Module:Clade/converter

 --require('strict') -- comment out until clade also uses noglobals

 localp={}

 localpargs=mw.getCurrentFrame():getParent().args

 --[[ =================== parser for conversion to clade structure =============================

  Function p.newickConverter()
  convert Newick strings to clade format
 		Usage: {{#invoke:Module:Sandbox/Jts1882/CladeN|newickConverter|newickstring={{{NEWICK_STRING}}} }}

  Function p.listConverter()
  convert wikitext-like lists to clade format
  use @ instead of * in wikitext to avoid processing
  	Usage: {{#invoke:Module:Clade/converter|listConverter|list={{{LIST_STRING}}} }}
 ]]
 functionp.cladeConverter(frame)
 ifframe.args['newickstring']orpargs['newick']orpargs['newickstring']then
 returnp.newickConverter(frame)
 elseifframe.args['list']orpargs['list']then
 returnp.listConverter(frame)
 end
 end

 --[[ =================== Newick to clade parser function =============================

 	Function of convert Newick strings to clade format

 	Usage: {{#invoke:Module:Sandbox/Jts1882/CladeN|newickConverter|newickstring={{{NEWICK_STRING}}} }}
 ]]
 functionp.newickConverter(frame)

 localnewickString=frame.args['newickstring']orpargs['newick']orpargs['newickstring']

 --if newickString == '{{{newickstring}}}' then return newickString end

 newickString=require('Module:Clade').processNewickString(newickString,"")-- "childNumber")


 -- show the Newick string
 localcladeString=''
 locallevelNumber=1-- for depth of iteration
 localchildNumber=1-- number of sister elements on node (always one for root)

 -- converted the newick string to the clade structure
 cladeString=cladeString..'{{clade'
 cladeString=cladeString..p.newickParseLevel(newickString,levelNumber,childNumber)
 cladeString=cladeString..'\r}}'

 localresultString=''
 localoption=mw.getCurrentFrame():getParent().args['option']or''
 ifoption=='tree'then
 --show the transcluded clade diagram
 resultString=cladeString
 else
 -- show the Newick string
 resultString='<div>Modified Newick string:'
 ..'<pre>'..newickString..'</pre>'
 -- show the converted clade structure
 resultString=resultString..'Output of clade template structure:'
 ..'<pre>'..cladeString..'</pre></div>'
 end
 --resultString = frame:expandTemplate{ title = 'clade', frame:preprocess(cladeString) }

 returnresultString
 end

 --[[ Parse one level of Newick string
  This function receives a Newick string, which has two components
  1. the right hand term is a clade label: |labelN=labelname
  2. the left hand term in parenthesis has common delimited child nodes, each of which can be
  i. a taxon name which just needs: |N=leafname 
  ii. a Newick string which needs further processing through reiteration
 ]]
 functionp.newickParseLevel(newickString,levelNumber,childNumber)


 localcladeString=""
 localindent=p.getIndent(levelNumber)
 --levelNumber=levelNumber+1

 localj=0
 localk=0
 j,k=string.find(newickString,'%(.*%)')-- find location of outer parenthesised term
 localinnerTerm=string.sub(newickString,j+1,k-1)-- select content in parenthesis
 localouterTerm=string.gsub(newickString,"%b()","")-- delete parenthetic term

 cladeString=cladeString..indent..'|label'..childNumber..'='..outerTerm
 cladeString=cladeString..indent..'|'..childNumber..'='..'{{clade'

 levelNumber=levelNumber+1
 indent=p.getIndent(levelNumber)

 -- protect commas in inner parentheses from split; temporarily replace commas between parentheses
 localinnerTerm2=string.gsub(innerTerm,"%b()",function(n)
 returnstring.gsub(n,",%s*","XXX")-- also strip spaces after commas here
 end)

