs to be used, fontforge will look for similarities among among the glyphs and guess at classes. The class-distance argument to determine how precise the classes should match (1 is very tight matching, 20 is rather loose).

In the last format the font's selection will be used to specify the list of glyphs to be examined (and the same list will be used for both the left and right glyphs -- but fontforge will probably find different classes).

addLookup (new-lookup-name,type,flags,
 feature-script-lang-tuple
 [,after-lookup-name)
Creates a new lookup with the given name, type and flags. It will tag it with any indicated features. The type of one of
  • gsub_single
  • gsub_multiple
  • gsub_alternate
  • gsub_ligature
  • gsub_context
  • gsub_contextchain
  • gsub_revesechain
  • morx_indic
  • morx_context
  • morx_insert
  • gpos_single
  • gpos_pair
  • gpos_cursive
  • gpos_mark2base
  • gpos_mark2ligature
  • gpos_mark2mark
  • gpos_context
  • gpos_contextchain
  • kern_statemachine

The flags argument is a tuple of strings. At most one of these strings may be the name of a mark class. The others are:

  • right_to_left
  • ignore_bases
  • ignore_ligatures
  • ignore_marks

A feature-script-lang tuple is a tuple with one entry for each feature (there may be no entries if there are no features). Each entry is itself a two element tuple, the first entry is a string containing a 4 letter feature tag, and the second entry is another tuple (potentially empty) with an entry for each script for which the feature is active. Each entry here is itself a two element tuple. The first element is a 4 letter script tag and the second is a tuple of languages. Each entry in the language tuple is a four letter language. Example: (("liga",(("latn",("dflt")),)),)
The optional final argument allows you to specify the ordering of the lookup. If not specified the lookup will be come the first lookup in its table.

addLookupSubtable (lookup-name,
 new-subtable-name
 [,after-subtable-name])
Creates a new subtable within the specified lookup. The lookup name should be a string specifying an existing lookup. The subtable name should also be a string and should not match any currently existing subtable in the lookup. The optional final argument allows you to specify the ordering within the lookup. If not specified this subtable will be first in the lookup.

If you want to create a subtable in a contextual lookup, then use addContextualSubtable below. If you want to create a kerning class subtable, then use addKerningClass above.

addContextualSubtable (lookup-name,
 new-subtable-name
 type
 rule
 [,afterSubtable=]  [,bclasses=]  [,mclasses=]  [,fclasses=]  [,bclassnames=]  [,mclassnames=]  [,fclassnames=] )
Creates a new subtable within the specified contextual lookup (contextual, contextual chaining, or reverse contextual chaining). The lookup name should be a string specifying an existing lookup. The subtable name should also be a string and should not match any currently existing subtable in the lookup.

The type should be one of the strings "glyph", "class", "coverage" or "reversecoverage". The rule should be a string specifying a string to match and a set of lookups to apply once the match has been made. (See below for more details).

The remaining arguments are optional, keyword arguments.

  • afterSubtable=, if present this should be followed by a string, the name of a subtable after which this one is to be placed in the lookup. If not specified this subtable will be first in the lookup.
  • bclasses=, fclasses=, mclasses= these three arguments specify sets of glyph classes for when type="class". They should be a tuple of thingies where each thingy is either a string containing a list of space separated glyph names, or another tuple containing a set of strings, each a glyph name. Note that the first class is magic and should usually be left as a null string.
  • bclassnames=, fclassnames=, mclassnames= These provide names for the glyph classes described above. These names are optional, but can be convenient. These are tuples of strings. There should be the same number of entries in bclassnames as there are in bclasses.
When type="glyph"
The rule should look something like:
  glyph-name1 glyph-name2 | glyph-name3 @<lookup-name> | glyph-name4
The "|"s divide between backtrack, match and lookahead sections. So this example would match it the current glyph were named glyph-name3 and it were preceded by glyph-name2 and that by glyph-name1 and followed by glyph-name4. If the match were successful then the lookup named lookup-name would be applied. The @<> are litteral characters and should be present in the rule.

