3Ic@sdZddkZddkZddklZlZddklZddkl Z de fdYZ de fd YZ d e fd YZ d e fd YZde fdYZde efdYZde efdYZde efdYZde efdYZdZdZdZdZedjo endddddddd dddg ZdS( sh Class for representing hierarchical language structures, such as syntax trees and morphological trees. iN(t Productiont Nonterminal(tProbabilisticMixIn(t slice_boundstTreecBseZdZeedZedZdZdZdZdZ dZ dZ d Z d Z d Zd Zd ZdZdZdZdZdZddZedZdZdZdZdZdeddddZedddd Zeedd!Z d"Z!e"e!Z!ed#Z#d$Z$ed%Z%e"d&eeeeed'Z&e"d(Z'd)Z(d*Z)d+Z*d,dd-d&ed.Z+d/Z,d0Z-RS(1s A hierarchical structure. Each C{Tree} represents a single hierarchical grouping of leaves and subtrees. For example, each constituent in a syntax tree is represented by a single C{Tree}. A tree's children are encoded as a C{list} of leaves and subtrees, where a X{leaf} is a basic (non-tree) value; and a X{subtree} is a nested C{Tree}. Any other properties that a C{Tree} defines are known as X{node properties}, and are used to add information about individual hierarchical groupings. For example, syntax trees use a NODE property to label syntactic constituents with phrase tags, such as "NP" and"VP". Several C{Tree} methods use X{tree positions} to specify children or descendants of a tree. Tree positions are defined as follows: - The tree position M{i} specifies a C{Tree}'s M{i}th child. - The tree position C{()} specifies the C{Tree} itself. - If C{M{p}} is the tree position of descendant M{d}, then C{M{p}+(M{i})} specifies the C{M{i}}th child of M{d}. I.e., every tree position is either a single index C{M{i}}, specifying C{self[M{i}]}; or a sequence C{(M{i1}, M{i2}, ..., M{iN})}, specifying C{self[M{i1}][M{i2}]...[M{iN}]}. cCs|djoti|Sn|djo8t|tptd|in|i|SnLt|tpt|d otd|inti|||SdS(Ns;%s: Expected a node value and child list or a single stringt__iter__s.%s() argument 2 should be a list, not a string( tNonetlistt__new__t isinstancet basestringt TypeErrort__name__tparsethasattr(tclst node_or_strtchildren((s/p/zhu/06/nlp/nltk/nltk/tree.pyR:s  cCs2|djodSnti||||_dS(s Construct a new tree. This constructor can be called in one of two ways: - C{Tree(node, children)} constructs a new tree with the specified node value and list of children. - C{Tree(s)} constructs a new tree by parsing the string C{s}. It is equivalent to calling the class method C{Tree.parse(s)}. N(RRt__init__tnode(tselfRR((s/p/zhu/06/nlp/nltk/nltk/tree.pyRIscCs;t|tptSn|i|ijoti||S(N(R RtFalseRRt__eq__(Rtother((s/p/zhu/06/nlp/nltk/nltk/tree.pyRdscCs ||j S(N((RR((s/p/zhu/06/nlp/nltk/nltk/tree.pyt__ne__gscCs;t|tptSn|i|ijpti||S(N(R RRRRt__lt__(RR((s/p/zhu/06/nlp/nltk/nltk/tree.pyRiscCs;t|tptSn|i|ijpti||S(N(R RRRRt__le__(RR((s/p/zhu/06/nlp/nltk/nltk/tree.pyRlscCs;t|tptSn|i|ijpti||S(N(R RtTrueRRt__gt__(RR((s/p/zhu/06/nlp/nltk/nltk/tree.pyRoscCs;t|tptSn|i|ijpti||S(N(R RRRRt__ge__(RR((s/p/zhu/06/nlp/nltk/nltk/tree.pyRrscCstddS(Ns$Tree does not support multiplication(R (Rtv((s/p/zhu/06/nlp/nltk/nltk/tree.pyt__mul__zscCstddS(Ns$Tree does not support multiplication(R (RR((s/p/zhu/06/nlp/nltk/nltk/tree.pyt__rmul__|scCstddS(NsTree does not support addition(R (RR((s/p/zhu/06/nlp/nltk/nltk/tree.pyt__add__~scCstddS(NsTree does not support addition(R (RR((s/p/zhu/06/nlp/nltk/nltk/tree.pyt__radd__scCst|ttfoti||Sn_t|djo|SnDt|djo|t|dSn|t|d|dSdS(Nii(R tinttsliceRt __getitem__tlen(Rtindex((s/p/zhu/06/nlp/nltk/nltk/tree.pyR%scCst|ttfoti|||Sn_t|djotdn<t|djo|||d}t|tot||i}q t|d}q Wd|S(s3 @return: The height of this tree. The height of a tree containing no children is 1; the height of a tree containing only leaves is 2; and the height of any other tree is one plus the maximum of its children's heights. @rtype: C{int} ii(R Rtmaxtheight(Rtmax_child_heightR/((s/p/zhu/06/nlp/nltk/nltk/tree.pyR2s tpreordercsg}|djo|idnxjt|D]\\}t|to0|i|}|ifd|Dq1|ifq1W|djo|idn|S( sg @param order: One of: C{preorder}, C{postorder}, C{bothorder}, C{leaves}. R4t bothorderc3s x|]}f|VqWdS(N((t.0tp(ti(s/p/zhu/06/nlp/nltk/nltk/tree.pys st postorder(R4R5((R9R5((R.t enumerateR Rt treepositionsR,(Rtordert positionsR/tchildpos((R8s/p/zhu/06/nlp/nltk/nltk/tree.pyR;s !ccsf| p ||o |VnxA|D]9}t|to#x |i|D] }|VqKWq%q%WdS(s Generate all the subtrees of this tree, optionally restricted to trees matching the filter function. @type filter: C{function} @param filter: the function to filter all local trees N(R Rtsubtrees(RtfilterR/tsubtree((s/p/zhu/06/nlp/nltk/nltk/tree.pyR?s cCszt|itp tdntt|it|g}x2|D]*}t|to||i7}qHqHW|S(s Generate the productions that correspond to the non-terminal nodes of the tree. For each subtree of the form (P: C1 C2 ... Cn) this produces a production of the form P -> C1 C2 ... Cn. @rtype: list of C{Production}s sPProductions can only be generated from trees having node labels that are strings( R RtstrR RRt _child_namesRt productions(RtprodsR/((s/p/zhu/06/nlp/nltk/nltk/tree.pyRDs  !cCsXg}xK|D]C}t|to|i|iq |i||ifq W|S(s @return: a list of tuples containing leaves and pre-terminals (part-of-speech tags). The order reflects the order of the leaves in the tree's hierarchical structure. @rtype: C{list} of C{tuples} (R RR,tposR.R(RRFR/((s/p/zhu/06/nlp/nltk/nltk/tree.pyRFscCs|djotdn|dfg}x|o|i\}}t|tp#|djo|Sq|d8}q/xBtt|dddD]$}|i||||ffqWq/WtddS(s/ @return: The tree position of the C{index}-th leaf in this tree. I.e., if C{tp=self.leaf_treeposition(i)}, then C{self[tp]==self.leaves()[i]}. @raise IndexError: If this tree contains fewer than C{index+1} leaves, or if C{index<0}. isindex must be non-negativeiis-index must be less than or equal to len(self)N((R)tpopR RtrangeR&R.(RR'tstackR*ttreeposR8((s/p/zhu/06/nlp/nltk/nltk/tree.pytleaf_treepositions  'cCs||jotdn|i|}|i|d}xNtt|D]:}|t|jp||||jo || SqRqRW|S(s @return: The tree position of the lowest descendant of this tree that dominates C{self.leaves()[start:end]}. @raise ValueError: if C{end <= start} send must be greater than starti(t ValueErrorRKRHR&(Rtstarttendt start_treepost end_treeposR8((s/p/zhu/06/nlp/nltk/nltk/tree.pyttreeposition_spanning_leaves+s (trightit|t^cCs-ddkl}|||||||dS(s This method can modify a tree in three ways: 1. Convert a tree into its Chomsky Normal Form (CNF) equivalent -- Every subtree has either two non-terminals or one terminal as its children. This process requires the creation of more"artificial" non-terminal nodes. 