Add convert to recordio function

python/paddle/v2/dataset/cifar.py

@@ -31,10 +31,10 @@
 import cPickle
 import itertools
 import numpy
-from common import download
+import paddle.v2.dataset.common
 import tarfile
 
-__all__ = ['train100', 'test100', 'train10', 'test10']
+__all__ = ['train100', 'test100', 'train10', 'test10', 'convert']
 
 URL_PREFIX = 'https://www.cs.toronto.edu/~kriz/'
 CIFAR10_URL = URL_PREFIX + 'cifar-10-python.tar.gz'
@@ -75,7 +75,8 @@ def train100():
     :rtype: callable
     """
     return reader_creator(
-        download(CIFAR100_URL, 'cifar', CIFAR100_MD5), 'train')
+        paddle.v2.dataset.common.download(CIFAR100_URL, 'cifar', CIFAR100_MD5),
+        'train')
 
 
 def test100():
@@ -88,7 +89,9 @@ def test100():
     :return: Test reader creator.
     :rtype: callable
     """
-    return reader_creator(download(CIFAR100_URL, 'cifar', CIFAR100_MD5), 'test')
+    return reader_creator(
+        paddle.v2.dataset.common.download(CIFAR100_URL, 'cifar', CIFAR100_MD5),
+        'test')
 
 
 def train10():
@@ -102,7 +105,8 @@ def train10():
     :rtype: callable
     """
     return reader_creator(
-        download(CIFAR10_URL, 'cifar', CIFAR10_MD5), 'data_batch')
+        paddle.v2.dataset.common.download(CIFAR10_URL, 'cifar', CIFAR10_MD5),
+        'data_batch')
 
 
 def test10():
@@ -116,9 +120,20 @@ def test10():
     :rtype: callable
     """
     return reader_creator(
-        download(CIFAR10_URL, 'cifar', CIFAR10_MD5), 'test_batch')
+        paddle.v2.dataset.common.download(CIFAR10_URL, 'cifar', CIFAR10_MD5),
+        'test_batch')
 
 
 def fetch():
-    download(CIFAR10_URL, 'cifar', CIFAR10_MD5)
-    download(CIFAR100_URL, 'cifar', CIFAR100_MD5)
+    paddle.v2.dataset.common.download(CIFAR10_URL, 'cifar', CIFAR10_MD5)
+    paddle.v2.dataset.common.download(CIFAR100_URL, 'cifar', CIFAR100_MD5)
+
+
+def convert(path):
+    """
+    Converts dataset to recordio format
+    """
+    paddle.v2.dataset.common.convert(path, train100(), 10, "cifar_train100")
+    paddle.v2.dataset.common.convert(path, test100(), 10, "cifar_test100")
+    paddle.v2.dataset.common.convert(path, train10(), 10, "cifar_train10")
+    paddle.v2.dataset.common.convert(path, test10(), 10, "cifar_test10")

python/paddle/v2/dataset/common.py

@@ -23,7 +23,10 @@
 import cPickle
 import glob
 
-__all__ = ['DATA_HOME', 'download', 'md5file', 'split', 'cluster_files_reader']
+__all__ = [
+    'DATA_HOME', 'download', 'md5file', 'split', 'cluster_files_reader',
+    'convert'
+]
 
 DATA_HOME = os.path.expanduser('~/.cache/paddle/dataset')
 

python/paddle/v2/dataset/conll05.py

@@ -23,9 +23,9 @@
 import tarfile
 import gzip
 import itertools
-from common import download
+import paddle.v2.dataset.common
 
-__all__ = ['test, get_dict', 'get_embedding']
+__all__ = ['test, get_dict', 'get_embedding', 'convert']
 
 DATA_URL = 'http://www.cs.upc.edu/~srlconll/conll05st-tests.tar.gz'
 DATA_MD5 = '387719152ae52d60422c016e92a742fc'
@@ -182,17 +182,23 @@ def get_dict():
     """
     Get the word, verb and label dictionary of Wikipedia corpus.
     """
-    word_dict = load_dict(download(WORDDICT_URL, 'conll05st', WORDDICT_MD5))
-    verb_dict = load_dict(download(VERBDICT_URL, 'conll05st', VERBDICT_MD5))
-    label_dict = load_dict(download(TRGDICT_URL, 'conll05st', TRGDICT_MD5))
+    word_dict = load_dict(
+        paddle.v2.dataset.common.download(WORDDICT_URL, 'conll05st',
+                                          WORDDICT_MD5))
+    verb_dict = load_dict(
+        paddle.v2.dataset.common.download(VERBDICT_URL, 'conll05st',
+                                          VERBDICT_MD5))
+    label_dict = load_dict(
+        paddle.v2.dataset.common.download(TRGDICT_URL, 'conll05st',
+                                          TRGDICT_MD5))
     return word_dict, verb_dict, label_dict
 
 
 def get_embedding():
     """
     Get the trained word vector based on Wikipedia corpus.
     """
-    return download(EMB_URL, 'conll05st', EMB_MD5)
+    return paddle.v2.dataset.common.download(EMB_URL, 'conll05st', EMB_MD5)
 
