From 76866e8e7dcd89ab5e43ba53492c4a730a3e4638 Mon Sep 17 00:00:00 2001
From: Giuseppe Bilotta <giuseppe.bilotta@gmail.com>
Date: Sat, 28 Jul 2007 23:20:51 +0200
Subject: [PATCH] Initial checkout

---
 mark2.rb | 117 +++++++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 117 insertions(+)
 create mode 100755 mark2.rb

diff --git a/mark2.rb b/mark2.rb
new file mode 100755
index 0000000..731374b
--- /dev/null
+++ b/mark2.rb
@@ -0,0 +1,117 @@
+#! /usr/bin/ruby -w
+# vim: set sw=2 et:
+# Author: Giuseppe Bilotta <giuseppe.bilotta@gmail.com>
+# New markov chain plugin
+
+class Array
+  def butlast
+    self[0,self.size-1]
+  end
+  def butfirst
+    self[1,self.size]
+  end
+end
+
+class MarkovChainer
+  # Maximum depth
+  MAX_ORDER = 5
+
+  # Word or nonword regexp:
+  # can be used to scan a string splitting it into
+  # words and nonwords.
+  WNW = /\w+|\W/u
+
+  def initialize
+    # mkv[i] holds the chains of order i
+    @mkv = Array.new
+
+    MAX_ORDER.times { |i|
+      @mkv[i] = {}
+    }
+
+    # Each chain is in the form
+    # [:array, :of, :symbols] => {
+    #   :word => [chance before, chance after]
+    #   :word => [chance before, chance after]
+    # }
+    # except for order 0, which is just a hash of
+    # {:word => chance}
+  end
+
+  def add_one(sym)
+    s = sym.to_sym rescue nil
+    if @mkv[0].has_key?(s)
+      @mkv[0][s] += 1
+    else
+      @mkv[0][s] = 1
+    end
+  end
+
+  def add_before(array, prev)
+    raise "Not enough words in new data" if array.empty?
+    raise "Too many words in new data" if array.size > MAX_ORDER
+    size = array.size
+    if @mkv[size].has_key?(array)
+      h = @mkv[size][array]
+      if h.has_key?(prev)
+       h[prev][0] += 1
+      else
+       h[prev] = [1,0]
+      end
+    else
+      @mkv[size][array.dup] = { prev => [1, 0] }
+    end
+  end
+
+  def add_after(array, nxt)
+    raise "Not enough words in new data" if array.empty?
+    raise "Too many words in new data" if array.size > MAX_ORDER
+    size = array.size
+    if @mkv[size].has_key?(array)
+      h = @mkv[size][array]
+      if h.has_key?(nxt)
+       h[nxt][1] += 1
+      else
+       h[nxt] = [0,1]
+      end
+    else
+      @mkv[size][array.dup] = { nxt => [0, 1] }
+    end
+  end
+
+  def add_multi(array)
+    raise "Too many words in new data" if array.size > MAX_ORDER + 1
+    add_before(array.butfirst, array.first)
+    add_after(array.butlast, array.last)
+  end
+
+  def add(*data)
+    if data.size == 1
+      add_one(data.first)
+    else
+      add_multi(data)
+    end
+  end
+
+  def learn(text)
+    syms = text.scan(WNW).map { |w| w.intern } 
+    syms.unshift(nil)
+    syms.push(nil)
+
+    syms.size.times { |i|
+      [MAX_ORDER, syms.size-i].min.times { |ord|
+        v = syms[i, ord+1]
+        # puts "Learning #{v.inspect}"
+        add(*v)
+        # pp @mkv
+      }
+    }
+    pp @mkv if defined? pp
+  end
+
+end
+
+mkv = MarkovChainer.new
+
+mkv.learn("This is a test, this is a nice little test.")
+
-- 
2.32.0.93.g670b81a890