If the invoked lookup is a ligature lookup then it should be invoked after the first glyph that forms the lookup (rather than the last) and all glyphs that might make up the lookup should be in the match section. So...
  e | f @ f l | o
would only apply the ff-lig lookup if the ffl were preceeded by e and followed by o.

When type="class"
The rule should look something like:
  class-name1 class-name2 | class-name3 @<lookup-name> | class-name4
Very similar to the case of glyphs, except that instead of glyph names we have class names here. It is possible to have different sets of class names in the three different sections (backtrack, match and lookahead). If you don't specify any class names then you must use numbers instead, each number refering to the class at that position in the tuple (the first class will be class 0, the second class 1, and so on).
When type="coverage"
The rule should look something like:
  [g1 g2] [g3 g4] | [g5 g6 g7] @<lookup-name> | [g8 g9]
Each entry within brackets, [], represents a coverage table and should be a list of glyph names. The brackets are specified literally.
When type="reversecoverage"
The rule should look something like:
  [g1 g2] [g3 g4] | [g5 g6 g7] => [rg1 rg2 rg3] | [g8 g9]
Very similar to normal coverage tables except that instead of specifying a lookup there are replacement glyphs inline. There must be the same number of replacement glyphs (rg1, rg2, rg3) as match glyphs (g5, g6, g7) and there may be only one coverage table in the match section.
WARNING This format has some limitations, if there are multiple lookups they will be applied in textual order (First lookup in the string is the first one applied). This limitation is also present in Adobe's feature files so I hope it shan't be a severe limitation.
addSmallCaps (scheight=,
 capheight=,
 lcstem=,
 ucstem=,
 symbols=,
 letter_extension=,
 symbol_extension=,
 stem_height_factor=)
This function uses keyword parameters. None is required, if omitted a default value will be used (generally found by analyzing the font).

For each selected letter, this function will create a corresponding small caps glyph. If you set the symbol keyword to true it will also create small caps variants of digits and symbols.

The outlines of the new glyph will be based on the outlines of the upper-case variant of the letter. These will then be scaled so that a glyph which was capheight high will now be scheight high, and so that stems which were ucstem wide will be lcstem wide. Normally the ratio of stem heights is the same as the ratio of stem widths, but you may change that with stem_height_factor.

When it creates a new glyph it will name that glyph by appending ".sc" to the original lower case letter name (so "a" would become "a.sc") you may change the extension used with letter_extension. Similary symbols and digits will use the extension "taboldstyle", but you may change that with symbol_extension.

When it creates a glyph it also creates two lookups one bound to the feature "c2sc" and the other to "smcp". A mapping from the lower case letter to the small caps letter will be provided under "smcp", while a mapping from the upper case to the small caps under "c2sc". Symbols will have both lookup maps attached to the original glyph.

alterKerningClass (subtable-name,
 first-classes,
 second-classes,
 offsets)
Changes the kerning class in the named subtable. The classes arguments are tuples of tuples of glyph names (each sub-tuble of glyph names is a kerning class). The offsets argument is a tuple of kerning offsets. There must be as many entries as len(first-class)*len(second-class). The optional after argument is used to specify the order of the subtable within the lookup.
autoKern (subtable-name,
 separation
 [,minKern=,
 onlyCloser=,
 touch=])

or

(subtable-name,
 separation,
 glyph-list1,
 glyph-list2
 [,minKern=,
 onlyCloser=,
 touch=])

The named subtable must be a kerning pair subtable that already exists.

This command will automatically generate kerning pairs for the named subtable. If no glyph lists are specified it will look at all pairs of the glyphs in the selection; if glyph lists are specified then it will look at all pairs that can be made with one glyph from the first list and the second from the second list.