 
 def test():
@@ -209,15 +215,23 @@ def test():
     """
     word_dict, verb_dict, label_dict = get_dict()
     reader = corpus_reader(
-        download(DATA_URL, 'conll05st', DATA_MD5),
+        paddle.v2.dataset.common.download(DATA_URL, 'conll05st', DATA_MD5),
         words_name='conll05st-release/test.wsj/words/test.wsj.words.gz',
         props_name='conll05st-release/test.wsj/props/test.wsj.props.gz')
     return reader_creator(reader, word_dict, verb_dict, label_dict)
 
 
 def fetch():
-    download(WORDDICT_URL, 'conll05st', WORDDICT_MD5)
-    download(VERBDICT_URL, 'conll05st', VERBDICT_MD5)
-    download(TRGDICT_URL, 'conll05st', TRGDICT_MD5)
-    download(EMB_URL, 'conll05st', EMB_MD5)
-    download(DATA_URL, 'conll05st', DATA_MD5)
+    paddle.v2.dataset.common.download(WORDDICT_URL, 'conll05st', WORDDICT_MD5)
+    paddle.v2.dataset.common.download(VERBDICT_URL, 'conll05st', VERBDICT_MD5)
+    paddle.v2.dataset.common.download(TRGDICT_URL, 'conll05st', TRGDICT_MD5)
+    paddle.v2.dataset.common.download(EMB_URL, 'conll05st', EMB_MD5)
+    paddle.v2.dataset.common.download(DATA_URL, 'conll05st', DATA_MD5)
+
+
+def convert(path):
+    """
+    Converts dataset to recordio format
+    """
+    paddle.v2.dataset.common.convert(path, test(), 10, "conl105_train")
+    paddle.v2.dataset.common.convert(path, test(), 10, "conl105_test")

python/paddle/v2/dataset/imdb.py

@@ -28,7 +28,7 @@
 import string
 import threading
 
-__all__ = ['build_dict', 'train', 'test']
+__all__ = ['build_dict', 'train', 'test', 'convert']
 
 URL = 'http://ai.stanford.edu/%7Eamaas/data/sentiment/aclImdb_v1.tar.gz'
 MD5 = '7c2ac02c03563afcf9b574c7e56c153a'
@@ -166,3 +166,12 @@ def word_dict():
 
 def fetch():
     paddle.v2.dataset.common.download(URL, 'imdb', MD5)
+
+
+def convert(path):
+    """
+    Converts dataset to recordio format
+    """
+    w = word_dict()
+    paddle.v2.dataset.common.convert(path, lambda: train(w), 10, "imdb_train")
+    paddle.v2.dataset.common.convert(path, lambda: test(w), 10, "imdb_test")

python/paddle/v2/dataset/imikolov.py

@@ -22,7 +22,7 @@
 import collections
 import tarfile
 
-__all__ = ['train', 'test', 'build_dict']
+__all__ = ['train', 'test', 'build_dict', 'convert']
 
 URL = 'http://www.fit.vutbr.cz/~imikolov/rnnlm/simple-examples.tgz'
 MD5 = '30177ea32e27c525793142b6bf2c8e2d'
@@ -146,3 +146,15 @@ def test(word_idx, n, data_type=DataType.NGRAM):
 
 def fetch():
     paddle.v2.dataset.common.download(URL, "imikolov", MD5)
+
+
+def convert(path):
+    """
+    Converts dataset to recordio format
+    """
+    N = 5
+    word_dict = build_dict()
+    paddle.v2.dataset.common.convert(path,
+                                     train(word_dict, N), 10, "imikolov_train")
+    paddle.v2.dataset.common.convert(path,
+                                     test(word_dict, N), 10, "imikolov_test")

python/paddle/v2/dataset/mnist.py

@@ -21,7 +21,7 @@
 import subprocess
 import numpy
 import platform
-__all__ = ['train', 'test']
+__all__ = ['train', 'test', 'convert']
 
 URL_PREFIX = 'http://yann.lecun.com/exdb/mnist/'
 TEST_IMAGE_URL = URL_PREFIX + 't10k-images-idx3-ubyte.gz'
@@ -113,3 +113,11 @@ def fetch():
     paddle.v2.dataset.common.download(TRAIN_LABEL_URL, 'mnist', TRAIN_LABEL_MD5)
     paddle.v2.dataset.common.download(TEST_IMAGE_URL, 'mnist', TEST_IMAGE_MD5)
     paddle.v2.dataset.common.download(TEST_LABEL_URL, 'mnist', TRAIN_LABEL_MD5)
+
+
+def convert(path):
+    """
+    Converts dataset to recordio format
+    """
+    paddle.v2.dataset.common.convert(path, train(), 10, "minist_train")
+    paddle.v2.dataset.common.convert(path, test(), 10, "minist_test")

python/paddle/v2/dataset/movielens.py

@@ -23,14 +23,15 @@
 """
 
 import zipfile
-from common import download
+import paddle.v2.dataset.common
 import re
 import random
 import functools
 
