# Orca # # Copyright 2010 Joanmarie Diggs. # # This library is free software; you can redistribute it and/or # modify it under the terms of the GNU Lesser General Public # License as published by the Free Software Foundation; either # version 2.1 of the License, or (at your option) any later version. # # This library is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU # Lesser General Public License for more details. # # You should have received a copy of the GNU Lesser General Public # License along with this library; if not, write to the # Free Software Foundation, Inc., Franklin Street, Fifth Floor, # Boston MA 02110-1301 USA. """Commonly-required utility methods needed by -- and potentially customized by -- application and toolkit scripts. They have been pulled out from the scripts because certain scripts had gotten way too large as a result of including these methods.""" __id__ = "$Id$" __version__ = "$Revision$" __date__ = "$Date$" __copyright__ = "Copyright (c) 2010 Joanmarie Diggs." __license__ = "LGPL" import math import pyatspi import re import debug import input_event import mouse_review import orca_state import settings from orca_i18n import _ # for gettext support from orca_i18n import ngettext # for ngettext support ############################################################################# # # # Utilities # # # ############################################################################# class Utilities: EMBEDDED_OBJECT_CHARACTER = u'\ufffc' SUPERSCRIPT_DIGITS = \ [u'\u2070', u'\u00b9', u'\u00b2', u'\u00b3', u'\u2074', u'\u2075', u'\u2076', u'\u2077', u'\u2078', u'\u2079'] SUBSCRIPT_DIGITS = \ [u'\u2080', u'\u2081', u'\u2082', u'\u2083', u'\u2084', u'\u2085', u'\u2086', u'\u2087', u'\u2088', u'\u2089'] flags = re.UNICODE WORDS_RE = re.compile("(\W+)", flags) SUPERSCRIPTS_RE = re.compile("[%s]+" % "".join(SUPERSCRIPT_DIGITS), flags) SUBSCRIPTS_RE = re.compile("[%s]+" % "".join(SUBSCRIPT_DIGITS), flags) # generatorCache # DISPLAYED_LABEL = 'displayedLabel' DISPLAYED_TEXT = 'displayedText' KEY_BINDING = 'keyBinding' NESTING_LEVEL = 'nestingLevel' NODE_LEVEL = 'nodeLevel' REAL_ACTIVE_DESCENDANT = 'realActiveDescendant' def __init__(self, script): """Creates an instance of the Utilities class. Arguments: - script: the script with which this instance is associated. """ self._script = script ######################################################################### # # # Utilities for finding, identifying, and comparing accessibles # # # ######################################################################### @staticmethod def __hasLabelForRelation(label): """Check if label has a LABEL_FOR relation Arguments: - label: the label in question Returns TRUE if label has a LABEL_FOR relation. """ if (not label) or (label.getRole() != pyatspi.ROLE_LABEL): return False relations = label.getRelationSet() for relation in relations: if relation.getRelationType() == pyatspi.RELATION_LABEL_FOR: return True return False @staticmethod def __isLabeling(label, obj): """Check if label is connected via LABEL_FOR relation with object Arguments: - obj: the object in question - labeled: the label in question Returns TRUE if label has a relation LABEL_FOR for object. """ if not obj or not label or label.getRole() != pyatspi.ROLE_LABEL: return False relations = label.getRelationSet() if not relations: return False for relation in relations: if relation.getRelationType() == pyatspi.RELATION_LABEL_FOR: for i in range(0, relation.getNTargets()): target = relation.getTarget(i) if target == obj: return True return False @staticmethod def activeWindow(): """Traverses the list of known apps looking for one who has an immediate child (i.e., a window) whose state includes the active state. Returns the Python Accessible of the window that's active or None if no windows are active. """ apps = Utilities.knownApplications() for app in apps: for child in app: try: if child.getState().contains(pyatspi.STATE_ACTIVE): return child except: debug.printException(debug.LEVEL_FINEST) return None @staticmethod def allDescendants(root, onlyShowing=True): """Returns a list of all objects under the given root. Objects are returned in no particular order - this function does a simple tree traversal, ignoring the children of objects which report the MANAGES_DESCENDANTS state. Arguments: - root: the Accessible object to traverse - onlyShowing: examine only those objects that are SHOWING Returns: a list of all objects under the specified object """ if root.childCount <= 0: return [] objlist = [] for i, child in enumerate(root): debug.println(debug.LEVEL_FINEST, "Script.allDescendants looking at child %d" % i) if child \ and ((not onlyShowing) or (onlyShowing and \ (child.getState().contains(pyatspi.STATE_SHOWING)))): objlist.append(child) if (child.getState().contains( \ pyatspi.STATE_MANAGES_DESCENDANTS) == 0) \ and (child.childCount > 0): objlist.extend(Utilities.allDescendants(child, onlyShowing)) return objlist @staticmethod def ancestorWithRole(obj, ancestorRoles, stopRoles): """Returns the object of the specified roles which contains the given object, or None if the given object is not contained within an object the specified roles. Arguments: - obj: the Accessible object - ancestorRoles: the list of roles to look for - stopRoles: the list of roles to stop the search at """ if not obj: return None if not isinstance(ancestorRoles, [].__class__): ancestorRoles = [ancestorRoles] if not isinstance(stopRoles, [].__class__): stopRoles = [stopRoles] ancestor = None obj = obj.parent while obj and (obj != obj.parent): if obj.getRole() in ancestorRoles: ancestor = obj break elif obj.getRole() in stopRoles: break else: obj = obj.parent return ancestor def cellIndex(self, obj): """Returns the index of the cell which should be used with the table interface. This is necessary because in some apps we cannot count on getIndexInParent() returning the index we need. Arguments: -obj: the table cell whose index we need. """ return obj.getIndexInParent() def childNodes(self, obj): """Gets all of the children that have RELATION_NODE_CHILD_OF pointing to this expanded table cell. Arguments: -obj: the Accessible Object Returns: a list of all the child nodes """ try: table = obj.parent.queryTable() except: return [] else: if not obj.getState().contains(pyatspi.STATE_EXPANDED): return [] nodes = [] # First see if this accessible implements RELATION_NODE_PARENT_OF. # If it does, the full target list are the nodes. If it doesn't # we'll do an old-school, row-by-row search for child nodes. # relations = obj.getRelationSet() try: for relation in relations: if relation.getRelationType() == \ pyatspi.RELATION_NODE_PARENT_OF: for target in range(relation.getNTargets()): nodes.append(relation.getTarget(target)) return nodes except: pass index = self.cellIndex(obj) row = table.getRowAtIndex(index) col = table.getColumnAtIndex(index) nodeLevel = self.nodeLevel(obj) done = False # Candidates will be in the rows beneath the current row. # Only check in the current column and stop checking as # soon as the node level of a candidate is equal or less # than our current level. # for i in range(row+1, table.nRows): cell = table.getAccessibleAt(i, col) relations = cell.getRelationSet() for relation in relations: if relation.getRelationType() \ == pyatspi.RELATION_NODE_CHILD_OF: nodeOf = relation.getTarget(0) if self.isSameObject(obj, nodeOf): nodes.append(cell) else: currentLevel = self.nodeLevel(nodeOf) if currentLevel <= nodeLevel: done = True break if done: break return nodes def commonAncestor(self, a, b): """Finds the common ancestor between Accessible a and Accessible b. Arguments: - a: Accessible - b: Accessible """ debug.println(debug.LEVEL_FINEST, "script_utilities.commonAncestor...") if (not a) or (not b): return None if a == b: return a aParents = [a] try: parent = a.parent while parent and (parent.parent != parent): aParents.append(parent) parent = parent.parent aParents.reverse() except: debug.printException(debug.LEVEL_FINEST) bParents = [b] try: parent = b.parent while parent and (parent.parent != parent): bParents.append(parent) parent = parent.parent bParents.reverse() except: debug.printException(debug.LEVEL_FINEST) commonAncestor = None maxSearch = min(len(aParents), len(bParents)) i = 0 while i < maxSearch: if self.isSameObject(aParents[i], bParents[i]): commonAncestor = aParents[i] i += 1 else: break debug.println(debug.LEVEL_FINEST, "...script_utilities.commonAncestor") return commonAncestor def componentAtDesktopCoords(self, parent, x, y): """Get the descendant component at the given desktop coordinates. Arguments: - parent: The parent component we are searching below. - x: X coordinate. - y: Y coordinate. Returns end-node that contains the given coordinates, or None. """ acc = self.popupItemAtDesktopCoords(x, y) if acc: return acc container = parent while True: if container.getRole() == pyatspi.ROLE_PAGE_TAB_LIST: try: si = container.querySelection() container = si.getSelectedChild(0)[0] except NotImplementedError: pass try: ci = container.queryComponent() except: return None else: inner_container = container container = ci.getAccessibleAtPoint(x, y, pyatspi.DESKTOP_COORDS) if not container or container.queryComponent() == ci: break if inner_container == parent: return None else: return