#include "basics.h"\r
-//#include <cstring>\r
-#include <stack>\r
#include "XMLTree.h"\r
#include "timings.h"\r
+#include <errno.h>\r
+using std::cout;\r
+using std::endl;\r
+using std::min;\r
+using std::string;\r
\r
// functions to convert tag positions to the corresponding tree node and viceversa. \r
// These are implemented in order to be able to change the tree and Tags representations, \r
// Current implementation corresponds to balanced-parentheses representation for\r
// the tree, and storing 2 tags per tree node (opening and closing tags).\r
\r
-// tag position -> tree node\r
-static treeNode tagpos2node(int t) \r
- {\r
- return (treeNode) t;\r
- }\r
\r
static int bits8 (int t ) {\r
int r = bits(t);\r
return r;\r
}\r
\r
-// tree node -> tag position\r
-static int node2tagpos(treeNode x) \r
-{\r
- return (int)x;\r
-}\r
-\r
-static int fast_find_close(bp *b,int s)\r
-{\r
- return fwd_excess(b,s,-1);\r
-}\r
-\r
-static int fast_inspect(bp* Par,treeNode i)\r
-{\r
- int j,l;\r
- j = i >> logD;\r
- l = i & (D-1);\r
- return (Par->B[j] >> (D-1-l)) & 1;\r
-}\r
-\r
-static treeNode fast_first_child(bp *Par, treeNode x)\r
-{\r
- x = x+1;\r
- return (fast_inspect(Par,x) == OP) ? x : NULLT;\r
-}\r
-\r
-static treeNode fast_next_sibling(bp* Par,treeNode x)\r
-{\r
- x = fwd_excess(Par,x,0);\r
- return (fast_inspect(Par,x) == OP) ? x : NULLT;\r
-}\r
\r
\r
static treeNode fast_sibling(bp* Par,treeNode x,TagType tag){\r
\r
}\r
\r
-static bool fast_isleaf(bp* Par,treeNode x){\r
- return (fast_inspect(Par,x+1) == CP ? true : false);\r
-}\r
+\r
\r
\r
inline uint get_field_no_power(uint *A, uint len, uint index) {\r
\r
}\r
\r
-inline bool fast_is_ancestor(bp * Par,treeNode x,treeNode y){\r
- if (x > y) \r
- return false;\r
- else\r
- return (x==0) || (y <= fast_find_close(Par,x));\r
-}\r
+\r
\r
\r
XMLTree::XMLTree( pb * const par, uint npar, vector<string> * const TN, TagIdMap * const tim, \r
uint *empty_texts_bmp, TagType *tags,\r
TextCollection * const TC, bool dis_tc)\r
{\r
+ buffer = 0;\r
// creates the data structure for the tree topology\r
Par = (bp *)umalloc(sizeof(bp));\r
STARTTIMER();\r
}\r
\r
// Save: saves XML tree data structure to file. \r
-void XMLTree::Save(int fd) \r
+void XMLTree::Save(int fd, char *filename) \r
{\r
FILE *fp;\r
char filenameaux[1024];\r
\r
// stores the texts \r
if (!disable_tc) {\r
- Text->Save(fp);\r
+ Text->Save(fp, filename);\r
};\r
-\r
-\r
}\r
\r
\r
// Load: loads XML tree data structure from file. Returns\r
// a pointer to the loaded data structure\r
-XMLTree *XMLTree::Load(int fd, bool load_tc,int sample_factor) \r
+XMLTree *XMLTree::Load(int fd, char *filename, bool load_tc,int sample_factor) \r
{\r
+\r
FILE *fp;\r
char buffer[1024];\r
XMLTree *XML_Tree;\r
int i;\r
\r
-\r
+ buffer[1023] = '\0';\r
\r
fp = fdopen(fd, "r");\r
\r
PRINTTIME("Loading parenthesis struct", Loading);\r
STARTTIMER();\r
\r
- XML_Tree->TagName = new vector<string>();\r
- XML_Tree->tIdMap = new std::unordered_map<string,int>();\r
- \r
- string s;\r
+ XML_Tree->TagName = new std::vector<std::string>();\r
+ XML_Tree->tIdMap = new std::unordered_map<std::string,int>();\r