It will attempt to guess a good kerning value between the two glyphs -- a value which will make the optical separation between the two appear to be separation em-units. If minkern is specified then and the (absolute value of the) kerning correction is less than this number then no kerning pair will be generated. If onlyCloser is set true then only negative kerning offsets will be generated (only thing which move two glyphs closer together). If touch is set to 1 then the kerning offset will not be based on optical distance but on the closest approach between two the two glyphs.

appendSFNTName (language,strid,string) Adds a new (or replaces an old) string in the sfnt 'name' table. Language may be either the english name of the language/locale as a string, or the number representing that language in MicroSoft's specification. Strid may be one of the (english) string names (Copyright, Family, SubFamily, etc.) or the numeric value of that item. The string itself is in UTF-8.
buildOrReplaceAALTFeatures () Removes any existing AALT features (and any lookups solely controled by such features) and creates new ones containing all possible single and alternate substutions available for each glyph.
cidConvertByCMap (cmap-filename) Converts a normal font into a CID-keyed font with one subfont using

the CMAP to determine the mapping.

cidConvertTo (registry,ordering,supplement) Converts a normal font into a CID-keyed font with one subfont.
cidFlatten () Converts a CID font into a normal font (glyphs will be in CID order).
cidFlattenByCMap (cmap-filename) Converts a CID font into a normal font using the encoding specified in the CMAP file.
cidInsertBlankSubFont () Adds a new (blank) sub-font into a cid-keyed font and changes the current sub-font to be it.
cidRemoveSubFont () Removes the current subfont from a cid-keyed font.
close () Frees memory for the current font.

Warning: Any python pointers to it will become invalid.

compareFonts (other-font,filename,
 flags-tuple)
This will compare the current font with the font in other-font (which must already have been opened). It will write the results to the filename, you may use "-" to send the output to stdout. The flags argument is a tuple of strings and controls what will be compared.
flags
outlines compare outlines
outlines-exactly compare outlines exactly (otherwise allow slight errors and the unlinking of references)
warn-outlines-mismatch warn if the outlines don't exactly match (but are pretty close)
hints compare hints
warn-refs-unlink warn if references need to be unlinked before a match is found
strikes compare bitmap strikes
fontnames compare font names
gpos compare glyph positioning
gsub compare glyph substitutions
add-outlines for any glyphs whose outlines differ, add the outlines of the glyph in the second font to the background of the glyph in the first
create-glyphs if a glyph exists in the second font but not the first, create that glyph in the first and add the outlines from the second into the backgroun layer
createChar (uni[,name]) Create (and return) a character at the specified unicode codepoint in this font and optionally name it. If you wish to create an glyph with no unicode codepoint set the first argument to -1 and specify a name. If there is already a character there, return it (it will not be renamed).
createInterpolatedGlyph (glyph1,glyph2,amount) Create (and return) a glyph with the same unicode code point as glyph1. The glyph may not already exist. The contents of the glyph will be formed by interpolating between glyph1 and glyph2. If amount==0 the result will look like glyph1, or 1 then like glyph2.
createMappedChar (enc)
(name)
Create (and return) a character at the specified encoding in this font. If there is already a character there, return it
find (contour[,error-bound,search_flags]) Searches the font for all glyphs containing the contour (or layer) and returns an iterator which returns those glyphs.

error-bound: defaults to 0.01.
search_flags: tuple of the strings: reverse, flips, rotate, scale.

When found, the glyph.temporary is set to a dict of:

	  {
	   "findMatchedRefs": matched_refs_bit_map,
	   "findMatchedContours": matched_contours_bit_map,
	   "findMatchedContoursStart": matched_contours_start_bit_map,
	  }

findEncodingSlot (uni)
(name)
Tests whether a glyph with this codepoint or name is in the font's encoding and returns the encoding slot. If the glyph is not present it returns -1.

(If a glyph with that name/unicode is in the font, but is not in the encoding, then an value beyond the end of the encoding will be returned).

glyphs ([type]) Returns an iterator which will return the glyphs in the font. By default they will be returned in "GID" order, but if type is specified as "encoding" they will be returned in encoding order. If returned in encoding order it is possible that a glyph will be returned more than once if there are multiple encoding slots which reference it.
generate (filename
 [,bitmap_type=,
 flags=,
 bitmap_resolution=,
 subfont_directory=,
 namelist=,
 layer=])
Generates a font. The type is determined by the font's extension. The bitmap type (if specified) is also an extension. If layer is specified, then the splines and references in that layer will be used instead of the foreground layer.

