//------------------------------------------------------------------------------
// Make a sibling of CurNode and make CurNode the new node.
void Tree::MakeSibling ()
{
NodePtr q = NewNode ();
NodePtr ancestor = CurNode->GetAnc();
CurNode->SetSibling(q);
q->SetAnc(ancestor);
ancestor->AddWeight(CurNode->GetWeight());
ancestor->IncrementDegree();
CurNode = q;
}
//------------------------------------------------------------------------------
// Uses a simple pushdown automaton to read Newick-style trees
bool TreeReader::Read (TreePtr t)
{
// States of pushdown automaton that reads trees
enum statetype
{
GETNAME,
GETINTERNODE,
NEXTMOVE,
DOSIBLING,
FINISHCHILDREN,
ACCEPTED,
CLEANUP,
QUIT
} state;
std::stack< NodePtr, std::vector<NodePtr> > stk;
Tokeniser::tokentype token;
tree = t;
tree->MakeRoot();
token = parser.GetNextToken ();
if (token == Tokeniser::EMPTY)
return false;
// Parse the tree description
state = GETNAME;
while ((state != QUIT) && (state != ACCEPTED))
{
switch (state)
{
case GETNAME:
switch (token)
{
case Tokeniser::STRING:
case Tokeniser::NUMBER:
LabelLeaf (parser.GetToken());
token = parser.GetNextToken ();
state = GETINTERNODE;
break;
case Tokeniser::LPAR:
state = NEXTMOVE;
break;
default:
errormsg = "Syntax error [GETNAME]: expecting a \"(\" or leaf name, got \"";
errormsg += parser.GetToken();
errormsg += "\" instead";
state = QUIT;
break;
}
break;
case GETINTERNODE:
switch (token)
{
case Tokeniser::COLON:
case Tokeniser::COMMA:
case Tokeniser::RPAR:
state = NEXTMOVE;
break;
default:
errormsg = "Syntax error [GETINTERNODE]: expecting one of \":,)\", got ";
errormsg += parser.GetToken();
errormsg += " instead";
state = QUIT;
break;
}
break;
case NEXTMOVE:
switch (token)
{
case Tokeniser::COLON:
if (LabelEdge ())
token = parser.GetNextToken ();
else
state = QUIT;
break;
// The next node encountered will be a sibling
// of Curnode and a descendant of the node on
// the top of the node stack.
case Tokeniser::COMMA:
if (stk.empty())
{
errormsg = "Tree description unbalanced, this \")\" has no matching \"(\"";
state = QUIT;
}
else
{
tree->MakeSibling ();
//token = parser.GetNextToken ();
token = GetTaxonName();
state = GETNAME;
}
break;
// The next node will be a child of CurNode, hence
// we create the node and push CurNode onto the
// node stack.
case Tokeniser::LPAR:
stk.push (tree->GetCurNode());
tree->MakeChild();
//token = parser.GetNextToken ();
token = GetTaxonName();
state = GETNAME;
break;
// We've finished ready the descendants of the node
// at the top of the node stack so pop it off.
case Tokeniser::RPAR:
if (stk.empty())
{
errormsg = "Tree description unbalanced (an extra \")\")";
state = QUIT;
}
else
{
NodePtr q = stk.top();
q->AddWeight(tree->GetCurNode()->GetWeight());
tree->SetCurNode (q);
stk.pop ();
token = parser.GetNextToken ();
state = FINISHCHILDREN;
}
break;
// We should have finished the tree
case Tokeniser::SEMICOLON:
if (stk.empty())
{
state = ACCEPTED;
}
else
{
errormsg = "Tree description ended prematurely (stack not empty)";
state = QUIT;
}
break;
default:
errormsg = "Syntax error [NEXTMOVE]: expecting one of \":,();\", got ";
errormsg += parser.GetToken();
errormsg += " instead";
state = QUIT;
break;
}
break;
case FINISHCHILDREN:
switch (token)
{
case Tokeniser::STRING:
case Tokeniser::NUMBER:
LabelInternalNode (parser.GetToken());
token = parser.GetNextToken ();
break;
case Tokeniser::COLON:
if (LabelEdge ())
token = parser.GetNextToken ();
else
state = QUIT;
break;
// We've completed traversing the descendants of the
// node at the top of the stack, so pop it off.