 __all__ = [
     'train', 'test', 'get_movie_title_dict', 'max_movie_id', 'max_user_id',
-    'age_table', 'movie_categories', 'max_job_id', 'user_info', 'movie_info'
+    'age_table', 'movie_categories', 'max_job_id', 'user_info', 'movie_info',
+    'convert'
 ]
 
 age_table = [1, 18, 25, 35, 45, 50, 56]
@@ -99,7 +100,7 @@ def __repr__(self):
 
 
 def __initialize_meta_info__():
-    fn = download(URL, "movielens", MD5)
+    fn = paddle.v2.dataset.common.download(URL, "movielens", MD5)
     global MOVIE_INFO
     if MOVIE_INFO is None:
         pattern = re.compile(r'^(.*)\((\d+)\)$')
@@ -246,7 +247,15 @@ def unittest():
 
 
 def fetch():
-    download(URL, "movielens", MD5)
+    paddle.v2.dataset.common.download(URL, "movielens", MD5)
+
+
+def convert(path):
+    """
+    Converts dataset to recordio format
+    """
+    paddle.v2.dataset.common.convert(path, train(), 10, "movielens_train")
+    paddle.v2.dataset.common.convert(path, test(), 10, "movielens_test")
 
 
 if __name__ == '__main__':

python/paddle/v2/dataset/sentiment.py

@@ -26,9 +26,9 @@
 import nltk
 from nltk.corpus import movie_reviews
 
-import common
+import paddle.v2.dataset.common
 
-__all__ = ['train', 'test', 'get_word_dict']
+__all__ = ['train', 'test', 'get_word_dict', 'convert']
 NUM_TRAINING_INSTANCES = 1600
 NUM_TOTAL_INSTANCES = 2000
 
@@ -39,12 +39,13 @@ def download_data_if_not_yet():
     """
     try:
         # make sure that nltk can find the data
-        if common.DATA_HOME not in nltk.data.path:
-            nltk.data.path.append(common.DATA_HOME)
+        if paddle.v2.dataset.common.DATA_HOME not in nltk.data.path:
+            nltk.data.path.append(paddle.v2.dataset.common.DATA_HOME)
         movie_reviews.categories()
     except LookupError:
         print "Downloading movie_reviews data set, please wait....."
-        nltk.download('movie_reviews', download_dir=common.DATA_HOME)
+        nltk.download(
+            'movie_reviews', download_dir=paddle.v2.dataset.common.DATA_HOME)
         print "Download data set success....."
         print "Path is " + nltk.data.find('corpora/movie_reviews').path
 
@@ -128,4 +129,13 @@ def test():
 
 
 def fetch():
-    nltk.download('movie_reviews', download_dir=common.DATA_HOME)
+    nltk.download(
+        'movie_reviews', download_dir=paddle.v2.dataset.common.DATA_HOME)
+
+
+def convert(path):
+    """
+    Converts dataset to recordio format
+    """
+    paddle.v2.dataset.common.convert(path, train, 10, "sentiment_train")
+    paddle.v2.dataset.common.convert(path, test, 10, "sentiment_test")

python/paddle/v2/dataset/uci_housing.py

@@ -14,22 +14,22 @@
 """
 UCI Housing dataset.
 
-This module will download dataset from
+This module will paddle.v2.dataset.common.download dataset from
 https://archive.ics.uci.edu/ml/machine-learning-databases/housing/ and
 parse training set and test set into paddle reader creators.
 """
 
 import numpy as np
 import os
-from common import download
+import paddle.v2.dataset.common
 
 __all__ = ['train', 'test']
 
 URL = 'https://archive.ics.uci.edu/ml/machine-learning-databases/housing/housing.data'
 MD5 = 'd4accdce7a25600298819f8e28e8d593'
 feature_names = [
     'CRIM', 'ZN', 'INDUS', 'CHAS', 'NOX', 'RM', 'AGE', 'DIS', 'RAD', 'TAX',
-    'PTRATIO', 'B', 'LSTAT'
+    'PTRATIO', 'B', 'LSTAT', 'convert'
 ]
 
 UCI_TRAIN_DATA = None
@@ -82,7 +82,7 @@ def train():
     :rtype: callable
     """
     global UCI_TRAIN_DATA
-    load_data(download(URL, 'uci_housing', MD5))
+    load_data(paddle.v2.dataset.common.download(URL, 'uci_housing', MD5))
 
     def reader():
         for d in UCI_TRAIN_DATA:
@@ -102,7 +102,7 @@ def test():
     :rtype: callable
     """
     global UCI_TEST_DATA
-    load_data(download(URL, 'uci_housing', MD5))
+    load_data(paddle.v2.dataset.common.download(URL, 'uci_housing', MD5))
 
     def reader():
         for d in UCI_TEST_DATA:
@@ -112,4 +112,12 @@ def reader():
 
 
 def fetch():
-    download(URL, 'uci_housing', MD5)
+    paddle.v2.dataset.common.download(URL, 'uci_housing', MD5)
+
+
+def convert(path):
+    """
+    Converts dataset to recordio format
+    """
+    paddle.v2.dataset.common.convert(path, train(), 10, "uci_housing_train")
+    paddle.v2.dataset.common.convert(path, test(), 10, "uci_houseing_test")