2. Markov (vertical) smoothing of children in new artificial nodes 3. Horizontal (parent) annotation of nodes @param factor: Right or left factoring method (default = "right") @type factor: C{string} = [left|right] @param horzMarkov: Markov order for sibling smoothing in artificial nodes (None (default) = include all siblings) @type horzMarkov: C{int} | None @param vertMarkov: Markov order for parent smoothing (0 (default) = no vertical annotation) @type vertMarkov: C{int} | None @param childChar: A string used in construction of the artificial nodes, separating the head of the original subtree from the child nodes that have yet to be expanded (default = "|") @type childChar: C{string} @param parentChar: A string used to separate the node representation from its vertical annotation @type parentChar: C{string} i(tchomsky_normal_formN(ttreetransformsRU(Rtfactort horzMarkovt vertMarkovt childChart parentCharRU((s/p/zhu/06/nlp/nltk/nltk/tree.pyRUAst+cCs*ddkl}||||||dS(s This method modifies the tree in three ways: 1. Transforms a tree in Chomsky Normal Form back to its original structure (branching greater than two) 2. Removes any parent annotation (if it exists) 3. (optional) expands unary subtrees (if previously collapsed with collapseUnary(...) ) @param expandUnary: Flag to expand unary or not (default = True) @type expandUnary: C{boolean} @param childChar: A string separating the head node from its children in an artificial node (default = "|") @type childChar: C{string} @param parentChar: A sting separating the node label from its parent annotation (default = "^") @type parentChar: C{string} @param unaryChar: A string joining two non-terminals in a unary production (default = "+") @type unaryChar: C{string} i(tun_chomsky_normal_formN(RVR](Rt expandUnaryRZR[t unaryCharR]((s/p/zhu/06/nlp/nltk/nltk/tree.pyR]\scCs'ddkl}|||||dS(s( Collapse subtrees with a single child (ie. unary productions) into a new non-terminal (Tree node) joined by 'joinChar'. This is useful when working with algorithms that do not allow unary productions, and completely removing the unary productions would require loss of useful information. The Tree is modified directly (since it is passed by reference) and no value is returned. @param collapsePOS: 'False' (default) will not collapse the parent of leaf nodes (ie. Part-of-Speech tags) since they are always unary productions @type collapsePOS: C{boolean} @param collapseRoot: 'False' (default) will not modify the root production if it is unary. For the Penn WSJ treebank corpus, this corresponds to the TOP -> productions. @type collapseRoot: C{boolean} @param joinChar: A string used to connect collapsed node values (default = "+") @type joinChar: C{string} i(tcollapse_unaryN(RVR`(Rt collapsePOSt collapseRoottjoinCharR`((s/p/zhu/06/nlp/nltk/nltk/tree.pyR`rscCsVt|to>g}|D]}||i|q~}||i|Sn|SdS(s Convert a tree between different subtypes of Tree. C{cls} determines which class will be used to encode the new tree. @type val: L{Tree} @param val: The tree that should be converted. @return: The new C{Tree}. N(R RtconvertR(Rtvalt_[1]R/R((s/p/zhu/06/nlp/nltk/nltk/tree.pyRds *cCs2|p|i|i|Sn|ii|SdS(N(t __class__RRd(Rtdeep((s/p/zhu/06/nlp/nltk/nltk/tree.pytcopyscCstS(N(t ImmutableTree(R((s/p/zhu/06/nlp/nltk/nltk/tree.pyt _frozen_classscCs|i}|djo|i|}nP|idt}x+|idD]}|||||>> parse_node(str) -> value For example, these functions could be used to parse nodes and leaves whose values should be some type other than string (such as L{FeatStruct }). Note that by default, node strings and leaf strings are delimited by whitespace and brackets; to override this