+ std::string s;\r
int ntags;\r
\r
// Load the tag names\r
ufread(&ntags, sizeof(int), 1, fp);\r
\r
for (i=0; i<ntags;i++) {\r
- char * r = fgets(buffer,1023,fp);\r
- if (r==NULL)\r
+ if (fgets(buffer,1022,fp) != buffer)\r
throw "Cannot read tag list";\r
- s = (const char*) buffer;\r
+ s = buffer;\r
// remove the trailing \n\r
s.erase(s.size()-1); \r
- XML_Tree->TagName->push_back(s);\r
+ XML_Tree->TagName->push_back(s); \r
XML_Tree->tIdMap->insert(std::make_pair(s,i));\r
\r
};\r
// Not used \r
// loads the texts\r
if (!XML_Tree->disable_tc){\r
- XML_Tree->Text = TextCollection::Load(fp,sample_factor);\r
+ XML_Tree->Text = TextCollection::Load(fp, filename, TextCollection::index_mode_default, sample_factor);\r
}\r
else XML_Tree->Text = NULL;\r
STOPTIMER(Loading);\r
\r
\r
// SubtreeSize(x): the number of nodes (and attributes) in the subtree of node x.\r
-int XMLTree::SubtreeSize(treeNode x) \r
+/*int XMLTree::SubtreeSize(treeNode x) \r
{\r
return subtree_size(Par, x);\r
}\r
-\r
+*/\r
// SubtreeTags(x,tag): the number of occurrences of tag within the subtree of node x.\r
+/*\r
int XMLTree::SubtreeTags(treeNode x, TagType tag) \r
{\r
if (x == Root())\r
\r
int s = x + 2*subtree_size(Par, x) - 1;\r
\r
- return Tags->rank(tag, s) - Tags->rank(tag, node2tagpos(x)-1);\r
+ return (Tags->rank(tag, s) - Tags->rank(tag, node2tagpos(x)-1))+1;\r
}\r
+*/\r
int XMLTree::SubtreeElements(treeNode x) \r
{\r
\r
}\r
\r
// IsFirstChild(x): returns whether node x is the first child of its parent.\r
-bool XMLTree::IsFirstChild(treeNode x)\r
+/*bool XMLTree::IsFirstChild(treeNode x)\r
{\r
return ((x != NULLT)&&(x==Root() || prev_sibling(Par,x) == (treeNode)-1));\r
}\r
-\r
+*/\r
\r
// NumChildren(x): number of children of node x. Constant time with the data structure\r
// of Sadakane.\r
}\r
\r
// Parent(x): returns the parent node of node x.\r
-\r
+/*\r
treeNode XMLTree::Parent(treeNode x) \r
{\r
if (x == Root())\r
return NULLT;\r
else\r
return parent(Par, x);;\r
- }\r
+ }*/\r
\r
// Child(x,i): returns the i-th child of node x, assuming it exists.\r
treeNode XMLTree::Child(treeNode x, int i) \r
}\r
\r
// FirstChild(x): returns the first child of node x, assuming it exists. Very fast in BP.\r
-\r
+/*\r
treeNode XMLTree::FirstChild(treeNode x) \r
{\r
NULLT_IF(x==NULLT);\r
return fast_first_child(Par, x);\r
}\r
-\r
+*/\r
+/*\r
treeNode XMLTree::FirstElement(treeNode x) \r
{\r
NULLT_IF(x==NULLT);\r
return x;\r
}\r
}\r
-\r
+*//*\r
treeNode XMLTree::NextElement(treeNode x) \r
{\r
NULLT_IF(x==NULLT);\r
return (fast_inspect(Par,x)==OP)? x : NULLT;\r
}\r
else return x; \r
-}\r
-value XMLTree::CamlFirstElement(value x)\r
-{\r
- return Val_int(FirstElement(Int_val(x)));\r
-}\r
-value XMLTree::CamlNextElement(value x)\r
-{\r
- return Val_int(NextElement(Int_val(x)));\r
-}\r
-\r
-extern "C" value caml_cpp_fast_first_element(value xmltree, value node){\r
- return XMLTREE(xmltree)->CamlFirstElement(node);\r
-}\r
-\r
-extern "C" value caml_cpp_fast_next_element(value xmltree, value node){\r
- return XMLTREE(xmltree)->CamlNextElement(node);\r
-}\r
+ }*/\r
\r
// LastChild(x): returns the last child of node x.\r
