sm-transcript 0.0.4 → 0.0.6

data/lib/sm_transcript/optparseExample.rb

@@ -1,113 +0,0 @@
-#!/usr/bin/env ruby
-
-require 'rubygems'
-require 'optparse'
-require 'optparse/time'
-require 'ostruct'
-require 'pp'
-
- class OptparseExample
-
-   CODES = %w[iso-2022-jp shift_jis euc-jp utf8 binary]
-   CODE_ALIASES = { "jis" => "iso-2022-jp", "sjis" => "shift_jis" }
-
-   #
-   # Return a structure describing the options.
-   #
-   def self.parse(args)
-     # The options specified on the command line will be collected in *options*.
-     # We set default values here.
-     options = OpenStruct.new
-     options.library = []
-     options.inplace = false
-     options.encoding = "utf8"
-     options.transfer_type = :auto
-     options.verbose = false
-
-     opts = OptionParser.new do |opts|
-       opts.banner = "Usage: example.rb [options]"
-
-       opts.separator ""
-       opts.separator "Specific options:"
-
-       # Mandatory argument.
-       opts.on("-r", "--require LIBRARY",
-               "Require the LIBRARY before executing your script") do |lib|
-         options.library << lib
-       end
-
-       # Optional argument; multi-line description.
-       opts.on("-i", "--inplace [EXTENSION]",
-               "Edit ARGV files in place",
-               "  (make backup if EXTENSION supplied)") do |ext|
-         options.inplace = true
-         options.extension = ext || ''
-         options.extension.sub!(/\A\.?(?=.)/, ".")  # Ensure extension begins with dot.
-       end
-
-       # Cast 'delay' argument to a Float.
-       opts.on("--delay N", Float, "Delay N seconds before executing") do |n|
-         options.delay = n
-       end
-
-       # Cast 'time' argument to a Time object.
-       opts.on("-t", "--time [TIME]", Time, "Begin execution at given time") do |time|
-         options.time = time
-       end
-
-       # Cast to octal integer.
-       opts.on("-F", "--irs [OCTAL]", OptionParser::OctalInteger,
-               "Specify record separator (default \\0)") do |rs|
-         options.record_separator = rs
-       end
-
-       # List of arguments.
-       opts.on("--list x,y,z", Array, "Example 'list' of arguments") do |list|
-         options.list = list
-       end
-
-       # Keyword completion.  We are specifying a specific set of arguments (CODES
-       # and CODE_ALIASES - notice the latter is a Hash), and the user may provide
-       # the shortest unambiguous text.
-       code_list = (CODE_ALIASES.keys + CODES).join(',')
-       opts.on("--code CODE", CODES, CODE_ALIASES, "Select encoding",
-               "  (#{code_list})") do |encoding|
-         options.encoding = encoding
-       end
-
-       # Optional argument with keyword completion.
-       opts.on("--type [TYPE]", [:text, :binary, :auto],
-               "Select transfer type (text, binary, auto)") do |t|
-         options.transfer_type = t
-       end
-
-       # Boolean switch.
-       opts.on("-v", "--[no-]verbose", "Run verbosely") do |v|
-         options.verbose = v
-       end
-
-       opts.separator ""
-       opts.separator "Common options:"
-
-       # No argument, shows at tail.  This will print an options summary.
-       # Try it and see!
-       opts.on_tail("-h", "--help", "Show this message") do
-         puts opts
-         exit
-       end
-
-       # Another typical switch to print the version.
-       opts.on_tail("--version", "Show version") do
-         puts OptionParser::Version.join('.')
-         exit
-       end
-     end
-
-     opts.parse!(args)
-     options
-   end  # parse()
-
- end  # class OptparseExample
-
- options = OptparseExample.parse(ARGV)
- pp options