case Tokeniser::RPAR:
if (stk.empty())
{
errormsg = "Tree description unbalanced, this \")\" has no matching \"(\"";
state = QUIT;
}
else
{
NodePtr q = stk.top();
q->AddWeight(tree->GetCurNode()->GetWeight());
tree->SetCurNode (q);
stk.pop ();
token = parser.GetNextToken ();
}
break;
// The node at the top of the stack still has some
// descendants.
case Tokeniser::COMMA:
if (stk.empty())
{
errormsg = "Tree description unbalanced, missing a \"(\"";
state = QUIT;
}
else
{
tree->MakeSibling ();
//token = parser.GetNextToken ();
token = GetTaxonName();
state = GETNAME;
}
break;
case Tokeniser::SEMICOLON:
state = NEXTMOVE;
break;
default:
if (stk.empty())
{
errormsg = "Tree description unbalanced";
state = QUIT;
}
else
{
errormsg = "Syntax error [FINISHCHILDREN]: expecting one of \":,();\" or internal label, got ";
errormsg += parser.GetToken();
errormsg += " instead";
}
state = QUIT;
break;
}
break;
}
}
// Handle errors
if (state == QUIT)
{
// Clean memory here...
// ... then throw exception
throw XTokeniser (errormsg, parser.GetFilePosition(),
parser.GetFileLine (), parser.GetFileColumn());
}
else
{
tree->GetRoot()->SetWeight(tree->GetNumLeaves());
doAdjust ();
}
return true;
}
//------------------------------------------------------------------------------
int Tree::Parse (const char *TreeDescr)
{
enum {stGETNAME, stGETINTERNODE, stNEXTMOVE,
stFINISHCHILDREN, stQUIT, stACCEPTED} state;
stack< NodePtr, vector<NodePtr> > stk;
NodePtr q;
Parser p ((char *)TreeDescr);
tokentype token;
float f;
// Initialise tree variables
Root = NULL;
Leaves = 0;
Internals = 0;
Error = 0;
// Create first node
CurNode = NewNode();
Root = CurNode;
// Initialise FSA that reads tree
state = stGETNAME;
//std::cout<<endl<<"400 Current token is "<<p.GetTokenAsCstr()<<endl; //added by BCO
token = p.NextToken ();
while ((state != stQUIT) && (state != stACCEPTED))
{
switch (state)
{
case stGETNAME:
switch (token)
{
case SPACE:
case TAB:
case NEWLINE:
//std::cout<<endl<<"413 Current token is "<<p.GetTokenAsCstr()<<endl; //added by BCO
token = p.NextToken ();
break;
case STRING:
// to do: handle translation
Leaves++;
CurNode->SetLeaf (true);
CurNode->SetLeafNumber (Leaves);
CurNode->SetWeight (1);
//std::cout<<endl<<"425 Current token is "<<p.GetTokenAsCstr()<<endl; //added by BCO
CurNode->SetLabel (p.GetToken());
CurNode->SetDegree (0);
//CurNode->SetModelCategory(vector<double>(1,1.0)); //Added by BCO
//std::cout<<endl<<"425 Current token is "<<p.GetTokenAsCstr()<<endl; //added by BCO
token = p.NextToken ();
state = stGETINTERNODE;
break;
case NUMBER:
// to do: handle translation
Leaves++;
CurNode->SetLeaf (true);
CurNode->SetLeafNumber (Leaves);
CurNode->SetWeight (1);
CurNode->SetLabel (p.GetToken());
CurNode->SetDegree (0);
// CurNode->SetModelCategory(vector<double>(1,1.0)); //Added by BCO
//std::cout<<endl<<"439 Current token is "<<p.GetTokenAsCstr()<<endl; //added by BCO
token = p.NextToken ();
state = stGETINTERNODE;
break;
case LPAR:
state = stNEXTMOVE;
break;
case ENDOFSTRING:
Error = errENDOFSTRING;
state = stQUIT;
break;
default:
Error = errSYNTAX;
std::cerr<<"Had error in Parse() in TreeLib.cpp line 447"; //added by BCO
state = stQUIT;
break;
}
break;
case stGETINTERNODE:
switch (token)
{
case SPACE:
case TAB:
case NEWLINE:
//std::cout<<endl<<"464 Current token is "<<p.GetTokenAsCstr()<<endl; //added by BCO
token = p.NextToken ();
break;
case COLON:
case COMMA:
case RPAR:
state = stNEXTMOVE;
break;
case ENDOFSTRING:
Error = errENDOFSTRING;
state = stQUIT;
break;
default:
Error = errSYNTAX;
std::cerr<<"Had error in Parse() in TreeLib.cpp line 472"; //added by BCO
state = stQUIT;
break;
}
break;
case stNEXTMOVE:
switch (token)
{
case COLON: //BCO, have it here deal separately with the next token being a number (use default behavior of treating it as a brlen) and having it a left curly bracket, indicating that simmap output is following.