treeNode XMLTree::LastChild(treeNode x)\r
}\r
\r
// NextSibling(x): returns the next sibling of node x, assuming it exists.\r
-treeNode XMLTree::NextSibling(treeNode x) \r
+/*treeNode XMLTree::NextSibling(treeNode x) \r
{\r
NULLT_IF(x==NULLT || x == Root() );\r
x = fast_find_close(Par,x)+1;\r
return (fast_inspect(Par,x) == CP ? NULLT : x);\r
}\r
-\r
+*/\r
\r
// PrevSibling(x): returns the previous sibling of node x, assuming it exists.\r
treeNode XMLTree::PrevSibling(treeNode x) \r
// TaggedChild(x,tag): returns the first child of node x tagged tag, or NULLT if there is none.\r
// Because of the balanced-parentheses representation of the tree, this operation is not supported\r
// efficiently, just iterating among the children of node x until finding the desired child.\r
+/*\r
treeNode XMLTree::TaggedChild(treeNode x, TagType tag) \r
{\r
\r
}\r
return NULLT; // no such sibling was found \r
}\r
-\r
+*/\r
treeNode XMLTree::SelectChild(treeNode x, TagIdSet *tags)\r
{\r
\r
\r
// TaggedDescendant(x,tag): returns the first node tagged tag with larger preorder than x and within\r
// the subtree of x. Returns NULLT if there is none.\r
+/*\r
treeNode XMLTree::TaggedDescendant(treeNode x, TagType tag) \r
{\r
//NULLT_IF(x==NULLT || fast_isleaf(Par,x));\r
\r
return (fast_is_ancestor(Par,x,y) ? y : NULLT);\r
}\r
-\r
-\r
+*/\r
+/*\r
treeNode XMLTree::SelectDescendant(treeNode x, TagIdSet *tags)\r
{\r
NULLT_IF (x ==NULLT || fast_isleaf(Par,x));\r
return min;\r
}\r
\r
-\r
+*/\r
\r
// TaggedPrec(x,tag): returns the first node tagged tag with smaller preorder than x and not an\r
// ancestor of x. Returns NULLT if there is none.\r
\r
// TaggedFollBelow(x,tag,root): returns the first node tagged tag with larger preorder than x \r
// and not in the subtree of x. Returns NULLT if there is none.\r
+/*\r
treeNode XMLTree::TaggedFollowingBelow(treeNode x, TagType tag, treeNode ancestor)\r
{\r
// NULLT_IF (x == NULLT || x == Root() || x == ancestor); \r
\r
//Special optimisation, test for the following sibling first\r
treeNode close = fast_find_close(Par, x);\r
- /*\r
- treeNode ns = close+1;\r
- if (fast_inspect(Par,ns) == OP) {\r
- TagType tagns = Tag(ns);\r
- // cout << GetTagNameByRef(tagns) << endl;\r
- //cout.flush();\r
- if (tagns == PCDATA_TAG_ID){\r
- close = ns+1;\r
- ns = ns+2;\r
- if (fast_inspect(Par,ns) != OP)\r
- goto after;\r
- tagns = Tag(ns); \r
- };\r
- if (tagns == tag)\r
- return ns;\r
- };\r
- after:\r
- */\r
treeNode s = tagpos2node(Tags->select_next(tag, close));\r
\r
if (ancestor == Root() || s==NULLT || s < fast_find_close(Par,ancestor)) return s;\r
else return NULLT;\r
} \r
-\r
+*/\r
+ /*\r
treeNode XMLTree::TaggedFollowingBefore(treeNode x, TagType tag, treeNode closing)\r
{\r
\r
\r
return s;\r
} \r
-\r
+ */\r
/* Here we inline TaggedFoll to find the min globally, and only at the end\r
we check if the min is below the context node */\r
treeNode XMLTree::SelectFollowingBelow(treeNode x, TagIdSet *tags, treeNode ancestor)\r
\r
\r
TagIdSet::const_iterator tagit;\r
- for (tagit = tags->begin(); tagit != tags->end(); tagit++) {\r