inner_container def defaultButton(self, obj): """Returns the default button in the dialog which contains obj. Arguments: - obj: the top-level object (e.g. window, frame, dialog) for which the status bar is sought. """ # There are some objects which are not worth descending. # skipRoles = [pyatspi.ROLE_TREE, pyatspi.ROLE_TREE_TABLE, pyatspi.ROLE_TABLE] if obj.getState().contains(pyatspi.STATE_MANAGES_DESCENDANTS) \ or obj.getRole() in skipRoles: return defaultButton = None # The default button is likely near the bottom of the window. # for i in range(obj.childCount - 1, -1, -1): if obj[i].getRole() == pyatspi.ROLE_PUSH_BUTTON \ and obj[i].getState().contains(pyatspi.STATE_IS_DEFAULT): defaultButton = obj[i] elif not obj[i].getRole() in skipRoles: defaultButton = self.defaultButton(obj[i]) if defaultButton: break return defaultButton @staticmethod def descendantsWithRole(root, role, onlyShowing=True): """Returns a list of all objects of a specific role beneath the given root. Arguments: - root: the Accessible object to traverse - role: the string describing the Accessible role of the object - onlyShowing: examine only those objects that are SHOWING Returns a list of descendants of the root that are of the given role. """ allObjs = Utilities.allDescendants(root, onlyShowing) return filter(lambda o: o.getRole() == role, allObjs) def displayedLabel(self, obj): """If there is an object labelling the given object, return the text being displayed for the object labelling this object. Otherwise, return None. Argument: - obj: the object in question Returns the string of the object labelling this object, or None if there is nothing of interest here. """ try: return self._script.generatorCache[self.DISPLAYED_LABEL][obj] except: if not self._script.generatorCache.has_key(self.DISPLAYED_LABEL): self._script.generatorCache[self.DISPLAYED_LABEL] = {} labelString = None labels = self.labelsForObject(obj) for label in labels: labelString = \ self.appendString(labelString, self.displayedText(label)) self._script.generatorCache[self.DISPLAYED_LABEL][obj] = labelString return self._script.generatorCache[self.DISPLAYED_LABEL][obj] def _displayedTextInComboBox(self, combo): """Returns the text being displayed in a combo box. If nothing is displayed, then None is returned. Arguments: - combo: the combo box Returns the text in the combo box or an empty string if nothing is displayed. """ displayedText = None # Find the text displayed in the combo box. This is either: # # 1) The last text object that's a child of the combo box # 2) The selected child of the combo box. # 3) The contents of the text of the combo box itself when # treated as a text object. # # Preference is given to #1, if it exists. # # If the label of the combo box is the same as the utterance for # the child object, then this utterance is only displayed once. # # [[[TODO: WDW - Combo boxes are complex beasts. This algorithm # needs serious work. Logged as bugzilla bug 319745.]]] # textObj = None for child in combo: if child and child.getRole() == pyatspi.ROLE_TEXT: textObj = child if textObj: [displayedText, caretOffset, startOffset] = \ self._script.getTextLineAtCaret(textObj) #print "TEXTOBJ", displayedText else: try: comboSelection = combo.querySelection() selectedItem = comboSelection.getSelectedChild(0) except: selectedItem = None if selectedItem: displayedText = self.displayedText(selectedItem) #print "SELECTEDITEM", displayedText elif combo.name and len(combo.name): # We give preference to the name over the text because # the text for combo boxes seems to never change in # some cases. The main one where we see this is in # the gaim "Join Chat" window. # displayedText = combo.name #print "NAME", displayedText else: [displayedText, caretOffset, startOffset] = \ self._script.getTextLineAtCaret(combo) # Set to None instead of empty string. displayedText = displayedText or None #print "TEXT", displayedText return displayedText def displayedText(self, obj): """Returns the text being displayed for an object. Arguments: - obj: the object Returns the text being displayed for an object or None if there isn't any text being shown. """ try: return self._script.generatorCache[self.DISPLAYED_TEXT][obj] except: displayedText = None role = obj.getRole() if role == pyatspi.ROLE_COMBO_BOX: displayedText = self._displayedTextInComboBox(obj) if not self._script.generatorCache.has_key(self.DISPLAYED_TEXT): self._script.generatorCache[self.DISPLAYED_TEXT] = {} self._script.generatorCache[self.DISPLAYED_TEXT][obj] = \ displayedText return self._script.generatorCache[self.DISPLAYED_TEXT][obj] # The accessible text of an object is used to represent what is # drawn on the screen. # try: text = obj.queryText() except NotImplementedError: pass else: displayedText = text.getText(0, self._script.getTextEndOffset(text)) # [[[WDW - HACK to account for things such as Gecko that want # to use the EMBEDDED_OBJECT_CHARACTER on a label to hold the # object that has the real accessible text for the label. We # detect this by the specfic case where the text for the # current object is a single EMBEDDED_OBJECT_CHARACTER. In # this case, we look to the child for the real text.]]] # unicodeText = displayedText.decode("UTF-8") if unicodeText \ and len(unicodeText) == 1 \ and unicodeText[0] == self.EMBEDDED_OBJECT_CHARACTER \ and obj.childCount > 0: try: displayedText = self.displayedText(obj[0]) except: debug.printException(debug.LEVEL_WARNING) elif unicodeText: # [[[TODO: HACK - Welll.....we'll just plain ignore any # text with EMBEDDED_OBJECT_CHARACTERs here. We still need a # general case to handle this stuff and expand objects # with EMBEDDED_OBJECT_CHARACTERs.]]] # for i in xrange(len(unicodeText)): if unicodeText[i] == self.EMBEDDED_OBJECT_CHARACTER: displayedText = None break if not displayedText: displayedText = obj.name # [[[WDW - HACK because push buttons can have labels as their # children. An example of this is the Font: button on the General # tab in the Editing Profile dialog in gnome-terminal. # if not displayedText and role == pyatspi.ROLE_PUSH_BUTTON: for child in obj: if child.getRole() == pyatspi.ROLE_LABEL: childText = self.displayedText(child) if childText and len(childText): displayedText = \ self.appendString(displayedText, childText) if not self._script.generatorCache.has_key(self.DISPLAYED_TEXT): self._script.generatorCache[self.DISPLAYED_TEXT] = {} self._script.generatorCache[self.DISPLAYED_TEXT][obj] = displayedText return self._script.generatorCache[self.DISPLAYED_TEXT][obj] def documentFrame(self): """Returns the document frame which is displaying the content. Note that this is intended primarily for web content.""" return None def documentFrameURI(self): """Returns the URI of the document frame that is active.""" return None @staticmethod def focusedObject(root): """Returns the accessible that has focus under or including the given root. TODO: This will currently traverse all children, whether they are visible or not and/or whether they are children of parents that manage their descendants. At some point, this method should be optimized to take such things into account. Arguments: - root: the root object where to start searching Returns the object with the FOCUSED state or None if no object with the FOCUSED state can be found. """ if root.getState().contains(pyatspi.STATE_FOCUSED): return root for child in root: try: candidate = Utilities.focusedObject(child) if candidate: return candidate except: pass return None def frameAndDialog(self, obj): """Returns the frame and (possibly) the dialog containing obj.""" results = [None, None] parent = obj.parent while parent and (parent.parent != parent): if parent.getRole() == pyatspi.ROLE_FRAME: results[0] = parent if parent.getRole() in [pyatspi.ROLE_DIALOG, pyatspi.ROLE_FILE_CHOOSER]: results[1] = parent parent = parent.parent return results def grabFocusBeforeRouting(self, obj, offset): """Whether or not we should perform a grabFocus before routing the cursor via the braille cursor routing keys. Arguments: - obj: the accessible object where the cursor should be routed - offset: the offset to which it should be routed Returns True if we should do an explicit grabFocus on obj prior to routing the cursor. """ return False def hasMatchingHierarchy(self, obj, rolesList): """Called to determine if the given object and it's hierarchy of parent objects, each have the desired roles. Please note: You should strongly consider an alternative means for determining that a given object is the desired item. Failing that, you should include only enough of the hierarchy to make the determination. If the developer of the application you are providing access to does so much as add an Adjustment to reposition a widget, this method can fail. You have been warned. Arguments: - obj: the accessible object to check. - rolesList: the list of desired roles for the components and the hierarchy of its parents. Returns True if all roles match. """ current = obj for role in rolesList: if current is None: return False if not isinstance(role, list): role = [role] if isinstance(role[0], str): current_role = current.getRoleName() else: current_role = current.getRole() if not current_role in role: return False current = self.validParent(current) return True def inFindToolbar(self, obj=None): """Returns