data/lib/sm_transcript/process_csv_files_to_html.rb

@@ -1,58 +0,0 @@
-#!/usr/bin/env ruby -wKU
-$KCODE = "UTF8"
-# p "Résumé"
-
-# require "rexml/document"
-require "rubygems"
-require "htmlentities"
-
-if __FILE__ == $0
-
-#  Dir.glob("/Users/pwilkins/work/Spoken\ Lecture/IIHS/*transcript\ ????09.csv") { |x|
-  Dir.glob("/Users/pwilkins/Downloads/*transcript\ ????09.csv") { |x|
-#      reg = Regexp.new('[\w\-_]*\.seg$')
-      coder = HTMLEntities.new
-      span_element = ""
-      prev_start_time = 0
-      f = File.new(x)
-      puts f.path
-      f.each("\n") { |line|
-        arr = line.split(",")
-#        p arr[8]
-        start_time = arr[2].to_i/1000
-      if start_time == prev_start_time # append word
-#        span_element << " #{coder.encode(arr[8].strip, :basic)}"
-        span_element << " #{arr[8].strip}"
-      else # create a new span_element
-        puts span_element << "</span> "
-#        span_element = "<span id='T#{start_time}'>#{coder.encode(arr[8].strip, :basic)}"
-        span_element = "<span id='T#{start_time}'>#{arr[8].strip}"
-        prev_start_time = start_time
-      end
-#       puts line.dump
-#        puts "<span id='T#{start_time}'>#{arr[6]}</span>"
-
-      }
-
-#      puts f
-#      root.elements.each("/document/lecture/segment") do |s|
-#        span_element = ""
-#        prev_start_time = 0
-#        s.text.scan(/^\d* \d* [\w']*$/) { |t|
-## get the start time and reduce its granularity so that multiple words fall
-## within a <span> element.
-#          start_time = t.scan(/^\d*/)[0].to_i/1000
-#          if start_time == prev_start_time # append word
-#            span_element << " #{t.scan(/[\w']*$/)}"
-#          else # create a new span_element
-#            puts span_element << "</span> "
-#            span_element = "<span id='T#{start_time}'>#{t.scan(/[\w']*$/)}"
-#            prev_start_time = start_time
-#          end
-##          puts "<span id='T#{t.scan(/^\d*/)[0].to_i/1000}'>#{t.scan(/[\w']*$/)}</span> "
-#        }
-#      end
-      puts
-  }
-end
-

data/lib/sm_transcript/process_seg_files.rb

@@ -1,21 +0,0 @@
-require "rexml/document"
-require "optparse"
-
-if __FILE__ == $0
-
-  Dir.glob("/Users/pwilkins/work/Spoken\ Lecture/IIHS/*.seg") { |x|
-    reg = Regexp.new('[\w\-_]*\.seg$')
-    doc = REXML::Document.new File.open(x)
-    root = doc.root
-    f = reg.match( root.attributes.get_attribute("fileName").value )
-    root.elements.each("/document/lecture/segment") do |s|
-      s.text.scan(/^\d* \d* [\w']*$/) { |t|
-      puts f[0] + "," + s.attributes.get_attribute("id").value + "," + t.gsub(/ /, ',')
-    }
-  end
-  puts
-}
-end
-
-
-

data/lib/sm_transcript/process_seg_files_to_csv.rb

@@ -1,24 +0,0 @@
-require "rexml/document"
-
-if __FILE__ == $0
-
-  Dir.glob("/Users/pwilkins/work/Spoken\ Lecture/IIHS/*.seg") { |x|
-      reg = Regexp.new('[\w\-_]*\.seg$')
-      doc = REXML::Document.new File.open(x)
-      root = doc.root
-      f = reg.match( root.attributes.get_attribute("fileName").value )
-      root.elements.each("/document/lecture/segment") do |s|
-      print "Filename (do not edit),Segment (do not edit),Start Time (do not edit),End Time (do not edit),Unedited Word (do not edit),Edited English Word,Hindi Word(s)"
-      puts
-        s.text.scan(/^\d* \d* [\w']*$/) { |t|
-          puts f[0] + "," + s.attributes.get_attribute("id").value + "," + t.gsub(/ /, ',')
-        }
-      end
-      puts
-  }
-end
-
-# Header for csv file:
-# Filename (do not edit),Segment (do not edit),Start Time (do not edit),End Time (do not edit),Unedited Word (do not edit),Edited English Word,Hindi Word(s)
-
-