default, use the C{node_pattern} and C{leaf_pattern} arguments. @type node_pattern: C{str} @type leaf_pattern: C{str} @param node_pattern, leaf_pattern: Regular expression patterns used to find node and leaf substrings in C{s}. By default, both nodes patterns are defined to match any sequence of non-whitespace non-bracket characters. @type remove_empty_top_bracketing: C{bool} @param remove_empty_top_bracketing: If the resulting tree has an empty node label, and is length one, then return its single child instead. This is useful for treebank trees, which sometimes contain an extra level of bracketing. @return: A tree corresponding to the string representation C{s}. If this class method is called using a subclass of C{Tree}, then it will return a tree of that type. @rtype: C{Tree} is"brackets must be a length-2 strings\sswhitespace brackets not alloweds [^\s%s%s]+s%s\s*(%s)?|%s|(%s)iis end-of-stringit(R R R&R tretsearchtescapeRtcompiletfinditertgroupt _parse_errortlstripR.RGtAssertionErrorR(Rtstbracketst parse_nodet parse_leaft node_patternt leaf_patterntremove_empty_top_bracketingtopen_btclose_bt open_patternt close_patternttoken_reRItmatchttokenRRttree((s/p/zhu/06/nlp/nltk/nltk/tree.pyR sV5$     . ""*c Cs|djot|d}}n|i|i}}d|i||d|f}|iddidd}|}t||djo||d d}n|djod||d}d }n|d d|dd |f7}t|d S(s Display a friendly error message when parsing a tree string fails. @param s: The string we're parsing. @param match: regexp match of the problem token. @param expecting: what we expected to see instead. s end-of-strings1%s.parse(): expected %r but got %r %sat index %d.t i s s i s...i s %s"%s" %s^iiNs s (R&RMRwR treplaceRL(RR{Rt expectingRFRtmsgtoffset((s/p/zhu/06/nlp/nltk/nltk/tree.pyRx s   cCsddkl}||dS(sP Open a new window containing a graphical diagram of this tree. i(t draw_treesN(tnltk.draw.treeR(RR((s/p/zhu/06/nlp/nltk/nltk/tree.pytdraw=scCs3did|D}d|ii|i|fS(Ns, cssx|]}t|VqWdS(N(trepr(R6tc((s/p/zhu/06/nlp/nltk/nltk/tree.pys Ess %s(%r, [%s])(tjoinRgR R(Rtchildstr((s/p/zhu/06/nlp/nltk/nltk/tree.pyt__repr__DscCs |iS(N(tpprint(R((s/p/zhu/06/nlp/nltk/nltk/tree.pyt__str__HsiFRqc Csr|i|||}t|||jo|Snt|itod|d|i|f}nd|d|i|f}x|D]}t|to7|dd|d|i||d|||7}qt|to'|dd|ddi|7}qt|t o*| o"|dd|dd|7}q|dd|dd |7}qW||d S( sS @return: A pretty-printed string representation of this tree. @rtype: C{string} @param margin: The right margin at which to do line-wrapping. @type margin: C{int} @param indent: The indentation level at which printing begins. This number is used to decide how far to indent subsequent lines. @type indent: C{int} @param nodesep: A string that is used to separate the node from the children. E.g., the default value C{':'} gives trees like C{(S: (NP: I) (VP: (V: saw) (NP: it)))}. s%s%s%sis%s%r%ss Rit/s%ss%ri( t _pprint_flatR&R RR RRttupleRRB(RtmargintindenttnodeseptparenstquotesR{R/((s/p/zhu/06/nlp/nltk/nltk/tree.pyRKs""'""c Cs d|idddddd S( s Returns a representation of the tree compatible with the LaTeX qtree package. This consists of the string C{\Tree} followed by the parse tree represented in bracketed notation. For example, the following result was generated from a parse tree of the sentence C{The announcement astounded us}:: \Tree [.I'' [.N'' [.D The ] [.N' [.N announcement ] ] ] [.I' [.V'' [.V' [.V astounded ] [.N'' [.N' [.N