True if the given object is in the Find toolbar. Arguments: - obj: an accessible object """ if not obj: obj = orca_state.locusOfFocus if obj and obj.getRole() == pyatspi.ROLE_ENTRY \ and obj.parent.getRole() == pyatspi.ROLE_TOOL_BAR: return True return False def isFunctionalDialog(self, obj): """Returns True if the window is a functioning as a dialog. This method should be subclassed by application scripts as needed. """ return False def isLayoutOnly(self, obj): """Returns True if the given object is a table and is for layout purposes only.""" layoutOnly = False if obj: attributes = obj.getAttributes() else: attributes = None if obj and (obj.getRole() == pyatspi.ROLE_TABLE) and attributes: for attribute in attributes: if attribute == "layout-guess:true": layoutOnly = True break elif obj and (obj.getRole() == pyatspi.ROLE_PANEL): text = self.displayedText(obj) label = self.displayedLabel(obj) if not ((label and len(label)) or (text and len(text))): layoutOnly = True if layoutOnly: debug.println(debug.LEVEL_FINEST, "Object deemed to be for layout purposes only: %s" \ % obj) return layoutOnly @staticmethod def isInActiveApp(obj): """Returns True if the given object is from the same application that currently has keyboard focus. Arguments: - obj: an Accessible object """ if not obj or not orca_state.locusOfFocus: return False return orca_state.locusOfFocus.getApplication() == obj.getApplication() def isLink(self, obj): """Returns True if obj is a link.""" return obj and obj.getRole() == pyatspi.ROLE_LINK def isReadOnlyTextArea(self, obj): """Returns True if obj is a text entry area that is read only.""" if not self.isTextArea(obj): return False state = obj.getState() readOnly = state.contains(pyatspi.STATE_FOCUSABLE) \ and not state.contains(pyatspi.STATE_EDITABLE) details = debug.getAccessibleDetails(debug.LEVEL_ALL, obj) debug.println(debug.LEVEL_ALL, "isReadOnlyTextArea=%s for %s" \ % (readOnly, details)) return readOnly def isSameObject(self, obj1, obj2): if (obj1 == obj2): return True elif (not obj1) or (not obj2): return False try: if (obj1.name != obj2.name) or (obj1.getRole() != obj2.getRole()): return False else: # Gecko sometimes creates multiple accessibles to represent # the same object. If the two objects have the same name # and the same role, check the extents. If those also match # then the two objects are for all intents and purposes the # same object. # extents1 = \ obj1.queryComponent().getExtents(pyatspi.DESKTOP_COORDS) extents2 = \ obj2.queryComponent().getExtents(pyatspi.DESKTOP_COORDS) if (extents1.x == extents2.x) and \ (extents1.y == extents2.y) and \ (extents1.width == extents2.width) and \ (extents1.height == extents2.height): return True # When we're looking at children of objects that manage # their descendants, we will often get different objects # that point to the same logical child. We want to be able # to determine if two objects are in fact pointing to the # same child. # If we cannot do so easily (i.e., object equivalence), we examine # the hierarchy and the object index at each level. # parent1 = obj1 parent2 = obj2 while (parent1 and parent2 and \ parent1.getState().contains( \ pyatspi.STATE_TRANSIENT) and \ parent2.getState().contains(pyatspi.STATE_TRANSIENT)): if parent1.getIndexInParent() != parent2.getIndexInParent(): return False parent1 = parent1.parent parent2 = parent2.parent if parent1 and parent2 and parent1 == parent2: return self.realActiveDescendant(obj1).name == \ self.realActiveDescendant(obj2).name except: pass # [[[TODO - JD: Why is this here? If it is truly limited to the # Java toolkit, it should be dealt with in Orca's Java toolkit # script. If it applies more broadly we should update the comment # to reflect that.]]] # # In java applications, TRANSIENT state is missing for tree items # (fix for bug #352250) # try: parent1 = obj1 parent2 = obj2 while parent1 and parent2 and \ parent1.getRole() == pyatspi.ROLE_LABEL and \ parent2.getRole() == pyatspi.ROLE_LABEL: if parent1.getIndexInParent() != parent2.getIndexInParent(): return False parent1 = parent1.parent parent2 = parent2.parent if parent1 and parent2 and parent1 == parent2: return True except: pass return False def isTextArea(self, obj): """Returns True if obj is a GUI component that is for entering text. Arguments: - obj: an accessible """ if self.isLink(obj): return False return obj and obj.getRole() in (pyatspi.ROLE_TEXT, pyatspi.ROLE_ENTRY, pyatspi.ROLE_PARAGRAPH, pyatspi.ROLE_TERMINAL) @staticmethod def knownApplications(): """Retrieves the list of currently running apps for the desktop as a list of Accessible objects. """ debug.println(debug.LEVEL_FINEST, "knownApplications...") apps = filter(lambda x: x is not None, pyatspi.Registry.getDesktop(0)) debug.println(debug.LEVEL_FINEST, "...knownApplications") return apps def labelsForObject(self, obj): """Return a list of the objects that are labelling this object. Argument: - obj: the object in question Returns a list of the objects that are labelling this object. """ # For some reason, some objects are labelled by the same thing # more than once. Go figure, but we need to check for this. # label = [] relations = obj.getRelationSet() allTargets = [] for relation in relations: if relation.getRelationType() \ == pyatspi.RELATION_LABELLED_BY: # The object can be labelled by more than one thing, so we just # get all the labels (from unique objects) and append them # together. An example of such objects live in the "Basic" # page of the gnome-accessibility-keyboard-properties app. # The "Delay" and "Speed" objects are labelled both by # their names and units. # for i in range(0, relation.getNTargets()): target = relation.getTarget(i) if not target in allTargets: allTargets.append(target) label.append(target) # [[[TODO: HACK - we've discovered oddness in hierarchies such as # the gedit Edit->Preferences dialog. In this dialog, we have # labeled groupings of objects. The grouping is done via a FILLER # with two children - one child is the overall label, and the # other is the container for the grouped objects. When we detect # this, we add the label to the overall context. # # We are also looking for objects which have a PANEL or a FILLER as # parent, and its parent has more children. Through these children, # a potential label with LABEL_FOR relation may exists. We want to # present this label. # This case can be seen in FileChooserDemo application, in Open dialog # window, the line with "Look In" label, a combobox and some # presentation buttons. # # Finally, we are searching the hierarchy of embedded components for # children that are labels.]]] # if not len(label): potentialLabels = [] useLabel = False if (obj.getRole() == pyatspi.ROLE_EMBEDDED): candidate = obj while candidate.childCount: candidate = candidate[0] # The parent of this object may contain labels # or it may contain filler that contains labels. # candidate = candidate.parent for child in candidate: if child.getRole() == pyatspi.ROLE_FILLER: candidate = child break # If there are labels in this embedded component, # they should be here. # for child in candidate: if child.getRole() == pyatspi.ROLE_LABEL: useLabel = True potentialLabels.append(child) elif ((obj.getRole() == pyatspi.ROLE_FILLER) \ or (obj.getRole() == pyatspi.ROLE_PANEL)) \ and (obj.childCount == 2): child0, child1 = obj child0_role = child0.getRole() child1_role = child1.getRole() if child0_role == pyatspi.ROLE_LABEL \ and not self.__hasLabelForRelation(child0) \ and child1_role in [pyatspi.ROLE_FILLER, \ pyatspi.ROLE_PANEL]: useLabel = True potentialLabels.append(child0) elif child1_role == pyatspi.ROLE_LABEL \ and not self.__hasLabelForRelation(child1) \ and child0_role in [pyatspi.ROLE_FILLER, \ pyatspi.ROLE_PANEL]: useLabel = True potentialLabels.append(child1) else: parent = obj.parent if parent and \ ((parent.getRole() == pyatspi.ROLE_FILLER) \ or (parent.getRole() == pyatspi.ROLE_PANEL)): for potentialLabel in parent: try: useLabel = self.