+ for (tagit = tags->begin(); tagit != tags->end(); ++tagit) {\r
\r
aux = tagpos2node(Tags->select_next(*tagit, close));\r
- \r
- // The next sibling of x is guaranteed to be below ctx\r
- // and is the node with lowest preorder which is after ctx.\r
- // if we find it, we return early; \r
if (aux == NULLT) continue;\r
if ((min == NULLT) || (aux < min)) min = aux;\r
};\r
// found the smallest node in preorder which is after x.\r
// if ctx is the root node, just return what we found.\r
\r
- if (ancestor == Root()) return min;\r
- // else check whether if is in below the ctx node\r
-\r
- NULLT_IF (min == NULLT || min >= fast_find_close(Par, ancestor));\r
+ if (ancestor == Root() || min == NULLT || min < fast_find_close(Par, ancestor)) return min;\r
+ else return NULLT;\r
\r
- return min;\r
+ }\r
+/*\r
+treeNode XMLTree::SelectFollowingBefore(treeNode x, TagIdSet *tags, treeNode ancestor_closing)\r
+ {\r
+\r
+ NULLT_IF(x==NULLT || x==Root());\r
+\r
+ treeNode close = fast_find_close(Par,x);\r
+ treeNode ns = close+1;\r
+ if ( (fast_inspect(Par,ns) == OP) && (tags->find(Tag(ns)) != tags->end()))\r
+ return ns;\r
+\r
+ int i;\r
+ treeNode min = NULLT;\r
+ treeNode aux;\r
+ \r
+\r
+ TagIdSet::const_iterator tagit;\r
+ for (tagit = tags->begin(); tagit != tags->end(); ++tagit) {\r
+\r
+ aux = tagpos2node(Tags->select_next(*tagit, close));\r
+ if (aux == NULLT) continue;\r
+ if ((min == NULLT) || (aux < min)) min = aux;\r
+ };\r
+ \r
+ // found the smallest node in preorder which is after x.\r
+ // if ctx is the root node, just return what we found.\r
+\r
+ if (ancestor_closing == Root() || min == NULLT || min < ancestor_closing) return min;\r
+ else return NULLT;\r
\r
}\r
+*/\r
+/*\r
treeNode XMLTree::SelectFollowingBefore(treeNode x, TagIdSet *tags, treeNode closing)\r
{\r
\r
return min;\r
\r
}\r
-\r
+*/\r
\r
// TaggedAncestor(x, tag): returns the closest ancestor of x tagged tag. Return\r
// NULLT is there is none.\r
\r
//WARNING this uses directly the underlying implementation for plain text\r
\r
-void XMLTree::Print(int fd,treeNode x){\r
- \r
- int newfd = dup(fd);\r
- stream = fdopen(newfd,"wa");\r
- /* if (stream_fd != fd){\r
- if (stream != NULL)\r
- fclose(stream);\r
- int newfd = dup(fd);\r
- stream = fdopen(newfd,"wa");\r
- stream_fd = fd;\r
- };\r
- */\r
\r
- FILE* fp = stream;\r
+void XMLTree::Print(int fd,treeNode x, bool no_text){\r
+ \r
+ if (buffer == 0)\r
+ buffer = new string();\r
+ \r
treeNode fin = fast_find_close(Par,x);\r
treeNode n = x;\r
TagType tag = Tag(n);\r
uchar * tagstr;\r
range r = DocIds(x);\r
treeNode first_idx;\r
- treeNode first_text = (tag == PCDATA_TAG_ID ? x : TaggedDescendant(x,PCDATA_TAG_ID));\r
- treeNode first_att = NULLT;//TaggedDesc(x,ATTRIBUTE_DATA_TAG_ID);\r
+ treeNode first_text = (tag == PCDATA_TAG_ID ? x : ParentNode(r.min-1));\r
+ treeNode first_att = NULLT;\r
\r
if (first_att == NULLT)\r
- first_idx = first_text;\r
+ first_idx = first_text;\r
else if (first_text == NULLT)\r
- first_idx = first_att;\r
+ first_idx = first_att;\r
else\r
- first_idx = min(first_att,first_text);\r
+ first_idx = min(first_att,first_text);\r
\r
uchar * current_text=NULL;\r
