From d9d72e06db07087aa54401fae8b259f0c4ccd649 Mon Sep 17 00:00:00 2001 From: Patrick Simianer Date: Wed, 29 Jan 2014 19:22:56 +0100 Subject: first usable version, name change => nlp_ruby --- lib/nlp_ruby/stringutil.rb | 34 ++++++++++++++++++++++++++++++++++ 1 file changed, 34 insertions(+) create mode 100644 lib/nlp_ruby/stringutil.rb (limited to 'lib/nlp_ruby/stringutil.rb') diff --git a/lib/nlp_ruby/stringutil.rb b/lib/nlp_ruby/stringutil.rb new file mode 100644 index 0000000..e9a3bc9 --- /dev/null +++ b/lib/nlp_ruby/stringutil.rb @@ -0,0 +1,34 @@ +# whitespace 'tokenizer' +def tokenize s + s.strip.split +end + +def splitpipe s + s.strip.split(/\s*\|\|\|\s*/) +end + +def downcase? s + s[/[[:lower:]]/] +end + +# iterator over n-grams +def ngrams(s, n, fix=false) + a = tokenize s + a.each_with_index { |tok, i| + tok.strip! + 0.upto([n-1, a.size-i-1].min) { |m| + yield a[i..i+m] if !fix||(fix&&a[i..i+m].size==n) + } + } +end + +# a=1.0 b=2.0 => { 'a' => 1.0, 'b' => 2.0 } +def read_feature_string s + map = SparseVector.new + tokenize(s).each { |i| + key, value = i.split '=' + map[key] = value.to_f + } + return map +end + -- cgit v1.2.3