__isLabeling(potentialLabel, obj) if useLabel: potentialLabels.append(potentialLabel) break except: pass if useLabel and len(potentialLabels): label = potentialLabels return label @staticmethod def linkIndex(obj, characterIndex): """A brute force method to see if an offset is a link. This is provided because not all Accessible Hypertext implementations properly support the getLinkIndex method. Returns an index of 0 or greater of the characterIndex is on a hyperlink. Arguments: -obj: the object with the Accessible Hypertext specialization -characterIndex: the text position to check """ if not obj: return -1 try: obj.queryText() except NotImplementedError: return -1 try: hypertext = obj.queryHypertext() except NotImplementedError: return -1 for i in xrange(hypertext.getNLinks()): link = hypertext.getLink(i) if (characterIndex >= link.startIndex) \ and (characterIndex <= link.endIndex): return i return -1 def nestingLevel(self, obj): """Determines the nesting level of this object in a list. If this object is not in a list relation, then 0 will be returned. Arguments: -obj: the Accessible object """ if not obj: return 0 try: return self._script.generatorCache[self.NESTING_LEVEL][obj] except: if not self._script.generatorCache.has_key(self.NESTING_LEVEL): self._script.generatorCache[self.NESTING_LEVEL] = {} nestingLevel = 0 parent = obj.parent while parent.parent.getRole() == pyatspi.ROLE_LIST: nestingLevel += 1 parent = parent.parent self._script.generatorCache[self.NESTING_LEVEL][obj] = nestingLevel return self._script.generatorCache[self.NESTING_LEVEL][obj] def nodeLevel(self, obj): """Determines the node level of this object if it is in a tree relation, with 0 being the top level node. If this object is not in a tree relation, then -1 will be returned. Arguments: -obj: the Accessible object """ if not obj: return -1 try: return self._script.generatorCache[self.NODE_LEVEL][obj] except: if not self._script.generatorCache.has_key(self.NODE_LEVEL): self._script.generatorCache[self.NODE_LEVEL] = {} nodes = [] node = obj done = False while not done: try: relations = node.getRelationSet() except (LookupError, RuntimeError): debug.println(debug.LEVEL_SEVERE, "nodeLevel() - Error getting RelationSet") return -1 node = None for relation in relations: if relation.getRelationType() \ == pyatspi.RELATION_NODE_CHILD_OF: node = relation.getTarget(0) break # We want to avoid situations where something gives us an # infinite cycle of nodes. Bon Echo has been seen to do # this (see bug 351847). # if (len(nodes) > 100) or nodes.count(node): debug.println(debug.LEVEL_WARNING, "nodeLevel detected a cycle!!!") done = True elif node: nodes.append(node) debug.println(debug.LEVEL_FINEST, "nodeLevel %d" % len(nodes)) else: done = True self._script.generatorCache[self.NODE_LEVEL][obj] = len(nodes) - 1 return self._script.generatorCache[self.NODE_LEVEL][obj] def popupItemAtDesktopCoords(self, x, y): """Since pop-up items often don't contain nested components, we need a way to efficiently determine if the cursor is over a menu item. Arguments: - x: X coordinate. - y: Y coordinate. Returns a menu item the mouse is over, or None. """ suspect_children = [] if self._script.lastSelectedMenu: try: si = self._script.lastSelectedMenu.querySelection() except NotImplementedError: return None if si.nSelectedChildren > 0: suspect_children = [si.getSelectedChild(0)] else: suspect_children = self._script.lastSelectedMenu for child in suspect_children: try: ci = child.queryComponent() except NotImplementedError: continue if ci.contains(x, y, pyatspi.DESKTOP_COORDS) \ and ci.getLayer() == pyatspi.LAYER_POPUP: return child @staticmethod def pursueForFlatReview(obj): """Determines if we should look any further at the object for flat review.""" try: state = obj.getState() except: debug.printException(debug.LEVEL_WARNING) return False else: return state.contains(pyatspi.STATE_SHOWING) def realActiveDescendant(self, obj): """Given an object that should be a child of an object that manages its descendants, return the child that is the real active descendant carrying useful information. Arguments: - obj: an object that should be a child of an object that manages its descendants. """ try: return self._script.\ generatorCache[self.REAL_ACTIVE_DESCENDANT][obj] except: if not self._script.\ generatorCache.has_key(self.REAL_ACTIVE_DESCENDANT): self._script.generatorCache[self.REAL_ACTIVE_DESCENDANT] = {} activeDescendant = None # If obj is a table cell and all of it's children are table cells # (probably cell renderers), then return the first child which has # a non zero length text string. If no such object is found, just # fall through and use the default approach below. See bug #376791 # for more details. # if obj.getRole() == pyatspi.ROLE_TABLE_CELL and obj.childCount: nonTableCellFound = False for child in obj: if child.getRole() != pyatspi.ROLE_TABLE_CELL: nonTableCellFound = True if not nonTableCellFound: for child in obj: try: text = child.queryText() except NotImplementedError: continue else: if text.getText(0, self._script.getTextEndOffset(text)): activeDescendant = child # [[[TODO: WDW - this is an odd hacky thing I've somewhat drawn # from Gnopernicus. The notion here is that we get an active # descendant changed event, but that object tends to have children # itself and we need to decide what to do. Well...the idea here # is that the last child (Gnopernicus chooses child(1)), tends to # be the child with information. The previous children tend to # be non-text or just there for spacing or something. You will # see this in the various table demos of gtk-demo and you will # also see this in the Contact Source Selector in Evolution. # # Just note that this is most likely not a really good solution # for the general case. That needs more thought. But, this # comment is here to remind us this is being done in poor taste # and we need to eventually clean up our act.]]] # if not activeDescendant and obj and obj.childCount: activeDescendant = obj[-1] self._script.generatorCache[self.REAL_ACTIVE_DESCENDANT][obj] = \ activeDescendant or obj return self._script.generatorCache[self.REAL_ACTIVE_DESCENDANT][obj] def showingDescendants(self, parent): """Given a parent that manages its descendants, return a list of Accessible children that are actually showing. This algorithm was inspired a little by the srw_elements_from_accessible logic in Gnopernicus, and makes the assumption that the children of an object that manages its descendants are arranged in a row and column format. Arguments: - parent: The accessible which manages its descendants Returns a list of Accessible descendants which are showing. """ import sys if not parent: return [] if not parent.getState().contains(pyatspi.STATE_MANAGES_DESCENDANTS) \ or parent.childCount <= 50: return [] try: icomponent = parent.queryComponent() except NotImplementedError: return [] descendants = [] parentExtents = icomponent.getExtents(0) # [[[TODO: WDW - HACK related to GAIL bug where table column # headers seem to be ignored: # http://bugzilla.gnome.org/show_bug.cgi?id=325809. The # problem is that this causes getAccessibleAtPoint to return # the cell effectively below the real cell at a given point, # making a mess of everything. So...we just manually add in # showing headers for now. The remainder of the logic below # accidentally accounts for this offset, yet it should also # work when bug 325809 is fixed.]]] # try: table = parent.queryTable() except NotImplementedError: table = None if table: for i in range(0, table.nColumns): header = table.getColumnHeader(i) if header: extents = header.queryComponent().getExtents(0) stateset = header.getState() if stateset.contains(pyatspi.STATE_SHOWING) \ and (extents.x >= 0) and (extents.y >= 0) \ and (extents.width > 0) and (extents.height > 0) \ and self.isVisibleRegion( extents.x, extents.y, extents.width, extents.height, parentExtents.x, parentExtents.y, parentExtents.width, parentExtents.height): descendants.append(header) # This algorithm goes left to right, top to bottom while attempting # to do *some* optimization over queries. It could definitely be # improved. The gridSize is a minimal chunk to jump around in the # table. # gridSize = 7 currentY = parentExtents.y while currentY < (parentExtents.y + parentExtents.height): currentX = parentExtents.x minHeight = sys.maxint index = -1 while currentX < (parentExtents.x + parentExtents.width): child = \ icomponent.getAccessibleAtPoint(currentX, currentY + 1, 0) if child: index = child.getIndexInParent() extents = child.queryComponent().getExtents(0) if extents.x >= 0 and extents.y >= 0: newX = extents.x + extents.width minHeight = min(minHeight, extents.height) if not descendants.count(child): descendants.append(child) else: newX = currentX + gridSize else: debug.println(debug.LEVEL_FINEST, "script_utilities.showingDescendants failed. " \ "Last valid child at index %d" % index) break if newX <= currentX: currentX += gridSize else: currentX = newX if minHeight == sys.maxint: minHeight = gridSize currentY += minHeight return descendants def statusBar(self, obj): """Returns the status bar in the window which contains obj. Arguments: - obj: the top-level object (e.g. window, frame, dialog) for which the status bar is sought. """ # There are some objects which are not worth descending. # skipRoles = [pyatspi.ROLE_TREE, pyatspi.ROLE_TREE_TABLE, pyatspi.ROLE_TABLE] if obj.getState().contains(pyatspi.STATE_MANAGES_DESCENDANTS) \ or obj.getRole() in skipRoles: return statusBar = None # The status bar is likely near the bottom of the window. # for i in range(obj.childCount - 1, -1, -1): if obj[i].getRole() == pyatspi.ROLE_STATUS_BAR: statusBar = obj[i] elif not obj[i] in skipRoles: statusBar = self.statusBar(obj[i]) if statusBar: break return statusBar @staticmethod def topLevelObject(obj): """Returns the top-level object (frame, dialog ...) containing obj, or None if obj is not inside a top-level object. Arguments: - obj: the Accessible object """ debug.println(debug.LEVEL_FINEST, "Finding top-level object for source.name=" + obj.name or "None") while obj and obj.parent and (obj != obj.parent) \ and (obj.parent.getRole() != pyatspi.ROLE_APPLICATION): obj = obj.parent debug.println(debug.LEVEL_FINEST, "--> obj.name=" + obj.name or "None") if obj and obj.parent and \ (obj.parent.getRole() == pyatspi.ROLE_APPLICATION): pass else: obj = None return obj def unrelatedLabels(self, root): """Returns a list containing all the unrelated (i.e., have no relations to anything and are not a fundamental element of a more atomic component like a combo box) labels