us ] ] ] ] ] ] ] See U{http://www.ling.upenn.edu/advice/latex.html} for the LaTeX style file for the qtree package. @return: A latex qtree representation of this tree. @rtype: C{string} s\Tree RiRRqRs[.s ](s[.s ](R(R((s/p/zhu/06/nlp/nltk/nltk/tree.pytpprint_latex_qtreepsc Cs g}x|D]}t|to |i|i|||q t|to|idi|q t|to| o|id|q |id|q Wt|ito/d|d|i|t i||dfSn,d|d|i|t i||dfSdS(NRs%ss%rs %s%s%s %s%siis %s%r%s %s%s( R RR.RRRRBRR tstring(RRRRt childstrsR/((s/p/zhu/06/nlp/nltk/nltk/tree.pyRs (.R t __module__t__doc__RRRRRRRRRRR R!R"R%R(R+R-R0R2R;R?RDRFRKRQRURR]RR`Rdt classmethodRiRkRpR RxRRRRRR(((s/p/zhu/06/nlp/nltk/nltk/tree.pyRsX                           k   % RjcBseZedZdZdZdZdZdZdZ dZ dZ ed Z d Z d Zd Zd ZdZdZeeeZRS(cCs||djodSntt|i||y"t|it|f|_Wn%tt fj ot dnXdS(Ns9ImmutableTree's node value and children must be immutable( RtsuperRjRRlRRt_hashR RL(RRR((s/p/zhu/06/nlp/nltk/nltk/tree.pyRs "cCs tddS(Ns"ImmutableTrees may not be modified(RL(R((s/p/zhu/06/nlp/nltk/nltk/tree.pyR(scCs tddS(Ns"ImmutableTrees may not be modified(RL(R((s/p/zhu/06/nlp/nltk/nltk/tree.pyt __setslice__scCs tddS(Ns"ImmutableTrees may not be modified(RL(R((s/p/zhu/06/nlp/nltk/nltk/tree.pyR+scCs tddS(Ns"ImmutableTrees may not be modified(RL(R((s/p/zhu/06/nlp/nltk/nltk/tree.pyt __delslice__scCs tddS(Ns"ImmutableTrees may not be modified(RL(R((s/p/zhu/06/nlp/nltk/nltk/tree.pyt__iadd__scCs tddS(Ns"ImmutableTrees may not be modified(RL(R((s/p/zhu/06/nlp/nltk/nltk/tree.pyt__imul__scCs tddS(Ns"ImmutableTrees may not be modified(RL(RR((s/p/zhu/06/nlp/nltk/nltk/tree.pyR.scCs tddS(Ns"ImmutableTrees may not be modified(RL(RR((s/p/zhu/06/nlp/nltk/nltk/tree.pyR,scCs tddS(Ns"ImmutableTrees may not be modified(RL(RR((s/p/zhu/06/nlp/nltk/nltk/tree.pyRGscCs tddS(Ns"ImmutableTrees may not be modified(RL(RR((s/p/zhu/06/nlp/nltk/nltk/tree.pytremovescCs tddS(Ns"ImmutableTrees may not be modified(RL(R((s/p/zhu/06/nlp/nltk/nltk/tree.pytreversescCs tddS(Ns"ImmutableTrees may not be modified(RL(R((s/p/zhu/06/nlp/nltk/nltk/tree.pytsortscCs|iS(N(R(R((s/p/zhu/06/nlp/nltk/nltk/tree.pyt__hash__scCs*t|do tdn||_dS(s|Set self._node. This will only succeed the first time the node value is set, which should occur in Tree.__init__().Rs"ImmutableTrees may not be modifiedN(RRLt_node(RR((s/p/zhu/06/nlp/nltk/nltk/tree.pyt _set_nodes cCs|iS(N(R(R((s/p/zhu/06/nlp/nltk/nltk/tree.pyt _get_nodes(R RRRR(RR+RRRR.R,RGRRRRRRtpropertyR(((s/p/zhu/06/nlp/nltk/nltk/tree.pyRjs"               tAbstractParentedTreecBseZdZedZedZdZdZdZ dZ dZ dZ d d Z d Zeed od ZdZdZnRS(s An abstract base class for L{Tree}s that automatically maintain pointers to their parents. These parent pointers are updated whenever any change is made to a tree's structure. Two subclasses are currently defined: - L{ParentedTree} is used for tree structures where each subtree has at most one parent. This class should be used in cases where there is no"sharing" of subtrees. - L{MultiParentedTree} is used for tree structures where a subtree may have zero or more parents. This class should be used in cases where subtrees may be shared. Subclassing =========== The C{AbstractParentedTree} class redefines all operations that modify a tree's structure to call two methods, which are used by subclasses to update parent information: - L{_setparent()} is