Flags is a tuple containing some of

afm
output an afm file
pfm
output a pfm file
tfm
output a tfm file
ofm
output a ofm file
composites-in-afm
Store composite info in the afm file
glyph-map-file
Output a glyph map file giving the mapping between output gid and glyphnames
short-post
Do not include glyphnames in a ttf/otf file
apple
output apple advanced typography tables
opentype
output opentype tables
old-kern
output an old style 'kern' with opentype tables
dummy-dsig
output an empty DSIG table so MS will mark a font with .ttf extension as an OpenType font.
TeX-table
Include a 'TeX ' table in an ttf/otf file
round
Round PS coordinates to integers
no-hints
Do not include PS hints
no-flex
Do not include PS flex hints
omit-instructions
Do not include TrueType instructions
PfEd-comments
Include font and glyph comments in the 'PfEd' table
PfEd-colors
Include glyph colors in the 'PfEd' table
PfEd-lookups
Include lookup names in the 'PfEd' table
PfEd-guidelines
Include guideline locations in the 'PfEd' table
PfEd-background
Include background (and spiro) layers in the 'PfEd' table
symbol
Generate an sfnt with a Symbol cmap entry rather than a Unicode entry.

See also save().

generateTtc (filename,others,
 [flags=,
 ttcflags=,  namelist=,
 layer=])
Generates a truetype collection file containing the current font and all others listed -- the others argument may be None, a font, or a tuple (or list) of fonts.

Flags are as above,

Ttcflags is a tuple consisting of the following

merge
Try and share tables and glyphs among the various fonts.
cff
Use the CFF glyph format rather than the TrueType format (the OpenType documentation says that this does not work, but both the Mac and unix/linux accept it).
generateFeatureFile (filename[,lookup-name]) Generates an adobe feature file for the current font. If a lookup-name is specified then only data for that lookup will be generated.
genericGlyphChange (stemType=<str>,
 thickThreshold=<double>,
 stemScale=<double>,
 stemAdd=<double>,
 stemHeightScale=<double>,
 stemHeightAdd=<double>,
 stemWidthScale=<double>,
 stemWidthAdd=<double>,
 thinStemScale=<double>,
 thinStemAdd=<double>,
 thickStemScale=<double>,
 thickStemAdd=<double>,
 processDiagonalStems=<boolean>,
 
 hCounterType=<str>,
 hCounterScale=<double>,
 hCounterAdd=<double>,
 lsbScale=<double>,
 lsbAdd=<double>,
 rsbScale=<double>,
 rsbAdd=<double>,
 
 vCounterType=<str>,
 vCounterScale=<double>,
 vCounterAdd=<double>,
 vScale=<double>,
 vMap=<tuple of tuples>)
This function uses keyword parameters. Which ones are required depends on the three type arguments (stemType, hCounterType, vCounterType).

If stemType is omitted, or is the string "uniform", then the stemScale parameter must be specified (and stemAdd may be). stemScale specifies a scaling factor by which all stems (horizontal and vertical, thick and thin) will be scaled. A value of 1.0 means no change. While stemAdd specifies the number of em-units to add to the width of each stem.

If stemType is the string "horizontalVertical", then values must be specified for stemHeightScale and stemWidthScale (and may be for stemHeightAdd, stemWidthAdd). The first of these specifies scaling for the height of horizontal stems, and the second scaling for the width of vertical stems.

If stemType is the string "thickThin", then values must be specified for thinStemScale, thickStemScale and thickThreshold (and may be for thinStemAdd, thickStemAdd). The first of these specifies scaling for the width/height of thin stems, and the second scaling for the width/height of thick stems. While the thickThreshold argument specifies the size (in em-units) at which a stem is classified as "thick".

If hCounterType is omitted, or is the string "uniform", then horizontal counters, and the left and right side bearings will all be scaled using the same rules, and hCounterScale must be specified to provide the scaling factor (while hCounterAdd may be specified).