under the given root. Note that the labels must also be showing on the display. Arguments: - root the Accessible object to traverse Returns a list of unrelated labels under the given root. """ allLabels = self.descendantsWithRole(root, pyatspi.ROLE_LABEL) unrelatedLabels = [] for label in allLabels: relations = label.getRelationSet() if len(relations) == 0: parent = label.parent if parent and (parent.getRole() == pyatspi.ROLE_PUSH_BUTTON): pass elif parent and (parent.getRole() == pyatspi.ROLE_PANEL) \ and (parent.name == label.name): pass elif label.getState().contains(pyatspi.STATE_SHOWING): unrelatedLabels.append(label) # Now sort the labels based on their geographic position, top to # bottom, left to right. This is a very inefficient sort, but the # assumption here is that there will not be a lot of labels to # worry about. # sortedLabels = [] for label in unrelatedLabels: label_extents = \ label.queryComponent().getExtents(pyatspi.DESKTOP_COORDS) index = 0 for sortedLabel in sortedLabels: sorted_extents = \ sortedLabel.queryComponent().getExtents( pyatspi.DESKTOP_COORDS) if (label_extents.y > sorted_extents.y) \ or ((label_extents.y == sorted_extents.y) \ and (label_extents.x > sorted_extents.x)): index += 1 else: break sortedLabels.insert(index, label) return sortedLabels def unfocusedAlertAndDialogCount(self, obj): """If the current application has one or more alert or dialog windows and the currently focused window is not an alert or a dialog, return a count of the number of alert and dialog windows, otherwise return a count of zero. Arguments: - obj: the Accessible object Returns the alert and dialog count. """ alertAndDialogCount = 0 app = obj.getApplication() window = Utilities.topLevelObject(obj) if window and window.getRole() != pyatspi.ROLE_ALERT and \ window.getRole() != pyatspi.ROLE_DIALOG and \ not self.isFunctionalDialog(window): for child in app: if child.getRole() == pyatspi.ROLE_ALERT or \ child.getRole() == pyatspi.ROLE_DIALOG or \ self.isFunctionalDialog(child): alertAndDialogCount += 1 return alertAndDialogCount def uri(self, obj): """Return the URI for a given link object. Arguments: - obj: the Accessible object. """ try: return obj.queryHyperlink().getURI(0) except: return None def validParent(self, obj): """Returns the first valid parent/ancestor of obj. We need to do this in some applications and toolkits due to bogus hierarchies. Arguments: - obj: the Accessible object """ if not obj: return None return obj.parent ######################################################################### # # # Utilities for working with the accessible text interface # # # ######################################################################### @staticmethod def adjustTextSelection(obj, offset): """Adjusts the end point of a text selection Arguments: - obj: the Accessible object. - offset: the new end point - can be to the left or to the right depending on the direction of selection """ try: text = obj.queryText() except: return if not text.getNSelections(): caretOffset = text.caretOffset startOffset = min(offset, caretOffset) endOffset = max(offset, caretOffset) text.addSelection(startOffset, endOffset) else: startOffset, endOffset = text.getSelection(0) if offset < startOffset: startOffset = offset else: endOffset = offset text.setSelection(0, startOffset, endOffset) def allSelectedText(self, obj): """Get all the text applicable text selections for the given object. including any previous or next text objects that also have selected text and add in their text contents. Arguments: - obj: the text object to start extracting the selected text from. Returns: all the selected text contents plus the start and end offsets within the text for the given object. """ textContents = "" startOffset = 0 endOffset = 0 text = obj.queryText() if text.getNSelections() > 0: [textContents, startOffset, endOffset] = self.selectedText(obj) current = obj morePossibleSelections = True while morePossibleSelections: morePossibleSelections = False for relation in current.getRelationSet(): if relation.getRelationType() == pyatspi.RELATION_FLOWS_FROM: prevObj = relation.getTarget(0) prevObjText = prevObj.queryText() if prevObjText.getNSelections() > 0: [newTextContents, start, end] = \ self.selectedText(prevObj) textContents = newTextContents + " " + textContents current = prevObj morePossibleSelections = True else: displayedText = prevObjText.getText(0, self._script.getTextEndOffset(prevObjText)) if len(displayedText) == 0: current = prevObj morePossibleSelections = True break current = obj morePossibleSelections = True while morePossibleSelections: morePossibleSelections = False for relation in current.getRelationSet(): if relation.getRelationType() == pyatspi.RELATION_FLOWS_TO: nextObj = relation.getTarget(0) nextObjText = nextObj.queryText() if nextObjText.getNSelections() > 0: [newTextContents, start, end] = \ self.selectedText(nextObj) textContents += " " + newTextContents current = nextObj morePossibleSelections = True else: displayedText = nextObjText.getText(0, self._script.getTextEndOffset(nextObjText)) if len(displayedText) == 0: current = nextObj morePossibleSelections = True break return [textContents, startOffset, endOffset] @staticmethod def allTextSelections(obj): """Get a list of text selections in the given accessible object, equivalent to getNSelections()*texti.getSelection() Arguments: - obj: An accessible. Returns list of start and end offsets for multiple selections, or an empty list if nothing is selected or if the accessible does not support the text interface. """ try: text = obj.queryText() except: return [] rv = [] for i in xrange(text.getNSelections()): rv.append(text.getSelection(i)) return rv @staticmethod def characterOffsetInParent(obj): """Returns the character offset of the embedded object character for this object in its parent's accessible text. Arguments: - obj: an Accessible that should implement the accessible hyperlink specialization. Returns an integer representing the character offset of the embedded object character for this hyperlink in its parent's accessible text, or -1 something was amuck. """ try: hyperlink = obj.queryHyperlink() except NotImplementedError: offset = -1 else: # We need to make sure that this is an embedded object in # some accessible text (as opposed to an imagemap link). # try: obj.parent.queryText() except NotImplementedError: offset = -1 else: offset = hyperlink.startIndex return offset @staticmethod def clearTextSelection(obj): """Clears the text selection if the object supports it. Arguments: - obj: the Accessible object. """ try: text = obj.queryText() except: return for i in xrange(text.getNSelections()): text.removeSelection(0) def expandEOCs(self, obj, startOffset=0, endOffset=-1): """Expands the current object replacing EMBEDDED_OBJECT_CHARACTERS with their text. Arguments - obj: the object whose text should be expanded - startOffset: the offset of the first character to be included - endOffset: the offset of the last character to be included Returns the fully expanded text for the object. """ string = self.substring(obj, startOffset, endOffset) if string: unicodeText = string.decode("UTF-8") if unicodeText and self.EMBEDDED_OBJECT_CHARACTER in unicodeText: # If we're not getting the full text of this object, but # rather a substring, we need to figure out the offset of # the first child within this substring. # childOffset = 0 for child in obj: if Utilities.characterOffsetInParent(child) >= startOffset: break childOffset += 1 toBuild = list(unicodeText) count = toBuild.count(self.EMBEDDED_OBJECT_CHARACTER) for i in xrange(count): index = toBuild.index(self.EMBEDDED_OBJECT_CHARACTER) child = obj[i + childOffset] childText = self.expandEOCs(child) if not childText: childText = "" toBuild[index] = childText.decode("UTF-8") string = "".join(toBuild) return string def hasTextSelections(self, obj): """Return an indication of whether this object has selected text. Note that it's possible that this object has no selection, but is part of a selected text area. Because of this, we need to check the objects on either side to see if they are none zero length and have text selections. Arguments: - obj: the text object to start checking for selected text. Returns: an indication of whether this object has selected text, or adjacent text objects have selected text. """ currentSelected = False otherSelected = False text = obj.queryText() nSelections = text.getNSelections() if nSelections: currentSelected = True else: otherSelected = False text = obj.queryText() displayedText = text.getText(0, self._script.getTextEndOffset(text)) if (text.caretOffset == 0) or len(displayedText) == 0: current = obj morePossibleSelections = True while morePossibleSelections: morePossibleSelections = False for relation in current.getRelationSet(): if relation.getRelationType() == \ pyatspi.RELATION_FLOWS_FROM: prevObj = relation.getTarget(0) prevObjText = prevObj.queryText() if prevObjText.getNSelections() > 0: otherSelected = True else: displayedText = prevObjText.getText(0, self._script.getTextEndOffset(prevObjText)) if len(displayedText) == 0: current = prevObj morePossibleSelections = True break current = obj morePossibleSelections = True while morePossibleSelections: morePossibleSelections = False for relation