called whenever a new child is added. - L{_delparent()} is called whenever a child is removed. cCs|djodSntt|i||xDt|D]6\}}t|to|i||dtq;q;Wx>t|D]0\}}t|to|i||qqWdS(Ntdry_run( RRRRR:R Rt _setparentR(RRRR8R/((s/p/zhu/06/nlp/nltk/nltk/tree.pyRs  cCstddS(s Update C{child}'s parent pointer to point to self. This method is only called if C{child}'s type is L{Tree}; i.e., it is not called when adding a leaf to a tree. This method is always called before the child is actually added to C{self}'s child list. @type child: L{Tree} @type index: C{int} @param index: The index of C{child} in C{self}. @raise TypeError: If C{child} is a tree with an impropriate type. Typically, if C{child} is a tree, then its type needs to match C{self}'s type. This prevents mixing of different tree types (single-parented, multi-parented, and non-parented). @param dry_run: If true, the don't actually set the child's parent pointer; just check for any error conditions, and raise an exception if one is found. sAbstract base classN(Rz(RR/R'R((s/p/zhu/06/nlp/nltk/nltk/tree.pyRscCstddS(s Update C{child}'s parent pointer to not point to self. This method is only called if C{child}'s type is L{Tree}; i.e., it is not called when removing a leaf from a tree. This method is always called before the child is actually removed from C{self}'s child list. @type child: L{Tree} @type index: C{int} @param index: The index of C{child} in C{self}. sAbstract base classN(Rz(RR/R'((s/p/zhu/06/nlp/nltk/nltk/tree.pyt _delparents cCsut|tout||\}}xCt||D]2}t||to|i|||q5q5Wtt|i|nt|t o|djo|t |7}n|djot dnt||to|i|||ntt|i|nYt |djot dn6t |djo||d=n||d|d=dS(Nisindex out of ranges(The tree position () may not be deleted.i( R R$RtxrangeRRRRR+R#R&R)(RR'RMtstopR8((s/p/zhu/06/nlp/nltk/nltk/tree.pyR+s$!c Cstt|to.t||\}}t|ttfpt|}nxHt|D]:\}}t|to|i|||dtqXqXWxCt ||D]2}t||to|i |||qqWxBt|D]4\}}t|to|i|||qqWt t |i ||n3t|to|djo|t|7}n|djotdn|||jodSnt|to|i||nt||to|i |||nt t |i ||n_t|djotdn<t|djo|||dstdocsA The parent of this tree, or C{None} if it has no parent.s4 The index of this tree in its parent. I.e., C{ptree.parent[ptree.parent_index] is ptree}. Note that C{ptree.parent_index} is not necessarily equal to C{ptree.parent.index(ptree)}, since the C{index()} method returns the first child that is I{equal} to its argument.sB The left sibling of this tree, or C{None} if it has none.sC The right sibling of this tree, or C{None} if it has none.s The root of this tree. I.e., the unique ancestor of this tree whose parent is C{None}. If C{ptree.parent} is C{None}, then C{ptree} is its own root.s The tree position of this tree, relative to the root of the tree. I.e., C{ptree.root[ptree.treeposition] is ptree}.cCsSt|tpt|||jpt|i|jptd|_dS(N(R RRzRR(RR/R'((s/p/zhu/06/nlp/nltk/nltk/tree.pyRscCs\t|tptddn|idj otdn|p ||_ndS(Ns"Can not insert a non-ParentedTree sinto a ParentedTrees3Can not insert a subtree that already has a parent.