If hCounterType is the string "nonUniform", then horizontal counters, and the left and right side bearings may all be scaled using different rules, and hCounterScale, lsbScale and rsbScale must be specified to provide the scaling factors (while hCounterAdd, lsbAdd, and rsbAdd may be specified).

If hCounterType is the string "center", then the left and right side-bearings will be set so the new glyph is centered within the original glyph's width. (Probably more useful for CJK fonts than LGC fonts).

If hCounterType is the string "retainScale", then the left and right side-bearings will be set so the new glyph is within within the original glyph's width, and the side-bearings remain in the same proportion to each other as before.

If vCounterType is omitted, or is the string "mapped", then certain zones on the glyph may be placed at new (or the same) locations -- similar to BlueValues. So you can specify a zone for the baseline, one for the x-height and another for the top of capitals and ascenders (and perhaps a fourth for descenders). Each such zone is specified by the vMap argument which is a tuple of 3-tuples, each 3-tuple specifying a zone with: Original location, original width, and final location. No default value is provided for this argument you must figure out all the values yourself.

If vCounterType is the string "scaled", then vertical counters, and the top and bottom side bearings will all be scaled using the same rules, and vCounterScale must be specified to provide the scaling factor (while vCounterAdd may be specified). This is probably most useful for CJK fonts.

getKerningClass (subtable-name) Returns a tuple whose entries are: (first-classes, second-classes, offsets). The classes are themselves tuples of tuples of glyph names. The offsets will be a tuple of numeric kerning offsets.
getLookupInfo (lookup-name) Returns a tuple whose entries are: (lookup-type, lookup-flags, feature-script-lang-tuple) The lookup type is a string as described in addLookup, and the feature-script-lang tuple is also described there.
getLookupSubtables (lookup-name) Returns a tuple of all subtable names in that lookup.
getLookupSubtableAnchorClasses (subtable-name) Returns a tuple of all anchor class names in that subtable.
getLookupOfSubtable (subtable-name) Returns the name of the lookup containing this subtable.
getSubtableOfAnchor (anchor-class-name) Returns the name of the subtable containing this anchor class.
importBitmaps (bitmap-font-file
[,to-background])
Load any bitmap strikes out of the bitmap-font-file into the current font
importLookups (another-font,lookup-names
[,before-name])
The first argument must be a FontForge Font object, the second a string or a tuple of strings, and the third, another string. It will search the other font for the named lookup(s) and import it into the current font. (Contextual lookups which invoke other lookups will have any nested lookups imported as well). Lookups will be imported in the order listed. If a before-name is specified, then it is looked up in the current font and all lookups will be added before it, if not specified lookups will appear at the end of the list.
interpolateFonts (fraction,filename) Interpolates a font between the current font and the font contained in filename.
isKerningClass (subtable-name) Returns whether the named subtable contains a kerning class.
isVerticalKerning (subtable-name) Returns whether the named subtable contains a vertical kerning data
italicize (italic_angle=, ia=
 lc_condense=, lc=
 uc_condense=, uc=
 symbol_condense=, symbol=
 deserif_flat=,
 deserif_slant=,
 deserif_pen=,
 baseline_serifs=,
 xheight_serifs=,
 ascent_serifs=,
 descent_serifs=,
 diagonal_serifs=,
 a=,
 f=,
 u0438=,
 u043f=,
 u0442=,
 u0444=,
 u0448=,
 u0452=,
 u045f=)
This function uses keyword parameters. None is required, if omitted a default value will be used. Some keywords have abbreviations ("ia" for "italic_angle") you may use either.

This function will attempt to italicize each selected glyph. For a detailed explanation of what this entails please see the information on the Italic dialog.

The *_condense keywords should be 4 element tuples of floating point numbers; these numbers correspond to: Left side bearing condensation, stem condensation, counter condensation and right side bearing condensation. These numbers should be small numbers around 1 (scale factors, not percentages).

Set at most one of the deserif_* keywords.

Setting a to true will turn on the transformation that applies to the "a" glyph. Setting u0438 will control the transformation that applies to the glyph at unicode codepoint U+0438.