in current.getRelationSet(): if relation.getRelationType() == \ pyatspi.RELATION_FLOWS_TO: nextObj = relation.getTarget(0) nextObjText = nextObj.queryText() if nextObjText.getNSelections() > 0: otherSelected = True else: displayedText = nextObjText.getText(0, self._script.getTextEndOffset(nextObjText)) if len(displayedText) == 0: current = nextObj morePossibleSelections = True break return [currentSelected, otherSelected] @staticmethod def isTextSelected(obj, startOffset, endOffset): """Returns an indication of whether the text is selected by comparing the text offset with the various selected regions of text for this accessible object. Arguments: - obj: the Accessible object. - startOffset: text start offset. - endOffset: text end offset. Returns an indication of whether the text is selected. """ if startOffset == endOffset: return False try: text = obj.queryText() except: return False for i in xrange(text.getNSelections()): [startSelOffset, endSelOffset] = text.getSelection(i) if (startOffset >= startSelOffset) and (endOffset <= endSelOffset): return True return False def isWordMisspelled(self, obj, offset): """Identifies if the current word is flagged as misspelled by the application. Different applications and toolkits flag misspelled words differently. Thus each script will likely need to implement its own version of this method. Arguments: - obj: An accessible which implements the accessible text interface. - offset: Offset in the accessible's text for which to retrieve the attributes. Returns True if the word is flagged as misspelled. """ return False def offsetsForPhrase(self, obj): """Return the start and end offset for the given phrase Arguments: - obj: the Accessible object """ try: text = obj.queryText() except: return [0, 0] lastPos = self._script.pointOfReference.get("lastCursorPosition") startOffset = lastPos[1] endOffset = text.caretOffset # Swap values if in wrong order. # if (startOffset > endOffset and endOffset != -1) or startOffset == -1: temp = endOffset endOffset = startOffset startOffset = temp return [startOffset, endOffset] def offsetsForLine(self, obj): """Return the start and end offset for the given line Arguments: - obj: the Accessible object """ lineAndOffsets = self._script.getTextLineAtCaret(obj) return [lineAndOffsets[1], lineAndOffsets[2]] def offsetsForWord(self, obj): """Return the start and end offset for the given word Arguments: - obj: the Accessible object """ try: text = obj.queryText() except: return [0, 0] wordAndOffsets = text.getTextAtOffset( text.caretOffset, pyatspi.TEXT_BOUNDARY_WORD_START) return [wordAndOffsets[1], wordAndOffsets[2]] def offsetsForChar(self, obj): """Return the start and end offset for the given character Arguments: - obj: the Accessible object """ try: text = obj.queryText() except: return [0, 0] lastKey, mods = self.lastKeyAndModifiers() if mods & settings.SHIFT_MODIFIER_MASK and lastKey == "Right": startOffset = text.caretOffset - 1 endOffset = text.caretOffset else: startOffset = text.caretOffset endOffset = text.caretOffset + 1 return [startOffset, endOffset] @staticmethod def queryNonEmptyText(obj): """Get the text interface associated with an object, if it is non-empty. Arguments: - obj: an accessible object """ try: text = obj.queryText() except: pass else: if text.characterCount: return text return None @staticmethod def selectedText(obj): """Get the text selection for the given object. Arguments: - obj: the text object to extract the selected text from. Returns: the selected text contents plus the start and end offsets within the text. """ textContents = "" startOffset = endOffset = 0 textObj = obj.queryText() nSelections = textObj.getNSelections() for i in range(0, nSelections): [startOffset, endOffset] = textObj.getSelection(i) debug.println(debug.LEVEL_FINEST, "getSelectedText: selection start=%d, end=%d" % \ (startOffset, endOffset)) selectedText = textObj.getText(startOffset, endOffset) debug.println(debug.LEVEL_FINEST, "getSelectedText: selected text=<%s>" % selectedText) if i > 0: textContents += " " textContents += selectedText return [textContents, startOffset, endOffset] def setCaretOffset(self, obj, offset): """Set the caret offset on a given accessible. Similar to Accessible.setCaretOffset() Arguments: - obj: Given accessible object. - offset: Offset to hich to set the caret. """ try: texti = obj.queryText() except: return None texti.setCaretOffset(offset) def substring(self, obj, startOffset, endOffset): """Returns the substring of the given object's text specialization. Arguments: - obj: an accessible supporting the accessible text specialization - startOffset: the starting character position - endOffset: the ending character position. Note that an end offset of -1 means the last character """ try: text = obj.queryText() except: return "" return text.getText(startOffset, endOffset) def textAttributes(self, acc, offset, get_defaults=False): """Get the text attributes run for a given offset in a given accessible Arguments: - acc: An accessible. - offset: Offset in the accessible's text for which to retrieve the attributes. - get_defaults: Get the default attributes as well as the unique ones. Default is True Returns a dictionary of attributes, a start offset where the attributes begin, and an end offset. Returns ({}, 0, 0) if the accessible does not supprt the text attribute. """ rv = {} try: text = acc.queryText() except: return rv, 0, 0 if get_defaults: stringAndDict = \ self.stringToKeysAndDict(text.getDefaultAttributes()) rv.update(stringAndDict[1]) attrString, start, end = text.getAttributes(offset) stringAndDict = self.stringToKeysAndDict(attrString) rv.update(stringAndDict[1]) return rv, start, end def unicodeText(self, obj): """Returns the unicode text for an object or None if the object doesn't implement the accessible text specialization. """ text = self.queryNonEmptyText(obj) if text: unicodeText = text.getText(0, -1).decode("UTF-8") else: unicodeText = None return unicodeText def willEchoCharacter(self, event): """Given a keyboard event containing an alphanumeric key, determine if the script is likely to echo it as a character. """ if not orca_state.locusOfFocus or not settings.enableEchoByCharacter: return False # The check here in English is something like this: "If this # character echo is enabled, then character echo is likely to # happen if the locus of focus is a focusable editable text # area or terminal and neither of the Ctrl, Alt, or Orca # modifiers are pressed. If that's the case, then character # echo will kick in for us." # return (self.isTextArea(orca_state.locusOfFocus)\ or orca_state.locusOfFocus.getRole() \ == pyatspi.ROLE_ENTRY) \ and (orca_state.locusOfFocus.getRole() \ == pyatspi.ROLE_TERMINAL \ or (not self.isReadOnlyTextArea(orca_state.locusOfFocus) \ and (orca_state.locusOfFocus.getState().contains( \ pyatspi.STATE_FOCUSABLE)))) \ and not (event.modifiers & settings.ORCA_CTRL_MODIFIER_MASK) def wordAtCoords(self, acc, x, y): """Get the word at the given coords in the accessible. Arguments: - acc: Accessible that supports the Text interface. - x: X coordinate. - y: Y coordinate. Returns a tuple containing the word, start offset, and end offset. """ try: ti = acc.queryText() except NotImplementedError: return '', 0, 0 text_contents = ti.getText(0, self._script.getTextEndOffset(ti)) line_offsets = [] start_offset = 0 while True: try: end_offset = text_contents.index('\n', start_offset) except ValueError: line_offsets.append((start_offset, len(text_contents))) break line_offsets.append((start_offset, end_offset)) start_offset = end_offset + 1 for start, end in line_offsets: bx, by, bw, bh = \ ti.getRangeExtents(start, end, pyatspi.DESKTOP_COORDS) bb = mouse_review.BoundingBox(bx, by, bw, bh) if bb.isInBox(x, y): start_offset = 0 word_offsets = [] while True: try: end_offset = \ text_contents[start:end].index(' ', start_offset) except ValueError: word_offsets.append((start_offset, len(text_contents[start:end]))) break word_offsets.append((start_offset, end_offset)) start_offset = end_offset + 1 for a, b in word_offsets: bx, by, bw, bh = \ ti.getRangeExtents(start+a, start+b, pyatspi.DESKTOP_COORDS) bb = mouse_review.BoundingBox(bx, by, bw, bh) if bb.isInBox(x, y): return text_contents[start+a:start+b], start+a, start+b return '', 0, 0 ######################################################################### # # # Debugging and Reporting Utilities # # # ######################################################################### def _isInterestingObj(self, obj): import inspect interesting = False if getattr(obj, "__class__", None): name = obj.__class__.__name__ if name not in ["function", "type", "list", "dict", "tuple", "wrapper_descriptor", "module", "method_descriptor", "member_descriptor", "instancemethod", "builtin_function_or_method", "frame", "classmethod", "classmethod_descriptor", "_Environ", "MemoryError", "_Printer", "_Helper", "getset_descriptor", "weakref", "property", "cell", "staticmethod", "EventListener", "KeystrokeListener", "KeyBinding", "InputEventHandler", "Rolename"]: try: filename = inspect.getabsfile(obj.