(R RR RRRL(RR/R'R((s/p/zhu/06/nlp/nltk/nltk/tree.pyRs(R RRRRRkRRRRRRtparentRt left_siblingt right_siblingtroott treepositionRJRRR(((s/p/zhu/06/nlp/nltk/nltk/tree.pyRs.                    tMultiParentedTreecBseZdZedZdZdZdZdZdZ dZ e dd d Z e ed d Z e ed d Ze e d d ZdZdZdZedZRS(s. A L{Tree} that automatically maintains parent pointers for multi-parented trees. The following read-only property values are automatically updated whenever the structure of a multi-parented tree is modified: L{parents}, L{parent_indices}, L{left_siblings}, L{right_siblings}, L{roots}, L{treepositions}. Each C{MultiParentedTree} may have zero or more parents. In particular, subtrees may be shared. If a single C{MultiParentedTree} is used as multiple children of the same parent, then that parent will appear multiple times in its C{parents} property. C{MultiParentedTrees} should never be used in the same tree as C{Trees} or C{ParentedTrees}. Mixing tree implementations may result in incorrect parent pointers and in C{TypeError} exceptions. cCs;|djodSng|_tt|i||dS(N(Rt_parentsRRR(RRR((s/p/zhu/06/nlp/nltk/nltk/tree.pyR(s cCstS(N(tImmutableMultiParentedTree(R((s/p/zhu/06/nlp/nltk/nltk/tree.pyRk2scCsRg}|iD]=}t|D]*\}}||jo|||fqqq~S(N(RR:(RRfRR'R/((s/p/zhu/06/nlp/nltk/nltk/tree.pyt_get_parent_indices8s cCsDg}|iD],\}}|djo|||dqq~S(Nii(R(RRfRR'((s/p/zhu/06/nlp/nltk/nltk/tree.pyt_get_left_siblings>scCsNg}|iD]6\}}|t|djo|||dqq~S(Ni(RR&(RRfRR'((s/p/zhu/06/nlp/nltk/nltk/tree.pyt_get_right_siblingsCscCs|ihiS(N(t_get_roots_helpertvalues(R((s/p/zhu/06/nlp/nltk/nltk/tree.pyt _get_rootsHscCsC|io%x2|iD]}|i|qWn||t|<|S(N(RRtid(RtresultR((s/p/zhu/06/nlp/nltk/nltk/tree.pyRKs   cCs t|iS((RR(R((s/p/zhu/06/nlp/nltk/nltk/tree.pyRSsRs The set of parents of this tree. If this tree has no parents, then C{parents} is the empty set. To check if a tree is used as multiple children of the same parent, use the L{parent_indices} property. @type: C{list} of L{MultiParentedTree}s| A list of all left siblings of this tree, in any of its parent trees. A tree may be its own left sibling if it is used as multiple contiguous children of the same parent. A tree may appear multiple times in this list if it is the left sibling of this tree with respect to multiple parents. @type: C{list} of L{MultiParentedTree}s A list of all right siblings of this tree, in any of its parent trees. A tree may be its own right sibling if it is used as multiple contiguous children of the same parent. A tree may appear multiple times in this list if it is the right sibling of this tree with respect to multiple parents. @type: C{list} of L{MultiParentedTree}s The set of all roots of this tree. This set is formed by tracing all possible parent paths until trees with no parents are found. @type: C{list} of L{MultiParentedTree}cCsX||ijogSn=g}t|D]$\}}||jo ||q)q)~SdS(sW Return a list of the indices where this tree occurs as a child of C{parent}. If this child does not occur as a child of C{parent}, then the empty list is returned. The following is always true:: for parent_index in ptree.parent_indices(parent): parent[parent_index] is ptree N(RR:(RRRfR'R/((s/p/zhu/06/nlp/nltk/nltk/tree.pytparent_indicests cCs||jo dgSnjg}|iD]T}|i|D]>}t|D]+\}}||jo|||fqIqIq9q&~SdS(s Return a list of all tree positions that can be used to reach this multi-parented tree starting from C{root}. I.e., the following is always true:: for treepos in ptree.treepositions(root): root[treepos] is ptree N((RR;R:(RRRfRRJR'R/((s/p/zhu/06/nlp/nltk/nltk/tree.pyR;s   cCst|tpt|||jpttg}|iD]}||jo ||q@q@~djptxIt|D]+\}}||jo||joPqqW|ii|dS(Ni(R RRzR&RR:R(RR/R'RfR7R8R((s/p/zhu/06/nlp/nltk/nltk/tree.pyRsI #cCsit|tptddn|p:x7|iD]}||joPq5q5W|ii|ndS(Ns'Can not insert a non-MultiParentedTree sinto a MultiParentedTree(R RR RR.