data/lib/sm_transcript/process_seg_files_to_html.rb

@@ -1,31 +0,0 @@
-require "rexml/document"
-
-if __FILE__ == $0
-
-  Dir.glob("/Users/pwilkins/work/Spoken\ Lecture/IIHS/*.seg") { |x|
-      reg = Regexp.new('[\w\-_]*\.seg$')
-      doc = REXML::Document.new File.open(x)
-      root = doc.root
-      f = reg.match( root.attributes.get_attribute("fileName").value )
-      puts f
-      root.elements.each("/document/lecture/segment") do |s|
-        span_element = ""
-        prev_start_time = 0
-        s.text.scan(/^\d* \d* [\w']*$/) { |t|
-# get the start time and reduce its granularity so that multiple words fall
-# within a <span> element.
-          start_time = t.scan(/^\d*/)[0].to_i/1000
-          if start_time == prev_start_time # append word
-            span_element << " #{t.scan(/[\w']*$/)}"
-          else # create a new span_element
-            puts span_element << "</span> "  
-            span_element = "<span id='T#{start_time}'>#{t.scan(/[\w']*$/)}" 
-            prev_start_time = start_time 
-          end
-#          puts "<span id='T#{t.scan(/^\d*/)[0].to_i/1000}'>#{t.scan(/[\w']*$/)}</span> "
-        }
-      end
-      puts
-  }
-end
-

data/lib/sm_transcript/require_relative.rb

@@ -1,14 +0,0 @@
-# This file implements the Ruby 1.9 feature 'require_relative'.  It is needed
-# only when running Ruby 1.8, so that code doesn't break.  It is taken from 
-# the Pickaxe book for 1.9.
- 
-def require_relative(relative_feature)
-  c = caller.first
-  fail "Can't parse #{c}" unless c.rindex(/:\d+(:in '.*')?$/)
-  file = $`
-  if /\A\((.*)\)/ =~ file # eval, etc.
-    raise LoadError, "require_relative is called in #{$1}"
-  end
-  absolute = File.expand_path(relative_feature, File.dirname(file))
-  require absolute
-end