__class__) if filename.index("orca"): interesting = True except: pass return interesting def _detectCycle(self, obj, visitedObjs, indent=""): """Attempts to discover a cycle in object references.""" # [[[TODO: WDW - not sure this really works.]]] import gc visitedObjs.append(obj) for referent in gc.get_referents(obj): try: if visitedObjs.index(referent): if self._isInterestingObj(referent): print indent, "CYCLE!!!!", `referent` break except: pass self._detectCycle(referent, visitedObjs, " " + indent) visitedObjs.remove(obj) def printAncestry(self, child): """Prints a hierarchical view of a child's ancestry.""" if not child: return ancestorList = [child] parent = child.parent while parent and (parent.parent != parent): ancestorList.insert(0, parent) parent = parent.parent indent = "" for ancestor in ancestorList: line = indent + "+- " + \ debug.getAccessibleDetails(debug.LEVEL_OFF, ancestor) debug.println(debug.LEVEL_OFF, line) print line indent += " " def printApps(self): """Prints a list of all applications to stdout.""" apps = self.knownApplications() line = "There are %d Accessible applications" % len(apps) debug.println(debug.LEVEL_OFF, line) print line for app in apps: line = debug.getAccessibleDetails( debug.LEVEL_OFF, app, " App: ", False) debug.println(debug.LEVEL_OFF, line) print line for child in app: line = debug.getAccessibleDetails( debug.LEVEL_OFF, child, " Window: ", False) debug.println(debug.LEVEL_OFF, line) print line if child.parent != app: debug.println(debug.LEVEL_OFF, " WARNING: child's parent is not app!!!") return True def printHierarchy(self, root, ooi, indent="", onlyShowing=True, omitManaged=True): """Prints the accessible hierarchy of all children Arguments: -indent: Indentation string -root: Accessible where to start -ooi: Accessible object of interest -onlyShowing: If True, only show children painted on the screen -omitManaged: If True, omit children that are managed descendants """ if not root: return if root == ooi: line = indent + "(*) " + debug.getAccessibleDetails( debug.LEVEL_OFF, root) else: line = indent + "+- " + debug.getAccessibleDetails( debug.LEVEL_OFF, root) debug.println(debug.LEVEL_OFF, line) print line rootManagesDescendants = root.getState().contains( pyatspi.STATE_MANAGES_DESCENDANTS) for child in root: if child == root: line = indent + " " + "WARNING CHILD == PARENT!!!" debug.println(debug.LEVEL_OFF, line) print line elif not child: line = indent + " " + "WARNING Child IS NONE!!!" debug.println(debug.LEVEL_OFF, line) print line elif self.validParent(child) != root: line = indent + " " + "WARNING CHILD.PARENT != PARENT!!!" debug.println(debug.LEVEL_OFF, line) print line else: paint = (not onlyShowing) or (onlyShowing and \ child.getState().contains(pyatspi.STATE_SHOWING)) paint = paint \ and ((not omitManaged) \ or (omitManaged and not rootManagesDescendants)) if paint: self.printHierarchy(child, ooi, indent + " ", onlyShowing, omitManaged) def scriptInfo(self): """Output useful information on the current script via speech and braille. This information will be helpful to script writers. """ infoString = "SCRIPT INFO: Script name='%s'" % self._script.name app = orca_state.locusOfFocus.getApplication() if orca_state.locusOfFocus and app: infoString += " Application name='%s'" \ % app.name try: infoString += " Toolkit name='%s'" \ % app.toolkitName except: infoString += " Toolkit unknown" try: infoString += " Version='%s'" \ % app.version except: infoString += " Version unknown" debug.println(debug.LEVEL_OFF, infoString) print infoString self._script.speakMessage(infoString) self._script.displayBrailleMessage(infoString) return True ######################################################################### # # # Miscellaneous Utilities # # # ######################################################################### def _addRepeatSegment(self, segment, line, respectPunctuation=True): """Add in the latest line segment, adjusting for repeat characters and punctuation. Arguments: - segment: the segment of repeated characters. - line: the current built-up line to characters to speak. - respectPunctuation: if False, ignore punctuation level. Returns: the current built-up line plus the new segment, after adjusting for repeat character counts and punctuation. """ import punctuation_settings import chnames style = settings.verbalizePunctuationStyle isPunctChar = True try: level, action = punctuation_settings.getPunctuationInfo(segment[0]) except: isPunctChar = False count = len(segment) if (count >= settings.repeatCharacterLimit) \ and (not segment[0] in self._script.whitespace): if (not respectPunctuation) \ or (isPunctChar and (style <= level)): repeatChar = chnames.getCharacterName(segment[0]) # Translators: Orca will tell you how many characters # are repeated on a line of text. For example: "22 # space characters". The %d is the number and the %s # is the spoken word for the character. # line += " " \ + ngettext("%(count)d %(repeatChar)s character", "%(count)d %(repeatChar)s characters", count) \ % {"count" : count, "repeatChar": repeatChar} else: line += segment else: line += segment return line def _pronunciationForSegment(self, segment): """Adjust the word segment to potentially replace it with what those words actually sound like. Two pronunciation dictionaries are checked. First the application specific one (which might be empty), then the default (global) one. Arguments: - segment: the string to adjust for words in the pronunciation dictionaries. Returns: a new word segment adjusted for words found in the pronunciation dictionaries, or the original word segment if there was no dictionary entry. """ import pronunciation_dict newSegment = pronunciation_dict.getPronunciation( segment, self._script.app_pronunciation_dict) if newSegment == segment: newSegment = pronunciation_dict.getPronunciation(segment) return newSegment def adjustForLinks(self, obj, line, startOffset): """Adjust line to include the word "link" after any hypertext links. Arguments: - obj: the accessible object that this line came from. - line: the string to adjust for links. - startOffset: the caret offset at the start of the line. Returns: a new line adjusted to add the speaking of "link" after text which is also a link. """ import punctuation_settings line = line.decode("UTF-8") endOffset = startOffset + len(line) try: hyperText = obj.queryHypertext() nLinks = hyperText.getNLinks() except: nLinks = 0 adjustedLine = list(line) for n in range(nLinks, 0, -1): link = hyperText.getLink(n - 1) # We only care about links in the string, line: # if startOffset < link.endIndex <= endOffset: index = link.endIndex - startOffset elif startOffset <= link.startIndex < endOffset: index = len(line) if link.endIndex < endOffset: index -= 1 else: continue # Translators: this indicates that this piece of # text is a hypertext link. # linkString = " " + _("link").decode("UTF-8") # If the link was not followed by a whitespace or punctuation # character, then add in a space to make it more presentable. # nextChar = "" if index < len(line): nextChar = adjustedLine[index] if not (nextChar in self._script.whitespace \ or punctuation_settings.getPunctuationInfo(nextChar)): linkString += " " adjustedLine[index:index] = linkString return "".join(adjustedLine).encode("UTF-8") def adjustForPronunciation(self, line): """Adjust the line to replace words in the pronunciation dictionary, with what those words actually sound like. Arguments: - line: the string to adjust for words in the pronunciation dictionary. Returns: a new line adjusted for words found in the pronunciation dictionary. """ newLine = "" words = self.WORDS_RE.split(line.decode("UTF-8")) for word in words: if word.isalnum(): word = self._pronunciationForSegment(word) newLine += word if line != newLine: debug.println(debug.LEVEL_FINEST, "adjustForPronunciation: \n From '%s'\n To '%s'" \ % (line, newLine)) return newLine.encode("UTF-8") else: return line def adjustForRepeats(self, line): """Adjust line to include repeat character counts. As some people will want this and others might not, there is a setting in settings.py that determines whether this functionality is enabled. repeatCharacterLimit = If is 0, then there would be no repeat characters. Otherwise would be the number of same characters (or more) in a row that cause the repeat character count output. If the value is set to 1, 2 or 3 then it's treated as if it was zero. In other words, no repeat character count is given. Arguments: - line: the string to adjust for repeat character counts. Returns: a new line adjusted for repeat character counts (if enabled). """ line = line.decode("UTF-8") if (len(line) < 4) or (settings.repeatCharacterLimit < 4): return line.encode("UTF-8") newLine = u'' segment = lastChar = line[0] multipleChars = False for i in range(1, len(line)): if line[i] == lastChar: segment += line[i] else: multipleChars = True newLine = self._addRepeatSegment(segment, newLine) segment = line[i] lastChar = line[i] newLine = self._addRepeatSegment(segment, newLine, multipleChars) # Pylint is confused and flags this with the following error: # # E1103:5188:Script.adjustForRepeats: Instance of 'True' has # no 'encode' member (but some types could not be inferred) # # We know newLine is a unicode string, so we'll just tell pylint # that we know what we are doing. # # pylint: disable-msg=E1103 return newLine.encode("UTF-8") def adjustForDigits(self, string): """Adjusts the string to convert digit-like text, such as subscript and superscript numbers, into actual digits. Arguments: - string: the string to be adjusted Returns: a new string which contains actual digits. """ uString = string.decode("UTF-8") subscripted = set(re.findall(self.SUBSCRIPTS_RE, uString)) superscripted = set(re.findall(self.SUPERSCRIPTS_RE, uString)) for number in superscripted: new = map(lambda d: str(self.SUPERSCRIPT_DIGITS.index(d)), number) # Translators: This string is part of the presentation of an # item that includes one or several consequtive superscripted # characters, e.g. 