The f keyword is slightly more complex. Setting it to 0 turns off all transformations of glyphs like "f", setting it to 1 will give "f" a tail which looks like a rotated version of its head, and setting it to 2 will simply extend the main stem of "f" below the baseline.

lookupSetFeatureList (lookup-name,
feature-script-lang-tuple)
Sets the feature list of indicated lookup. The feature-script-lang tuple is described at addLookup.
lookupSetFlags (lookup-name,flags) Sets the lookup flags for the named lookup.
lookupSetStoreLigatureInAfm (lookup-name,boolean) Sets whether this ligature lookup contains data to store in the afm.
mergeFonts (filename[,
preserveCrossFontKerning])
Merges the font in the file into the current font.
mergeFeature (filename) Merge feature and lookup information from an adobe feature file, or metrics information from the (afm,tfm,etc) file into the current font.
mergeKern (filename) Deprecated name for mergeFeature above
mergeLookups (lookup-name1,lookup-name2) The lookups must be of the same type. All subtables from lookup-name2 will be moved to lookup-name1, the features list of lookup-name2 will be merged with that of lookup-name1, and lookup-name2 will be removed.
mergeLookupSubtables (subtable-name1,
subtable-name2)
The subtables must be in the same lookup. Not all lookup types allow their subtables to be merged (contextual subtables may not be merged, kerning classes may not be (kerning pairs may be)). Any information bound to subtable2 will be bound to subtable1 and subtable2 will be removed.
printSample (type,pointsize,
 sample,output-file)
Type is a string which must be one of
fontdisplay
Display all glyphs in the font in encoding order
chars
Display the selected glyphs scaled to fill a page
Ignores the pointsize argument.
waterfall
Displays the selected glyphs at many pointsizes.
The pointsize argument should be a tuple of pointsizes here.
fontsample
The third argument should contain a string which will be layed out and displayed as well as FontForge can.
fontsampleinfile
The third argument should contain the name of a file which contains text to be layed out and displayed.

If output is to a file (see printSetup) then the last argument specifies a file name in which to store output.

randomText (script[,lang]) Returns a random text sample using the letter frequencies of the specified script (and optionally language). Both script and language should be expressed as strings containing OpenType Script and Language tags. "dflt" is a reasonable language tag. If the language is not specified, one will be chosen at random. If ff has no frequency information for the script/language specified it will use the letters in the script with equal frequencies.
regenBitmaps (tuple-of-sizes) A tuple with an entry for each bitmap strike to be regenerated (rerasterized). Each strike is identified by pixelsize (if the strike is a grey scale font it will be indicated by (bitmap-depth<<16)|pixelsize.
removeAnchorClass (anchor-class-name) Removes the named AnchorClass (and all associated points) from the font.
removeLookup (lookup-name) Remove the lookup (and any subtables within it).
removeLookupSubtable (subtable-name) Remove the subtable (and all data associated with it).
removeGlyph (uni[,name])
(name)
(glyph)
You may either pass in a FontForge glyph object (from this font) or identify a glyph in the font by unicode code point or name. In any case the glyph will be removed from the font.

If you use (uni,name) to specify a name, set uni to -1.

WARNING: This frees fontforge's storage to this glyph. If you have any python pointers to that storage they will be looking at garbage. This does not go through the usual python reference mechanism.

replaceAll (srch,rpl[,error-bound]) Searches the font for all occurences of the srch contour (or layer) and replaces them with the replace contour (or layer).
revert () Reloads the font from the disk.

Caveat: if you have any pointers to glyphs which live in the font those pointers will no longer be valid, and using them will cause crashes. This is very un-python-like.

revertFromBackup () Reloads the font from the backup file on the disk.

Caveat: if you have any pointers to glyphs which live in the font those pointers will no longer be valid, and using them will cause crashes. This is very un-python-like.

save ([filename]) Saves the font to an sfd file. See also generate()
saveNamelist (filename) Saves the font's namelist to a file.
getTableData (table-name) Gets binary data from any saved table. FF will save 'fpgm', 'prep', 'cvt ' and 'maxp'. FF may also save tables which you explicitly request. Do not expect to get binary data for tables like 'GPOS' or 'glyf' which FF will generate when it creates a font... that information is not currently available.