(RR/R'RR((s/p/zhu/06/nlp/nltk/nltk/tree.pyRs (R RRRRRkRRRRRRtparentst left_siblingstright_siblingstrootsRR;RRR(((s/p/zhu/06/nlp/nltk/nltk/tree.pyRs(                 RcBseZedZRS(cCs2|djodSntt|i||dS(N(RRRR(RRR((s/p/zhu/06/nlp/nltk/nltk/tree.pyRs(R RRR(((s/p/zhu/06/nlp/nltk/nltk/tree.pyRsRcBseZedZRS(cCs2|djodSntt|i||dS(N(RRRR(RRR((s/p/zhu/06/nlp/nltk/nltk/tree.pyRs(R RRR(((s/p/zhu/06/nlp/nltk/nltk/tree.pyRstProbabilisticTreecBsweZedZedZdZdZdZdZdZ dZ e dZ d Z ee Z RS( cKstt|i|||S(N(RRR(RRRt prob_kwargs((s/p/zhu/06/nlp/nltk/nltk/tree.pyRscKs<|djodSnti|||ti||dS(N(RRRR(RRRR((s/p/zhu/06/nlp/nltk/nltk/tree.pyRscCstS(N(tImmutableProbabilisticTree(R((s/p/zhu/06/nlp/nltk/nltk/tree.pyRkscCsdti||ifS(Ns %s (p=%s)(RRtprob(R((s/p/zhu/06/nlp/nltk/nltk/tree.pyRscCs d|idd|ifS(Ns %s (p=%s)Ri<(RR(R((s/p/zhu/06/nlp/nltk/nltk/tree.pyRscCs@ti||}|djo|Snt|i|iS(Ni(Rt__cmp__tcmpR(RRR((s/p/zhu/06/nlp/nltk/nltk/tree.pyRscCsAt|tptSnti||o|i|ijS(N(R RRRR(RR((s/p/zhu/06/nlp/nltk/nltk/tree.pyRscCs ||j S(N((RR((s/p/zhu/06/nlp/nltk/nltk/tree.pyRscCs>|p#|i|i|d|iSn|ii|SdS(NR(RgRRRd(RRh((s/p/zhu/06/nlp/nltk/nltk/tree.pyRis*cCst|totg}|D]}||i|q~}t|to ||i|d|iSq||i|ddSn|SdS(NRg?(R RRdRRR(RReRfR/R((s/p/zhu/06/nlp/nltk/nltk/tree.pyRds * (R RRRRRkRRRRRRRiRdR(((s/p/zhu/06/nlp/nltk/nltk/tree.pyRs          RcBsweZedZedZdZdZdZdZdZ dZ e dZ d Z ee Z RS( cKstt|i|||S(N(RRR(RRRR((s/p/zhu/06/nlp/nltk/nltk/tree.pyRscKs<|djodSnti|||ti||dS(N(RRjRR(RRRR((s/p/zhu/06/nlp/nltk/nltk/tree.pyRscCstS(N(R(R((s/p/zhu/06/nlp/nltk/nltk/tree.pyRkscCsdti||ifS(Ns%s [%s](RRR(R((s/p/zhu/06/nlp/nltk/nltk/tree.pyRscCs d|idd|ifS(Ns%s [%s]Ri<(RR(R((s/p/zhu/06/nlp/nltk/nltk/tree.pyRscCs@ti||}|djo|Snt|i|iS(Ni(RRRR(RRR((s/p/zhu/06/nlp/nltk/nltk/tree.pyRscCsAt|tptSnti||o|i|ijS(N(R RRRR(RR((s/p/zhu/06/nlp/nltk/nltk/tree.pyRscCs ||j S(N((RR((s/p/zhu/06/nlp/nltk/nltk/tree.pyRscCs>|p#|i|i|d|iSn|ii|SdS(NR(RgRRRd(RRh((s/p/zhu/06/nlp/nltk/nltk/tree.pyRis*cCst|totg}|D]}||i|q~}t|to ||i|d|iSq||i|ddSn|SdS(NRi(R RRdRRR(RReRfR/R((s/p/zhu/06/nlp/nltk/nltk/tree.pyRds * (R RRRRRkRRRRRRRiRdR(((s/p/zhu/06/nlp/nltk/nltk/tree.pyRs          cCsRg}xE|D]=}t|to|it|iq |i|q W|S(N(R RR.RR(RtnamesR/((s/p/zhu/06/nlp/nltk/nltk/tree.pyRC scCsti|dtS(s$ Parse a treebank string and return a tree. Trees are represented as nested brackettings, e.g. (S (NP (NNP John)) (VP (V runs))). @return: A tree corresponding to the string representation. @rtype: C{tree} @param s: The string to be converted @type s: C{string} R(RR R(R{((s/p/zhu/06/nlp/nltk/nltk/tree.pyt bracket_parses cCstid|}xtt|D]}||djo)||||d||d<||} corpus, and shows the results of calling several of their methods. i(RsA(S (NP (DT the) (NN cat)) (VP (VBD ate) (NP (DT a) (NN cookie))))s#Convert bracketed string into tree:sDisplay tree properties:iis(JJ big)sTree modification:s (NN cake)sCollapse unary:sChomsky normal form:txtytzRg?sProbabilistic Tree:s0Convert tree to bracketed string and back again:s LaTeX output:sProduction output:ttestiN(ii(iii(iii(stesti(tnltkRRRRR2R-RRR`RURRRRD(RR{tttthe_cattpt((s/p/zhu/06/nlp/nltk/nltk/tree.pytdemoKsZ               t__main__RRR(RRrRtgrammarRRt probabilityRtutilRRRRjRRRRRRRRCRRRR t__all__(((s/p/zhu/06/nlp/nltk/nltk/tree.pys s4  ~4q&'  $ G