'X' followed by 'superscript 2' followed by # 'superscript 3' should be presented as 'X superscript 23'. # newString = _(" superscript %s") % "".join(new) uString = re.sub(number, newString, uString) for number in subscripted: new = map(lambda d: str(self.SUBSCRIPT_DIGITS.index(d)), number) # Translators: This string is part of the presentation of an # item that includes one or several consequtive subscripted # characters, e.g. 'X' followed by 'subscript 2' followed by # 'subscript 3', should be presented as 'X subscript 23.' # newString = _(" subscript %s") % "".join(new) uString = re.sub(number, newString, uString) return uString.encode("UTF-8") @staticmethod def absoluteMouseCoordinates(): """Gets the absolute position of the mouse pointer.""" import gtk rootWindow = gtk.Window().get_screen().get_root_window() x, y, modifiers = rootWindow.get_pointer() return x, y @staticmethod def appendString(text, newText, delimiter=" "): """Appends the newText to the given text with the delimiter in between and returns the new string. Edge cases, such as no initial text or no newText, are handled gracefully.""" if not newText or len(newText) == 0: return text elif text and len(text): return text + delimiter + newText else: return newText def isAutoTextEvent(self, event): """Returns True if event is associated with text being autocompleted or autoinserted or autocorrected or autosomethingelsed. Arguments: - event: the accessible event being examined """ return False def isSentenceDelimiter(self, currentChar, previousChar): """Returns True if we are positioned at the end of a sentence. This is determined by checking if the current character is a white space character and the previous character is one of the normal end-of-sentence punctuation characters. Arguments: - currentChar: the current character - previousChar: the previous character Returns True if the given character is a sentence delimiter. """ if not isinstance(currentChar, unicode): currentChar = currentChar.decode("UTF-8") if not isinstance(previousChar, unicode): previousChar = previousChar.decode("UTF-8") if currentChar == '\r' or currentChar == '\n': return True return currentChar in self._script.whitespace \ and previousChar in '!.?:;' def isWordDelimiter(self, character): """Returns True if the given character is a word delimiter. Arguments: - character: the character in question Returns True if the given character is a word delimiter. """ if not isinstance(character, unicode): character = character.decode("UTF-8") return character in self._script.whitespace \ or character in '!*+,-./:;<=>?@[\]^_{|}' \ or character == self._script.NO_BREAK_SPACE_CHARACTER @staticmethod def isVisibleRegion(ax, ay, awidth, aheight, bx, by, bwidth, bheight): """Returns true if any portion of region 'a' is in region 'b'""" highestBottom = min(ay + aheight, by + bheight) lowestTop = max(ay, by) leftMostRightEdge = min(ax + awidth, bx + bwidth) rightMostLeftEdge = max(ax, bx) if lowestTop <= highestBottom \ and rightMostLeftEdge <= leftMostRightEdge: return True elif aheight == 0: if awidth == 0: return lowestTop == highestBottom \ and leftMostRightEdge == rightMostLeftEdge else: return leftMostRightEdge <= rightMostLeftEdge elif awidth == 0: return lowestTop <= highestBottom return False @staticmethod def lastKeyAndModifiers(): """Convenience method which returns a tuple containing the event string and modifiers of the last non-modifier key event or ("", 0) if there is no such event.""" if isinstance(orca_state.lastInputEvent, input_event.KeyboardEvent) \ and orca_state.lastNonModifierKeyEvent: eventStr = orca_state.lastNonModifierKeyEvent.event_string mods = orca_state.lastInputEvent.modifiers else: eventStr = "" mods = 0 return (eventStr, mods) def mnemonicShortcutAccelerator(self, obj): """Gets the mnemonic, accelerator string and possibly shortcut for the given object. These are based upon the first accessible action for the object. Arguments: - obj: the Accessible object Returns: list containing strings: [mnemonic, shortcut, accelerator] """ try: return self._script.generatorCache[self.KEY_BINDING][obj] except: if not self._script.generatorCache.has_key(self.KEY_BINDING): self._script.generatorCache[self.KEY_BINDING] = {} try: action = obj.queryAction() except NotImplementedError: self._script.generatorCache[self.KEY_BINDING][obj] = ["", "", ""] return self._script.generatorCache[self.KEY_BINDING][obj] # Action is a string in the format, where the mnemonic and/or # accelerator can be missing. # # ;; # # The keybindings in should be separated by ":" # bindingStrings = action.getKeyBinding(0).decode("UTF-8").split(';') if len(bindingStrings) == 3: mnemonic = bindingStrings[0] fullShortcut = bindingStrings[1] accelerator = bindingStrings[2] elif len(bindingStrings) > 0: mnemonic = "" fullShortcut = bindingStrings[0] try: accelerator = bindingStrings[1] except: accelerator = "" else: mnemonic = "" fullShortcut = "" accelerator = "" fullShortcut = fullShortcut.replace("<","") fullShortcut = fullShortcut.replace(">"," ") fullShortcut = fullShortcut.replace(":"," ").strip() # If the accelerator or mnemonic strings includes a Space, # make sure we speak it. # if mnemonic.endswith(" "): # Translators: this is the spoken word for the space character # mnemonic += _("space") mnemonic = mnemonic.replace("<","") mnemonic = mnemonic.replace(">"," ").strip() if accelerator.endswith(" "): # Translators: this is the spoken word for the space character # accelerator += _("space") accelerator = accelerator.replace("<","") accelerator = accelerator.replace(">"," ").strip() debug.println(debug.LEVEL_FINEST, "script_utilities.getKeyBinding: " \ + repr([mnemonic, fullShortcut, accelerator])) if not self._script.generatorCache.has_key(self.KEY_BINDING): self._script.generatorCache[self.KEY_BINDING] = {} self._script.generatorCache[self.KEY_BINDING][obj] = \ [mnemonic, fullShortcut, accelerator] return self._script.generatorCache[self.KEY_BINDING][obj] @staticmethod def stringToKeysAndDict(string): """Converts a string made up of a series of :; pairs into a dictionary of keys and values. Text before the colon is the key and text afterwards is the value. The final semi-colon, if found, is ignored. Arguments: - string: the string of tokens containing :; pairs. Returns a list containing two items: A list of the keys in the order they were extracted from the string and a dictionary of key/value items. """ try: items = [s.strip() for s in string.split(";")] items = filter(lambda item: len(item.split(':')) == 2, items) keys = map(lambda item: item.split(':')[0].strip(), items) dictionary = dict(map(lambda item: item.split(':'), items)) except: return [], {} return [keys, dictionary] def textForValue(self, obj): """Returns the text to be displayed for the object's current value. Arguments: - obj: the Accessible object that may or may not have a value. Returns a string representing the value. """ # Use ARIA "valuetext" attribute if present. See # http://bugzilla.gnome.org/show_bug.cgi?id=552965 # attributes = obj.getAttributes() for attribute in attributes: if attribute.startswith("valuetext"): return attribute[10:] try: value = obj.queryValue() except NotImplementedError: return "" # OK, this craziness is all about trying to figure out the most # meaningful formatting string for the floating point values. # The number of places to the right of the decimal point should # be set by the minimumIncrement, but the minimumIncrement isn't # always set. So...we'll default the minimumIncrement to 1/100 # of the range. But, if max == min, then we'll just go for showing # them off to two meaningful digits. # try: minimumIncrement = value.minimumIncrement except: minimumIncrement = 0.0 if minimumIncrement == 0.0: minimumIncrement = (value.maximumValue - value.minimumValue) \ / 100.0 try: decimalPlaces = max(0, -math.log10(minimumIncrement)) except: try: decimalPlaces = max(0, -math.log10(value.minimumValue)) except: try: decimalPlaces = max(0, -math.log10(value.maximumValue)) except: decimalPlaces = 0 formatter = "%%.%df" % decimalPlaces valueString = formatter % value.currentValue #minString = formatter % value.minimumValue #maxString = formatter % value.maximumValue # [[[TODO: WDW - probably want to do this as a percentage at some # point? Logged as bugzilla bug 319743.]]] # return valueString @staticmethod def unicodeValueString(character): """ Returns a four hex digit representation of the given character Arguments: - The character to return representation Returns a string representaition of the given character unicode vlue """ try: if not isinstance(character, unicode): character = character.decode('UTF-8') return "%04x" % ord(character) except: debug.printException(debug.LEVEL_WARNING) return ""