--- a
+++ b/classification/RNN/prepare_data.py
@@ -0,0 +1,73 @@
+import os
+import six
+import collections
+
+
+
+# 创建一个数据字典。数字字典就是把每个字都对应一个一个数字，包括标点符号。
+def build_dict(train_data_path, test_data_path, dict_path, cutoff=0):
+    word_freq = collections.defaultdict(int)
+    # 读取已经训练数据
+    with open(train_data_path, 'r', encoding='utf-8') as f:
+        train_data = [line.strip().split("\t")[2] for line in f]
+    with open(test_data_path, 'r', encoding='utf-8') as f:
+        test_data = [line.strip().split("\t")[2] for line in f]
+
+    for data in train_data+test_data:
+        for word in data:
+            word_freq[word] += 1
+    # Not sure if we should prune less-frequent words here.
+    word_freq = [x for x in six.iteritems(word_freq) if x[1] > cutoff]
+    dictionary = sorted(word_freq, key=lambda x: (-x[1], x[0]))
+
+    words, _ = list(zip(*dictionary))
+    word_idx = dict(list(zip(words, six.moves.range(len(words)))))
+    word_idx['<unk>'] = len(words)
+
+    # 把这些字典保存到本地中
+    with open(dict_path, 'w', encoding='utf-8') as f:
+        f.write(str(word_idx))
+    print("数据字典生成完成！")
+
+# 获取字典的长度
+def get_dict_len(dict_path):
+    with open(dict_path, 'r', encoding='utf-8') as f:
+        line = eval(f.readlines()[0])
+    return len(line.keys())
+
+def create_data_list(data_root_path):
+    with open(os.path.join(data_root_path, 'dict.txt'), 'r', encoding='utf-8') as f_dict:
+        dict_txt = eval(f_dict.readlines()[0])
+    print("字典长度：{}".format(len(dict_txt)))
+    with open(os.path.join(data_root_path, 'tags.txt'), 'r', encoding='utf-8') as f_tag:
+        tag_txt = eval(f_tag.readlines()[0])
+    print("类别数目：{}".format(len(tag_txt)))
+    with open(os.path.join(data_root_path, 'train_data.txt'), 'r', encoding='utf-8') as f_train:
+        lines = f_train.readlines()
+    with open(os.path.join(data_root_path, 'train_list.txt'), 'w', encoding='utf-8') as f_train_list:
+        for line in lines:
+            labs = ""
+            l = line.strip().split("\t")
+            for word in l[2]:
+                labs += str(dict_txt[word]) + ','
+            f_train_list.write(labs + '\t' + str(tag_txt[l[1]]) + '\n')
+
+    with open(os.path.join(data_root_path, 'test_data.txt'), 'r', encoding='utf-8') as f_test:
+        lines = f_test.readlines()
+    with open(os.path.join(data_root_path, 'test_list.txt'), 'w', encoding='utf-8') as f_test_list:
+        for line in lines:
+            labs = ""
+            l = line.strip().split("\t")
+            for word in l[2]:
+                labs += str(dict_txt[word]) + ','
+            f_test_list.write(labs + '\t' + str(tag_txt[l[1]]) + '\n')
+    print("数据列表生成完成！")
+
+if __name__ == "__main__":
+    data_root_path = "./data/"
+    train_data_path = "./data/train_data.txt"
+    test_data_path = "./data/test_data.txt"
+    dict_path = "./data/dict.txt"
+    # build_dict(train_data_path, test_data_path, dict_path)
+    # print(get_dict_len(dict_path))
+    # create_data_list(data_root_path)