//std::cout<<endl<<"488 Current token is "<<p.GetTokenAsCstr()<<endl; //added by BCO
token = p.NextToken ();
char matchchar;
matchchar=p.GetTokenAsCstr()[0];
if (matchchar=='{') { //added by BCO
// wrote everything in this if statement
//Note that the simmap parser only deals with "morphological" data -- states should be integers, in other words. //BCO
//State order goes from rootmost end of the branch to the tipmost end //BCO
//StateTimes[i] is amount of time in StateOrder[i] //BCO
vector<double> modelvector(maxModelCategoryStates,0.0); //maxModelCategoryStates is in TreeLib.h
vector<int> stateordervector; //Added by BCO
vector<double> statetimesvector; //Added by BCO
double totaledgelength=0;
char internalmatchchar;
internalmatchchar=p.GetTokenAsCstr()[0];
while (internalmatchchar!='}') {
//std::cout<<endl<<"502 Current token is "<<p.GetTokenAsCstr()<<endl; //added by BCO
token=p.NextToken();
char colonmatchchar;
colonmatchchar=p.GetTokenAsCstr()[0];
if (colonmatchchar==':') {
//std::cout<<endl<<"506 Current token is "<<p.GetTokenAsCstr()<<endl; //added by BCO
token=p.NextToken(); //Get rid of colon
}
int state=atoi(p.GetTokenAsCstr());
if (state>=maxModelCategoryStates) {
Error=errTOOMANYMODELCATS;
break;
}
//std::cout<<endl<<"515 Current token is "<<p.GetTokenAsCstr()<<endl; //added by BCO
token=p.NextToken(); //get rid of comma
//std::cout<<endl<<"517 Current token is "<<p.GetTokenAsCstr()<<endl; //added by BCO
token=p.NextToken(); //This should have the branch length spent in the above state
double brleninstate = atof (p.GetTokenAsCstr());
modelvector[state]+=brleninstate;
stateordervector.push_back(state);
statetimesvector.push_back(brleninstate);
totaledgelength+=brleninstate;
//std::cout<<endl<<"523 Current token is "<<p.GetTokenAsCstr()<<endl; //added by BCO
token=p.NextToken(); //take in next token
internalmatchchar=p.GetTokenAsCstr()[0];
}
CurNode->SetEdgeLength (totaledgelength);
CurNode->SetModelCategory(modelvector); //Added by BCO
CurNode->SetStateOrder(stateordervector); //Added by BCO
CurNode->SetStateTimes(statetimesvector); //Added by BCO
EdgeLengths = true;
//std::cout<<endl<<"527 Current token is "<<p.GetTokenAsCstr()<<endl; //added by BCO
token = p.NextToken ();
break;
} //BCO
else { //BCO. Will work when not dealing with a simmap tree.