 --local s = p.strsplit(innerTerm2, ",")
 locals=mw.text.split(innerTerm2,",")
 locali=1
 whiles[i]do
 localrestoredString=string.gsub(s[i],"XXX",",")-- convert back to commas

 localouterTerm=string.gsub(restoredString,"%b()","")
 ifstring.find(restoredString,'%(.*%)')then
 --cladeString = cladeString .. indent .. '|y' .. i .. '=' .. p.newickParseLevel(restoredString,levelNumber+1,i) 
 cladeString=cladeString..p.newickParseLevel(restoredString,levelNumber,i)
 else
 cladeString=cladeString..indent..'|'..i..'='..restoredString--.. '(level=' .. levelNumber .. ')'
 end
 i=i+1
 end
 -- end -- end splitting of strings

 cladeString=cladeString..indent..'}}'
 returncladeString
 end

 functionp.getIndent(levelNumber)
 localindent="\r"
 localextraIndent=pargs['indent']ormw.getCurrentFrame().args['indent']or0

 whiletonumber(extraIndent)>0do
 indent=indent.." "-- an extra indent to make aligining compound trees easier
 extraIndent=extraIndent-1
 end

 whilelevelNumber>1do
 indent=indent.." "
 levelNumber=levelNumber-1
 end
 returnindent
 end


 --[[ =================== experimental list to clade parser function =============================

  Function of convert wikitext-like listss to clade format
  - use @ instead of * in wikitext to avoid processing

  Usage: {{#invoke:Module:Clade/converter|listConverter|list={{{LIST_STRING}}} }}
 ]]

 functionp.listConverter(frame)

 locallistString=frame.args['list']ormw.getCurrentFrame():getParent().args['list']

 -- show the list string
 localcladeString=''
 locallevelNumber=1-- for depth of iteration
 localchildNumber=1-- number of sister elements on node (always one for root)
 localindent=p.getIndent(levelNumber)
 -- converted the newick string to the clade structure
 cladeString=cladeString..indent..'{{clade'
 cladeString=cladeString..p.listParseLevel(listString,levelNumber,childNumber)
 --cladeString = cladeString .. '\r}}' 

 localresultString=''
 localoption=mw.getCurrentFrame():getParent().args['option']or''
 ifoption=='tree'then
 --show the transcluded clade diagram
 resultString=cladeString
 else
 -- show the list string
 --resultString = '<pre>'..listString..'</pre>'	
 -- show the converted clade structure
 resultString=resultString..'<pre>'..cladeString..'</pre>'
 end
 --resultString = frame:expandTemplate{ title = 'clade', frame:preprocess(cladeString) }

 returnresultString
 end

 functionp.listParseLevel(listString,levelNumber,childNumber)

 localcladeString=""
 localindent=p.getIndent(levelNumber)
 levelNumber=levelNumber+1

 locallist=mw.text.split(listString,"\n")
 locali=1
 localchild=1
 locallastNode=0

 whilelist[i]do
 list[i]=list[i]:gsub("^@","")-- strip the first @

 ifnotstring.match(list[i],"^@",1)then-- count children at this level (not beginning wiht @)
 lastNode=lastNode+1
 end
 i=i+1
 end
 i=1

 whilelist[i]do

 --[[ pseudocode: 
 	 if next value begins with @ we have a subtree, 
 	 	which must be recombined and past iteratively
 	 else we have a simple leaf
 	 ]]

 -- if the next value begins with @, we have a subtree which should be recombined
 iflist[i+1]andstring.match(list[i+1],"^@",1)then

 locallabel=list[i]
 i=i+1
 localrecombined=list[i]
 whilelist[i+1]andstring.match(list[i+1],"^@",1)do
 recombined=recombined.."\n"..list[i+1]
 i=i+1
 end
 cladeString=cladeString..indent..'|label'..child..'='..label
 cladeString=cladeString..indent..'|'..child..'='..'{{clade'
 ..p.listParseLevel(recombined,levelNumber,i)
 else
 cladeString=cladeString..indent..'|'..child..'='..list[i]
 end
 i=i+1
 child=child+1
 end


 cladeString=cladeString..indent..'}}'
 returncladeString
 end
 returnp