data/test/transcripts/GardnerRileyInterview.t1.html

@@ -1,247 +0,0 @@
-<span id='T1'>this is kathie</span> 
-<span id='T3'>lee</span> 
-<span id='T5'>it's just such</span> 
-<span id='T7'>today I'm going to</span> 
-<span id='T8'>be</span> 
-<span id='T9'>some of you</span> 
-<span id='T10'>already know which</span> 
-<span id='T11'>teaches</span> 
-<span id='T15'>child</span> 
-<span id='T18'>my research focuses</span> 
-<span id='T19'>on</span> 
-<span id='T20'>most nineteen</span> 
-<span id='T21'>forty</span> 
-<span id='T22'>cheer</span> 
-<span id='T23'>larry theory</span> 
-<span id='T24'>particularly</span> 
-<span id='T28'>systems feminist</span> 
-<span id='T29'>theory to</span> 
-<span id='T30'>analyze</span> 
-<span id='T31'>how contemporary</span> 
-<span id='T32'>parisian two</span> 
-<span id='T33'>issues right</span> 
-<span id='T35'>okay</span> 
-<span id='T36'>and it's fairly clear</span> 
-<span id='T37'>to me why</span> 
-<span id='T38'>research is</span> 
-<span id='T40'>engineering</span> 
-<span id='T42'>I think</span> 
-<span id='T43'>it's fair to</span> 
-<span id='T44'>this for a couple of</span> 
-<span id='T45'>reasons one it's my</span> 
-<span id='T46'>approach</span> 
-<span id='T47'>to just</span> 
-<span id='T48'>trying</span> 
-<span id='T49'>to theoretical</span> 
-<span id='T50'>perspectives</span> 
-<span id='T51'>engineering disciplines</span> 
-<span id='T52'>that</span> 
-<span id='T53'>really</span> 
-<span id='T54'>elaborate leases together</span> 
-<span id='T55'>so</span> 
-<span id='T56'>for my analysis</span> 
-<span id='T58'>also I think</span> 
-<span id='T60'>it's</span> 
-<span id='T61'>very</span> 
-<span id='T62'>ways</span> 
-<span id='T63'>just from where start</span> 
-<span id='T64'>my research</span> 
-<span id='T65'>questions to the absolute</span> 
-<span id='T66'>large exactly</span> 
-<span id='T67'>no kidding</span> 
-<span id='T68'>example I just</span> 
-<span id='T69'>recently</span> 
-<span id='T70'>conversely an</span> 
-<span id='T71'>article</span> 
-<span id='T72'>professor season</span> 
-<span id='T73'>volt is also</span> 
-<span id='T74'>interested</span> 
-<span id='T75'>in hearing master</span> 
-<span id='T76'>as</span> 
-<span id='T77'>opposed to circumvent</span> 
-<span id='T78'>weren't use</span> 
-<span id='T79'>a visiting</span> 
-<span id='T80'>professor of</span> 
-<span id='T81'>us who studies</span> 
-<span id='T82'>did in our</span> 
-<span id='T83'>group together</span> 
-<span id='T85'>vision of social</span> 
-<span id='T86'>justice</span> 
-<span id='T87'>integrate picture</span> 
-<span id='T88'>so we came</span> 
-<span id='T89'>together in a way</span> 
-<span id='T90'>the system that</span> 
-<span id='T91'>has that kind of justice</span> 
-<span id='T92'>background</span> 
-<span id='T93'>and issues the justice</span> 
-<span id='T94'>studies</span> 
-<span id='T95'>kathleen brothers</span> 
-<span id='T96'>rhetoric can</span> 
-<span id='T97'>help language works</span> 
-<span id='T98'>we conceptualize</span> 
-<span id='T99'>justice to</span> 
-<span id='T100'>language</span> 
-<span id='T101'>I wrote feminist theory</span> 
-<span id='T102'>and talked</span> 
-<span id='T103'>about how we can start</span> 
-<span id='T104'>thinking of social justice</span> 
-<span id='T105'>to</span> 
-<span id='T106'>balance principles</span> 
-<span id='T107'>suggests</span> 
-<span id='T108'>equality</span> 
-<span id='T109'>generous</span> 
-<span id='T110'>thinking about the</span> 
-<span id='T111'>justice</span> 
-<span id='T112'>is less individualized</span> 
-<span id='T113'>rationals</span> 
-<span id='T114'>thirty or about</span> 
-<span id='T116'>care</span> 
-<span id='T117'>we mentioned</span> 
-<span id='T118'>the ways</span> 
-<span id='T119'>in which we entrance</span> 
-<span id='T121'>social justice</span> 
-<span id='T122'>we did</span> 
-<span id='T123'>so by looking at</span> 
-<span id='T124'>work I actually</span> 
-<span id='T125'>person in</span> 
-<span id='T126'>charge</span> 
-<span id='T127'>and markings sort</span> 
-<span id='T128'>talk about what the</span> 
-<span id='T129'>messages to me see</span> 
-<span id='T130'>in this text</span> 
-<span id='T131'>about ways</span> 
-<span id='T132'>in which changes</span> 
-<span id='T134'>yeah that's</span> 
-<span id='T135'>interesting I</span> 
-<span id='T136'>hope everybody</span> 
-<span id='T137'>think there's no</span> 
-<span id='T138'>chance</span> 
-<span id='T140'>around</span> 
-<span id='T141'>the sort</span> 
-<span id='T142'>of sorry</span> 
-<span id='T144'>the patience</span> 
-<span id='T146'>to</span> 
-<span id='T147'>mainstream</span> 
