User:Richiez

From Wikipedia, the free encyclopedia

My talk page

my test page

[edit] Script to do automatic Harvard style referencing

The task of formatting references manually is not much fun and prone to all kinds of errors and accidents.. how nice would it be to write {{pmid 123456}} inline and have the rest done automatically? Turns out that it is not so hard at all. A version using footnote referencing would be even easier but in many situations I prefer harvard referencing.

This little ruby script works on local files and converts {{pmid XXXX}} and {{isbn XXXXX}} to harvard style references with links. Appropriate cite templates are added in the References section.

This is an early test version – the output will definitely require manual postprocessing in most cases. Output goes to stdout.

Uses User:Diberri's tool internally.

I found no easy way to generate {{citation ...}} templates automatically so cite templates are emitted instead and an extra anchor is generated. The links to this anchor however are generated by the Harvcol templates automagically so under unusual conditions this might break.. especially if my code and the harvcol templates do not agree on the name for an anchor. Wikiref/wikicite would be another possibility but then I could not use Harv/cite templates amd would have to do the formatting manually.

Here is the source code – requires some computer knowledge to use.

#!/usr/bin/ruby
#
# walk wiki article and replace {{pmid XXXX}} and {{isbn XXXX}} with 
# harvard style references.
# pmid/isbn generates {{Harvcol | Lastname [...]| year}}
#   and {{Anchor ..}}{{cite ...}} in References section
#  -txt, -nb suffix generates Harcoltxt resp Harvcolnb variants
#    eg {{pmid-nb 123456}}
#   using http://diberri.dyndns.org/
 
require 'cgi'
require 'net/http'
 
Host = "diberri.dyndns.org"
 
def get_session()
  return $session if $session
  if ENV["http_proxy"] =~ /http:\/\/(.*):(\d*)/
    $proxy_addr=$1
    $proxy_port=$2
    $session=Net::HTTP.new(Host, 80, $proxy_addr, $proxy_port)
  else
    $session=Net::HTTP.new(Host, 80)
  end
  return $session
end
 
def lookup_isbn(style,isbn)
  path = "/cgi-bin/templatefiller/index.cgi?ddb=&type=isbn&id="+isbn.to_s
  lookup(style,path)
end
def lookup_pmid(style,pmid)
  path = "/cgi-bin/templatefiller/index.cgi?ddb=&type=pubmed_id&id="+pmid.to_s
  lookup(style,path)
end
 
def lookup(style,path)
  #Net::HTTP.new(Host, 80, "127.0.0.1", 8080)
  session= get_session 
 
  headers = {
    "User-Agent" => "Dillo/0.8.5-i18n-misc",
    "Referer" => "http://www.google.com/language_tools?hl=en",
    "Accept-Language" => "en-us,en;q=0.5",
    "Accept-Encoding" => "gzip,deflate",
    "Accept-Charset" => "ISO-8859-1,utf-8;q=0.7,*;q=0.7",
    "Keep-Alive" => "300"
  }
  res=session.get( path, headers )
 
  unless /^2\d\d/ =~ res.code
    $stderr.printf "HTTP problem: %d %s\n", res.code, res.message
    `wwwoffle \"#{Host+path}\" ` if /^5\d\d/ =~ res.code
    return nil
  end
 
  txt=res.body
  if /<textarea.*?>\{\{(.*?)\}\}/ =~ txt then
    tmpl=$1.gsub(/&ndash;/,"--")
    tmpl=CGI.unescapeHTML(tmpl)
    tmpl=tmpl.gsub(/--/,"&ndash;")
    #printf "{{%s}}\n", tmpl
    authors=""
    year=""
    tmpl.split("|").each{|el|
      if /author=(.*)/ =~ el
        #printf "author:%s\n",$1
        authors=$1
      elsif  /year=(\d+)/ =~ el
        #printf "year:%s\n",$1
        year=$1
      end
    }
    xauthors=[]
    # Harv template expects max 4 authors without first names
    authors.split(",")[0..3].each{|el|
      unless /''et al''/ =~ el
        xauthors.push el.split(" ")[0..-2].join(" ") # get rid of first names
      end
    }
 
    case style
    when /-txt/ then stl="Harvcoltxt"
    when /-nb/ then stl="Harvcolnb"
    else stl="Harvcol"
    end
    #p xauthors
    harv="{{#{stl} |"+xauthors.flatten.join(" |")+" |"+year+" }}"
    reftag="CITEREF#{xauthors.flatten.join('_')+'_'+year}".gsub(" ","_")
 
    return ["{{"+tmpl+"}}",harv,reftag]
  end
end
 
def replace_ref(orig,style,id)
  case orig
  when /pmid.*/i 
    type="pmid"
    res=lookup_pmid(style,id)
  when /isbn.*/i 
    type="pmid"
    res=lookup_isbn(style,id)
  end
 
  return orig unless res
  template,harv,reftag=res
 
  tag=type+id
  $references.push [tag,template,reftag] unless $references.assoc(tag)
 
  harv = harv + "<!-- #{orig} -->"
  return harv
end
 
# gets {{cite ..}}
def do_cite(tref)
  tag,citation,reftag = tref
 
  return "\n* {{anchor|#{reftag}}} #{citation}"
end
 
def ref_file(file)
  lines=file.readlines
  # 1st may be title or meta-information
  lines[2..-1].each{|line|
    # split {{(pmid|isbn)(|-nb|-txt) id-number}} into parts
    line.gsub!(/\{\{(pmid|isbn)(.*?)\s+(.*?)\}\}/i){|orig|
      replace_ref orig,$2,$3
    }
  }
 
  txt=lines.to_s
  references="<!-- automatically generated references block -->\n"
 
  #references=references+$references.collect{|el| "\n* "+el[1]}.to_s
  references=references+$references.collect{|el| do_cite(el)}.to_s
  references=references+"\n<!-- end of automatically generated references block -->\n"
  # insert at top of References section, user must sort it manually
  if /^==References==/ =~ txt
    txt.sub!(/^==References==/){|orig| orig+references }
  # no reference section yet? Create one right before stubs and Categories
  elsif /(\{\{.*stub\}\}|\[\[Category:.*\]\])/ =~ txt
    txt.sub!(/(\{\{.*stub\}\}|\[\[Category:.*\]\]).*/){|orig| 
      "==References==\n"+references+"\n"+orig }
  else
    txt=txt+"==References==\n"+references
  end
  printf "%s", txt
end
 
ARGV.each{|arg|
  $references=[]
  fl=File.open(arg,"r")
  ref_file fl
}