Returns a binary string.

setTableData (table-name,sequence) Sets binary data of any saved table. FF will save 'fpgm', 'prep', 'cvt ' and 'maxp'. FF may also save tables which you explicitly request. Do not expect to set binary data for tables like 'GPOS' or 'glyf' which FF will generate when it creates a font... that information is not currently available.

If sequence is None, then the named table will be removed from the font.

validate ([force]) Validates the font and returns a bit mask of all errors from all glyphs (as defined in the validation_state of a glyph -- except bit 0x1 is clear). If the font passed the validation then the return value will be 0 (not 0x1). Otherwise the return value will be the set of errors found.

Note: The set of errors is slightly different for TrueType and PostScript output. The returned mask contains the list of potential errors. You must figure out which apply to you.

Normally each glyph will cache its validation_state and it will not be recalculated. If you pass a non-zero argument to the routine then it will force recalculation of each glyph -- this can be slow.s

Selection based interface
See the selection type for how to alter the selection
addExtrema () Extrema should be marked by on-curve points. If a curve in any selected glyph lacks a point at a significant extremum this command will add one.
addSmallCaps () For all selected upper or lower case letters in the latin, greek and cyrillic scripts this will try to create a small caps version of that glyph in a new glyph slot. So if you select "A" (or "a") then a glyph "a.sc" will be created (if "a.sc" already exists, it will be reused, and its current contents cleared). The contents of "a.sc" will be based on the upper case variant of this glyph (and that variant must be present for the command to work). FontForge will also create two lookups (unless appropriate ones already exist) one, bound to the feature 'c2sc' will map upper case letters to small caps, the other, bound to feature 'smcp' will map lower case letters to small caps.
autoHint () Generates PostScript hints for all selected glyphs.
autoInstr () Generates TrueType instructions for all selected glyphs.
autoWidth (separation [,minBearing=,maxBearing=]) Guesses at reasonable horizontal advance widths for the selected glyphs
autoTrace () Auto traces any background images in all selected glyphs
build () If any of the selected characters is a composite character, then this command will clear it and insert references to its components (this command can create new glyphs).
canonicalContours () Orders the contours in the selected glyphs by the x coordinate of their leftmost point. (This can reduce the size of the charstring needed to describe the glyph(s).
canonicalStart () Sets the start point of all the contours of the selected glyphs to be the leftmost point on the contour. (If there are several points with that value then use the one which is closest to the baseline). This can reduce the size of the charstring needed to describe the glyph(s). By regularizing things it can also make more things available to be put in subroutines.
changeWeight (stroke_width[,type,
serif_height,serif_fuzz,
counter_type,custom_zones])
See the Element->Style->Change Width command for a more complete description of these arguments.

Stroke_width is the amount by which all stems are expanded.

Type is one of "LCG", "CJK", "auto", "custom".

Serif_height tells ff not to expand serifs which are that much off the baseline, while serif_fuzz specifies the amount of fuzziness allowed in the match. If you don't want special serif behavior set this to 0.

Counter_type is one of "squish", "retain", "auto".

Custom_zones is only meaningful if the type argument were "custom". It may be either a number, which specifies the "top hint" value (bottom hint is assumed to be 0, others are between), or a tuple of 4 numbers (top hint, top zone, bottom zone, bottom hint).

condenseExtend (c_factor,c_add
[,sb_factor,sb_add,
correct])
Condenses or extends the size of the counters and side-bearings of the selected glyphs. The first two arguments provide information on shrinking/growing the counters, the second two the sidebearings. If the last two are omitted they default to the same values as the first two.