-<span id='T148'>education</span> 
-<span id='T158'>is english</span> 
-<span id='T159'>so I necessarily</span> 
-<span id='T160'>star cluster</span> 
-<span id='T161'>making</span> 
-<span id='T162'>any do</span> 
-<span id='T163'>switcher</span> 
-<span id='T164'>necessarily to start</span> 
-<span id='T166'>reading</span> 
-<span id='T167'>and</span> 
-<span id='T168'>then he was a very</span> 
-<span id='T170'>various feminist</span> 
-<span id='T171'>theory that might</span> 
-<span id='T172'>help me</span> 
-<span id='T173'>developed fourier</span> 
-<span id='T174'>analysis of intense</span> 
-<span id='T175'>exactly</span> 
-<span id='T178'>and you know article</span> 
-<span id='T179'>in</span> 
-<span id='T180'>the summer</span> 
-<span id='T181'>they've looked at</span> 
-<span id='T182'>nature writing environmental</span> 
-<span id='T183'>times their</span> 
-<span id='T184'>readers it's a very</span> 
-<span id='T185'>all over the</span> 
-<span id='T186'>famous for</span> 
-<span id='T187'>terrigenous</span> 
-<span id='T188'>millions of</span> 
-<span id='T189'>fiction writers</span> 
-<span id='T191'>the united</span> 
-<span id='T192'>states</span> 
-<span id='T193'>and let's look</span> 
-<span id='T194'>at the particularly</span> 
-<span id='T195'>think about</span> 
-<span id='T197'>what's happening in</span> 
-<span id='T198'>nature reading</span> 
-<span id='T199'>because</span> 
-<span id='T200'>people to react</span> 
-<span id='T201'>to the genes</span> 
-<span id='T202'>gauguin</span> 
-<span id='T203'>are</span> 
-<span id='T204'>willing to was running</span> 
-<span id='T205'>you know it's</span> 
-<span id='T206'>very to</span> 
-<span id='T207'>talk about how these</span> 
-<span id='T208'>writers are actually modeling</span> 
-<span id='T209'>first principles</span> 
-<span id='T210'>of collaboration</span> 
-<span id='T212'>principles that</span> 
-<span id='T213'>diffusing</span> 
-<span id='T215'>our relationships</span> 
-<span id='T216'>in</span> 
-<span id='T217'>our world</span> 
-<span id='T218'>into just had a breakdown</span> 
-<span id='T219'>those boundaries</span> 
-<span id='T220'>actually</span> 
-<span id='T221'>came from</span> 
-<span id='T222'>so that</span> 
-<span id='T223'>focus on interdependence</span> 
-<span id='T224'>collaboration</span> 
-<span id='T226'>no longer have a higher</span> 
-<span id='T228'>relationships</span> 
-<span id='T229'>as a teenager</span> 
-<span id='T230'>culture</span> 
-<span id='T231'>is so</span> 
-<span id='T232'>it's going</span> 
-<span id='T234'>argue is one of these</span> 
-<span id='T235'>contemporary parisian</span> 
-<span id='T236'>julius create</span> 
-<span id='T237'>a dipole kick</span> 
-<span id='T238'>up there</span> 
-<span id='T239'>a new way of</span> 
-<span id='T240'>thinking about</span> 
-<span id='T241'>nature</span> 
-<span id='T242'>was there was definitely</span> 
-<span id='T243'>for</span> 
-<span id='T244'>us</span> 
-<span id='T245'>here</span> 
-<span id='T246'>okay yeah</span> 
-<span id='T247'>it's</span> 
-<span id='T248'>it's very clear</span> 
-<span id='T249'>I mean how how james</span> 
-<span id='T251'>it's</span> 
-<span id='T252'>way</span> 
-<span id='T261'>oh jeez</span> 
-<span id='T263'>I</span> 
-<span id='T264'>think humanities</span> 
-<span id='T266'>perspective</span> 
-<span id='T268'>he anteks</span> 
-<span id='T269'>paralyzing</span> 
-<span id='T270'>where it says</span> 
-<span id='T271'>experiences</span> 
-<span id='T272'>to look at</span> 
-<span id='T273'>the issues of sexuality</span> 
-<span id='T275'>and difference</span> 
-<span id='T276'>of ways in</span> 
-<span id='T277'>which we think about identity</span> 
-<span id='T278'>in the</span> 
-<span id='T279'>united states</span> 
-<span id='T280'>how to</span> 
-<span id='T281'>change</span> 
-<span id='T282'>and its relationship</span> 
-<span id='T283'>to be our</span> 
-<span id='T284'>relationship</span> 
-<span id='T285'>to each other</span> 
-<span id='T286'>require</span> 
-<span id='T287'>individual or person</span> 
-<span id='T288'>rapes</span> 
-<span id='T289'>ethnicity</span> 
-<span id='T290'>sex</span> 
-<span id='T291'>nationality</span> 
-<span id='T292'>resonance</span> 
-<span id='T293'>challenge how</span> 
-<span id='T294'>we operate world</span> 
-<span id='T296'>as individuals and</span> 
-<span id='T297'>also we operate</span> 
-<span id='T302'>and later</span> 
-<span id='T303'>the semester</span> 
-<span id='T312'>it wasn't</span> 
-<span id='T314'>to see commonality</span> 
-<span id='T315'>between</span> 
-<span id='T316'>questions</span> 
-<span id='T321'>thank you</span>