- if (first_idx != NULLT)\r
- current_text = GetText(MyText(first_idx));\r
- int read = 0;\r
-\r
- std::stack<uchar*> st;\r
- while (n <= fin){\r
- if (fast_inspect(Par,n)){\r
- if (tag == PCDATA_TAG_ID) { \r
- // fputs((const char*) (GetText(MyTextUnsafe(n))),fp);\r
- \r
- read = fprintf(fp,"%s",(const char*) current_text);\r
- current_text += (read + 1);\r
\r
- n+=2; // skip closing $\r
- tag = Tag(n);\r
- }\r
- else {\r
-\r
- fputc('<',fp);\r
- tagstr = (uchar*) GetTagNameByRef(tag);\r
- fputs((const char*) tagstr ,fp);\r
- n++;\r
- if (fast_inspect(Par,n)) {\r
- st.push(tagstr);\r
- tag = Tag(n);\r
- if (tag == ATTRIBUTE_TAG_ID){\r
- n++;\r
- while (fast_inspect(Par,n)){\r
- fputc(' ',fp);\r
- fputs((const char*) &(GetTagNameByRef(Tag(n))[3]),fp);\r
- fputs("=\"",fp);\r
- n++;\r
- read = fprintf(fp,"%s",(const char*) current_text);\r
- current_text += (read + 1);\r
- //fputs((const char*) GetText(MyTextUnsafe(n)),fp);\r
- fputc('"',fp);\r
- n+=3; //close @$ @@ \r
- };\r
- fputc('>',fp);\r
- n++;\r
- tag=Tag(n);\r
- }\r
+ if (first_idx != NULLT)\r
+ current_text = GetText(MyText(first_idx));\r
+ \r
+ size_t read = 0;\r
+ std::vector<uchar*> st;\r
+ while (n <= fin){\r
+ if (fast_inspect(Par,n)){\r
+ if (tag == PCDATA_TAG_ID) { \r
+ \r
+ if (no_text)\r
+ _dputs("<$/>", fd);\r
else {\r
- fputc('>',fp);\r
+ read = _dprintf((const char*) current_text, fd);\r
+ current_text += (read + 1);\r
};\r
+ n+=2; // skip closing $\r
+ tag = Tag(n);\r
+ \r
}\r
- else {// <foo /> tag\r
- fputs("/>",fp);\r
+ else {\r
+ _dputc('<',fd);\r
+ tagstr = (uchar*) GetTagNameByRef(tag);\r
+ _dputs((const char*) tagstr, fd);\r
n++;\r
- tag=Tag(n); \r
- }; \r
- };\r
- }\r
- else\r
- do {\r
- fputs("</",fp);\r
- fputs((const char*)st.top(),fp);\r
- fputc('>', fp);\r
- st.pop();\r
- n++;\r
- }while (!fast_inspect(Par,n) && !st.empty());\r
- tag=Tag(n);\r
- };\r
- fputc('\n',fp);\r
- fflush(fp);\r
- fclose(fp);\r
+ if (fast_inspect(Par,n)) {\r
+ st.push_back(tagstr);\r
+ tag = Tag(n);\r
+ if (tag == ATTRIBUTE_TAG_ID){\r
+ n++;\r
+ if (no_text) _dputs("><@@>",fd);\r
+\r
+ while (fast_inspect(Par,n)){\r
+ if (no_text) {\r
+ _dputc('<', fd);\r
+ _dputs((const char*) &(GetTagNameByRef(Tag(n))[3]), fd);\r
+ _dputc('>', fd);\r
+ _dputs("<$@/></", fd);\r
+ _dputs((const char*) &(GetTagNameByRef(Tag(n))[3]), fd);\r
+ _dputc('>', fd);\r
+ n+= 4;\r
+ } else {\r
+ _dputc(' ', fd);\r
+ _dputs((const char*) &(GetTagNameByRef(Tag(n))[3]), fd);\r
+ n++;\r
+ _dputs("=\"", fd);\r
+ read = _dprintf((const char*) current_text, fd);\r
+ current_text += (read + 1);\r
+ _dputc('"', fd);\r
+ n+=3;\r
+ }\r
+ };\r
+ if (no_text) _dputs("</@@>", fd);\r
+ else _dputc('>', fd);\r
+ n++;\r
+ tag=Tag(n);\r
+ \r
+ } else \r
+ _dputc('>', fd);\r
+ \r
+ } else {// <foo /> tag\r
+ _dputs("/>", fd);\r
+ n++;\r
+ tag=Tag(n); \r
+ }; \r
+ };\r
+ }\r
+ else\r
+ do {\r
+ _dputs("</", fd);\r
+ _dputs((const char*)st.back(), fd);\r
+ _dputc('>', fd);\r
+ st.pop_back();\r
+ n++;\r
+ } while (!(fast_inspect(Par,n) || st.empty()));\r
+ tag = Tag(n);\r
+ };\r
+ _dputc('\n', fd);\r
+ _flush(fd);\r
}\r
projects / SXSI / XMLTree.git / blobdiff