f = atof (p.GetTokenAsCstr());
CurNode->SetEdgeLength (f);
CurNode->SetModelCategory(vector<double>(1,f)); //Added by BCO
CurNode->SetStateOrder(vector<int>(1,0)); //Added by BCO
CurNode->SetStateTimes(vector<double>(1,f)); //Added by BCO
EdgeLengths = true;
//std::cout<<endl<<"536 Current token is "<<p.GetTokenAsCstr()<<endl; //added by BCO
token = p.NextToken ();
} //BCO
break;
case SPACE:
case TAB:
case NEWLINE:
//std::cout<<endl<<"543 Current token is "<<p.GetTokenAsCstr()<<endl; //added by BCO
token = p.NextToken ();
break;
// The next node encountered will be a sibling
// of Curnode and a descendant of the node on
// the top of the node stack.
case COMMA:
q = NewNode();
CurNode->SetSibling (q);
if (stk.empty())
{
Error = errMISSINGLPAR;
state = stQUIT;
}
else
{
q->SetAnc (stk.top());
stk.top()->AddWeight (CurNode->GetWeight());
stk.top()->IncrementDegree ();
CurNode = q;
state = stGETNAME;
//std::cout<<endl<<"564 Current token is "<<p.GetTokenAsCstr()<<endl; //added by BCO
token = p.NextToken ();
}
break;
// The next node will be a child of CurNode, hence
// we create the node and push CurNode onto the
// node stack.
case LPAR:
Internals++;
stk.push (CurNode);
q = NewNode();
CurNode->SetChild (q);
q->SetAnc (CurNode);
CurNode->IncrementDegree ();
CurNode = q;
//std::cout<<endl<<"579 Current token is "<<p.GetTokenAsCstr()<<endl; //added by BCO
token = p.NextToken ();
state = stGETNAME;
break;
// We've finished ready the descendants of the node
// at the top of the node stack so pop it off.
case RPAR:
if (stk.empty ())
{
Error = errUNBALANCED;
state = stQUIT;
}
else
{
q = stk.top();
q->AddWeight (CurNode->GetWeight());
CurNode = q;
stk.pop();
state = stFINISHCHILDREN;
//std::cout<<endl<<"598 Current token is "<<p.GetTokenAsCstr()<<endl; //added by BCO
token = p.NextToken ();
}
break;
// We should have finished the tree
case SEMICOLON:
if (stk.empty())
{
state = stACCEPTED;
}
else
{
Error = errSTACKNOTEMPTY;
state = stQUIT;
}
break;
case ENDOFSTRING:
Error = errENDOFSTRING;
state = stQUIT;
break;
default:
Error = errSYNTAX;
std::cerr<<"Had error in Parse() in TreeLib.cpp line 597"; //added by BCO
state = stQUIT;
break;
}
break;
case stFINISHCHILDREN:
switch (token)
{
case STRING:
case NUMBER:
// internal label
InternalLabels = true;
CurNode->SetLabel (p.GetToken());
//std::cout<<endl<<"634 Current token is "<<p.GetTokenAsCstr()<<endl; //added by BCO
token = p.NextToken ();
break;
case COLON: //BCO, have it here deal separately with the next token being a number (use default behavior of treating it as a brlen) and having it a left curly bracket, indicating that simmap output is following.