A counter's width will become:
   new_width = c_factor * old_width + c_add

If present the correct argument allows you to specify whether you want to correct for the italic angle before condensing the glyph. (it defaults to True)

clear () Clears the contents of all selected glyphs
cluster ([within,max]) Moves clustered coordinates to a standard central value in all selected glyphs. See also round.
copy () Copies all selected glyphs into (fontforge's internal) clipboard.
copyReference () Copies all selected glyphs (as references) into (fontforge's internal) clipboard.
correctDirection () Orients all contours so that external ones are clockwise and internal counter-clockwise in all selected glyphs.
correctReferences () Checks a font for glyphs with mixed contours and references (or references with transformation matrices which cannot be represented truetype (ie. scaling by 2 or more)). If a mixed case is discovered fontforge will take the contours out of the glyph, put them in a new glyph, and make a reference to the new glyph.
cut () Copies all selected glyphs into (fontforge's internal) clipboard. And then clears them.
paste () Pastes the contents of (fontforge's internal) clipboard into the selected glyphs -- and removes what was there before.
intersect () Leaves only areas in the intersection of contours in all selected glyphs. See also removeOverlap.
pasteInto () Pastes the contents of (fontforge's internal) clipboard into the selected glyphs -- and retains what was there before.
removeOverlap () Removes overlapping areas in all selected glyphs. See also intersect.
replaceWithReference ([fudge]) Finds any glyph which contains an inline copy of one of the selected glyphs, and converts that copy into a reference to the appropriate glyph. Selection is changed to the set of glyphs which the command alters.

If specified the fudge argument specifies the error allowed for coordinate differences.

round ([factor]) Rounds the x and y coordinates of each point in all selected glyphs. If factor is specified then new-coord = round(factor*old-coord)/factor. See also cluster.
simplify ([error-bound,flags,tan_bounds,
linefixup,linelenmax])
Tries to remove excess points in all selected glyphs if doing so will not perturb the curve by more than error-bound. Flags is a tuple of the following strings
ignoreslopes
Allow slopes to change
ignoreextrema
Allow removal of extrema
smoothcurves
Allow curve smoothing
choosehv
Snap to horizontal or vertical
forcelines
flatten bumps on lines
nearlyhvlines
Make nearly horizontal/vertical lines be so
mergelines
Merge adjacent lines into one
setstarttoextremum
Rotate the point list so that the start point is on an extremum
removesingletonpoints
If the contour contains just one point then remove it
stroke ("circular",width[,
 linecap,linejoin,flags])
("eliptical",width,
 minor-width,angle
 [,linecap,linejoin,flags])
("caligraphic",width,
 height,angle[,flags])
("polygonal",contour[,flags])
Strokes the contours of all selected glyphs using one of the indicated pens. Line cap may be one of
  • butt
  • round
  • square

line join may be

  • miter
  • round
  • bevel

flags is a tuple containing some of the following strings

  • removeinternal
  • removeexternal
  • cleanup

If a polygonal pen is specified the contour must be a closed convex polygon (no curved edges) with fewer than 100 vertices.

transform (matrix) Transforms all selected glyphs by the matrix.
nltransform (xexpr,yexpr) xexpr and yexpr are strings specifying non-linear transformations that will be applied to all points in the selected glyphs of the font (with xexpr being applied to x values, and yexpr to y values, of course). The syntax for the expressions is explained in the non-linear transform dialog.
unlinkReferences () Unlinks all references in all selected glyphs and replaces them with splines.

Stupid example

import fontforge                                 #Load the module
amb=fontforge.open("Ambrosia.sfd")               #Open a font
amb.selection.select(("ranges",None),"A","Z")    #select A-Z
amb.copy()                                       #Copy those glyphs into the clipboard

n=fontforge.font()                               #Create a new font
n.selection.select(("ranges",None),"A","Z")      #select A-Z of it
n.paste()                                        #paste the glyphs above in
print n["A"].foreground                          #test to see that something
                                                 #  actually got pasted
n.fontname="NewFont"                             #Give the new font a name
n.save("NewFont.sfd")                            #and save it.

FontForge as a python extension

If you have configured fontforge with the --enable-pyextension argument, then when fontforge installs itself it will also set itself up as something that can be used inside of python (up until now we have been talking about using python inside of fontforge).

$ ./configure --enable-pyextension
$ make
$ sudo make install

once you do that you can invoke all of the above fontforge commands from inside of python by saying:

>>> import fontforge