summaryrefslogtreecommitdiff
path: root/feature-dict
diff options
context:
space:
mode:
authorPatrick Simianer <p@simianer.de>2016-07-05 11:01:46 +0200
committerPatrick Simianer <p@simianer.de>2016-07-05 11:01:46 +0200
commit2b1d7f881c19c4d4b5afae194e02d3300c7675d0 (patch)
tree5a06ee7de98640a39244b57bb369697176b44ebf /feature-dict
parent69949dda35c3ea21d8e926e5f0a596a0a0f61c6a (diff)
mv
Diffstat (limited to 'feature-dict')
-rwxr-xr-xfeature-dict24
1 files changed, 24 insertions, 0 deletions
diff --git a/feature-dict b/feature-dict
new file mode 100755
index 0000000..6849769
--- /dev/null
+++ b/feature-dict
@@ -0,0 +1,24 @@
+#!/usr/bin/env ruby
+
+not_quiet = ARGV[1]
+n = 0
+feature_dict = {}
+l_i = 1
+while line = STDIN.gets
+ STDERR.write "#{l_i}\n" if l_i%1000==0&&not_quiet
+ line.split.each { |i|
+ f, v = i.split('=', 2)
+ if !feature_dict.has_key? f
+ feature_dict[f] = n
+ n += 1
+ end
+ }
+ l_i += 1
+end
+
+f = File.new ARGV[0], 'w'
+f.write Marshal.dump feature_dict
+f.close
+
+STDERR.write "size = #{feature_dict.size}\n"
+