READY TO GO

jcyk · Apr 21, 2017 · e4f761e · e4f761e
commit e4f761e
Show file tree

Hide file tree

Showing 34 changed files with 1,947,626 additions and 0 deletions.
diff --git a/.gitignore b/.gitignore
@@ -0,0 +1,3 @@
+
+*.pyc
+/idioms/*
diff --git a/cleargit.sh b/cleargit.sh
@@ -0,0 +1,7 @@
+#!/bin/bash
+git checkout --orphan latest_branch
+git add -A
+git commit -am "READY TO GO"
+git branch -D master
+git branch -m master
+git push -f origin master
diff --git a/data/as_test b/data/as_test
diff --git a/data/as_train_all b/data/as_train_all
diff --git a/data/cityu_test b/data/cityu_test
diff --git a/data/cityu_train_all b/data/cityu_train_all
diff --git a/data/dic b/data/dic
diff --git a/data/dic.py b/data/dic.py
@@ -0,0 +1,10 @@
+# -*- coding: UTF-8 -*-
+words = set()
+with open('pku_train_all') as f:
+	for line in f.readlines():
+		sent =  unicode(line.decode('utf8')).split()
+		for word in sent:
+			words.add(word)
+for word in words:
+	print word.encode('utf8')
+
diff --git a/data/msr_dev b/data/msr_dev