//std::cout<<endl<<"640 Current token is "<<p.GetTokenAsCstr()<<endl; //added by BCO
token = p.NextToken ();
char matchchar;
matchchar=p.GetTokenAsCstr()[0];
if (matchchar=='{') { //added by BCO
//BCO wrote everything in this if statement
//Note that the simmap parser only deals with discrete data -- states should be integers, in other words. //BCO
vector<double> modelvector(maxModelCategoryStates,0.0); //maxModelCategoryStates is in TreeLib.h
vector<int> stateordervector; //added by BCO
vector<double> statetimesvector; //added by BCO
double totaledgelength=0;
char internalmatchchar;
internalmatchchar=p.GetTokenAsCstr()[0];
while (internalmatchchar!='}') {
//std::cout<<endl<<"654 Current token is "<<p.GetTokenAsCstr()<<endl; //added by BCO
token=p.NextToken();
char colonmatchchar;
colonmatchchar=p.GetTokenAsCstr()[0];
if (colonmatchchar==':') {
//std::cout<<endl<<"659 Current token is "<<p.GetTokenAsCstr()<<endl; //added by BCO
token=p.NextToken(); //Get rid of colon
}
int state=atoi(p.GetTokenAsCstr());
if (state>=maxModelCategoryStates) {
Error=errTOOMANYMODELCATS;
break;
}
//std::cout<<endl<<"668 Current token is "<<p.GetTokenAsCstr()<<endl; //added by BCO
token=p.NextToken(); //get rid of comma
//std::cout<<endl<<"670 Current token is "<<p.GetTokenAsCstr()<<endl; //added by BCO
token=p.NextToken(); //This should have the branch length spent in the above state
double brleninstate = atof (p.GetTokenAsCstr());
modelvector[state]+=brleninstate;
stateordervector.push_back(state);
statetimesvector.push_back(brleninstate);
totaledgelength+=brleninstate;
//std::cout<<endl<<"675 Current token is "<<p.GetTokenAsCstr()<<endl; //added by BCO
token=p.NextToken(); //take in next token
internalmatchchar=p.GetTokenAsCstr()[0];
}
CurNode->SetEdgeLength (totaledgelength);
CurNode->SetModelCategory(modelvector); //Added by BCO
CurNode->SetStateOrder(stateordervector); //Added by BCO
CurNode->SetStateTimes(statetimesvector); //Added by BCO
EdgeLengths = true;
//std::cout<<endl<<"682 Current token is "<<p.GetTokenAsCstr()<<endl; //added by BCO
token = p.NextToken ();
break;
} //BCO
else { //BCO. Will work when not dealing with a simmap tree.
f = atof (p.GetTokenAsCstr());
CurNode->SetEdgeLength (f);
CurNode->SetModelCategory(vector<double>(1,f)); //Added by BCO
CurNode->SetStateOrder(vector<int>(1,0)); //Added by BCO
CurNode->SetStateTimes(vector<double>(1,f)); //Added by BCO
EdgeLengths = true;
//std::cout<<endl<<"691 Current token is "<<p.GetTokenAsCstr()<<endl; //added by BCO
token = p.NextToken ();
} //BCO
break;
case SPACE:
case TAB:
case NEWLINE:
//std::cout<<endl<<"653 Current token is "<<p.GetTokenAsCstr()<<endl; //added by BCO
token = p.NextToken ();
break;
// We've completed traversing the descendants of the
// node at the top of the stack, so pop it off.
case RPAR:
if (stk.empty())
{
Error = errUNBALANCED;
state = stQUIT;
}
else
{
q = stk.top();
q->AddWeight (CurNode->GetWeight());
CurNode = q;
stk.pop();
//std::cout<<endl<<"671 Current token is "<<p.GetTokenAsCstr()<<endl; //added by BCO
token = p.NextToken ();
}
break;
// The node at the top of the stack still has some
// descendants.
case COMMA:
q = NewNode();
CurNode->SetSibling (q);
if (stk.empty())
{
Error = errMISSINGLPAR;
state = stQUIT;
}
else
{
q->SetAnc (stk.top());
stk.top()->AddWeight (CurNode->GetWeight());
stk.top()->IncrementDegree ();
CurNode = q;
state = stGETNAME;
//std::cout<<endl<<"693 Current token is "<<p.GetTokenAsCstr()<<endl; //added by BCO
token = p.NextToken ();
}
break;
case SEMICOLON:
state = stNEXTMOVE;
break;
default:
if (stk.empty())
{
Error = errSEMICOLON;
}
else
{
Error = errSYNTAX;
std::cerr<<"Had error in Parse() in TreeLib.cpp line 679"; //added by BCO
}
state = stQUIT;
break;
}
break;
}
}
// Handle errors
if (state == stQUIT)
{
// Clean up to go here
return (p.GetPos());
}
else
{
Root->SetWeight(Leaves);
getPathLengths(Root); //added by BCO
return (0);
}
}
