def test_setitem():
od = OrderedDict()
od['alice'] = 0
od['bob'] = 1
od['carol'] = 2
assert len(od) == 3
od._check()
def test_delitem():
od = OrderedDict(pairs)
assert len(od) == 10
for value in range(10):
del od[value]
assert len(od) == 0
od._check()
def test_delitem():
od = OrderedDict(pairs)
assert len(od) == 10
for value in range(10):
del od[value]
assert len(od) == 0
od._check()
def test_iloc():
od = OrderedDict(enumerate(range(10)))
for num in range(10):
assert od.iloc[num] == num
od.iloc[-1] == 9
assert len(od.iloc) == 10
od._check()
def test_setitem():
od = OrderedDict()
od['alice'] = 0
od['bob'] = 1
od['carol'] = 2
assert len(od) == 3
od._check()
def __init__(self, topic_formatter=None, topic_filter=None, reducer=None):
self.handlers = OrderedDict() # {topic: handler}
self.topic_formatter = topic_formatter
self.topic_filter = topic_filter or (
lambda topic, received_ropic: topic == received_ropic
)
self.reducer = reducer or delistify
def test_iloc():
od = OrderedDict(enumerate(range(10)))
iloc = od.keys()
for num in range(10):
assert iloc[num] == num
iloc[-1] == 9
assert len(iloc) == 10
od._check()
def test_pop():
od = OrderedDict(enumerate(range(10)))
for num in range(10):
assert od.pop(num) == num
od._check()
assert od.pop(0, 'thing') == 'thing'
assert od.pop(1, default='thing') == 'thing'
od._check()
def test_popitem():
od = OrderedDict(enumerate(range(10)))
for num in reversed(range(10)):
key, value = od.popitem()
assert num == key == value
od._check()
od = OrderedDict(enumerate(range(10)))
for num in range(10):
key, value = od.popitem(last=False)
assert num == key == value
od._check()
class TopicHandler(object):
""" Run handlers on events. Each event has topic as it's id, which is passed as argument to handler functions. """
def __init__(self, topic_formatter=None, topic_filter=None, reducer=None):
self.handlers = OrderedDict() # {topic: handler}
self.topic_formatter = topic_formatter
self.topic_filter = topic_filter or (
lambda topic, received_ropic: topic == received_ropic
)
self.reducer = reducer or delistify
def add(self, topic, callback=None, formatter=None, filter=None, reducer=None):
handler = Handler(reducer=reducer)
handler.add(callback=callback, filter=filter, formatter=formatter)
self.handlers.setdefault(topic, Handler()).add(callback=handler)
def has_coroutine(self, topic):
return self.handlers[topic].has_coroutine()
def set_topic_formatter(self, formatter):
self.topic_formatter = formatter
def __call__(self, topic, *args, **kwargs):
args = [topic] + list(args)
if self.topic_formatter:
topic = self.topic_formatter(topic)
res = []
for _topic, handler in self.handlers.items():
if self.topic_filter(_topic, topic):
res.append(handler(*args, **kwargs))
return self.reducer(res)
def call(self, topic, *args, **kwargs):
return self.__call__(topic, *args, **kwargs)
async def call_async(self, topic, *args, **kwargs):
args = [topic] + list(args)
if self.topic_formatter:
topic = self.topic_formatter(topic)
res = []
for _topic, handler in self.handlers.items():
if self.topic_filter(_topic, topic):
res.append(await handler.call_async(*args, **kwargs))
return self.reducer(res)
def has_topic(self, topic):
return topic in self.handlers
def __getitem__(self, item):
return self.handlers[item]
def test_setdefault():
od = OrderedDict()
od.setdefault(0, False)
assert od[0] == False
od.setdefault(1, default=True)
assert od[1] == True
od.setdefault(2)
assert od[2] == None
assert od.setdefault(0) == False
assert od.setdefault(1) == True
def test_pop():
od = OrderedDict(enumerate(range(10)))
for num in range(10):
assert od.pop(num) == num
od._check()
assert od.pop(0, 'thing') == 'thing'
assert od.pop(1, default='thing') == 'thing'
od._check()
def test_setdefault():
od = OrderedDict()
od.setdefault(0, False)
assert od[0] == False
od.setdefault(1, default=True)
assert od[1] == True
od.setdefault(2)
assert od[2] == None
assert od.setdefault(0) == False
assert od.setdefault(1) == True
def test_popitem():
od = OrderedDict(enumerate(range(10)))
for num in reversed(range(10)):
key, value = od.popitem()
assert num == key == value
od._check()
od = OrderedDict(enumerate(range(10)))
for num in range(10):
key, value = od.popitem(last=False)
assert num == key == value
od._check()
def start_posting_procedure(self):
# Get new file descriptors for writing the buffer
if self.term_posting_fd is None:
self.get_new_term_posting_file()
if self.doc_posting_fd is None:
self.get_new_doc_posting_file()
if self.stemmer is not None:
self.stemmer.reset_dictionary()
docs_thread = Thread(name='post_docs', target=self.post_docs())
docs_thread.start()
main_dictionary_thread = Thread(name='store_terms',
target=self.dictionary.add_terms(
self.global_term_dict.values()))
main_dictionary_thread.start()
# Sort the term buffer before writing to storage
self.global_term_dict = OrderedDict(
sorted(self.global_term_dict.items()))
# Post the data to storage
self.post_terms_and_reset_posting(docs_thread, main_dictionary_thread)
def gen_search_gzh_url(keyword, page=1):
"""拼接搜索 公众号 URL
Parameters
----------
keyword : str or unicode
搜索文字
page : int, optional
页数 the default is 1
Returns
-------
str
search_gzh_url
"""
assert isinstance(page, int) and page > 0
qs_dict = OrderedDict()
qs_dict['type'] = _search_type_gzh
qs_dict['page'] = page
qs_dict['ie'] = 'utf8'
qs_dict['query'] = keyword
return 'http://weixin.sogou.com/weixin?{}'.format(urlencode(qs_dict))
def test_clear():
od = OrderedDict(pairs)
assert len(od) == 10
od.clear()
assert len(od) == 0
od._check()
def test_iter_reversed():
od = OrderedDict([('b', 0), ('a', 1), ('c', 2)])
assert list(od) == ['b', 'a', 'c']
assert list(reversed(od)) == ['c', 'a', 'b']
od._check()
def test_iter_reversed():
od = OrderedDict([('b', 0), ('a', 1), ('c', 2)])
assert list(od) == ['b', 'a', 'c']
assert list(reversed(od)) == ['c', 'a', 'b']
od._check()
def test_equality():
od = OrderedDict.fromkeys('abc')
assert od == {'a': None, 'b': None, 'c': None}
assert od != {}
assert od != OrderedDict()
od._check()
def test_pop_error():
od = OrderedDict()
od.pop(0)
def test_values():
od = OrderedDict(enumerate(range(10)))
assert list(reversed(od.values())) == list(reversed(range(10)))
assert od.values()[:3] == [0, 1, 2]
od._check()
def test_init():
od = OrderedDict()
assert len(od) == 0
od._check()
od = OrderedDict(enumerate(range(10)))
assert len(od) == 10
od._check()
od = OrderedDict(a=0, b=1, c=2)
assert len(od) == 3
od._check()
od = OrderedDict(pairs)
assert len(od) == 10
od._check()
def test_clear():
od = OrderedDict(pairs)
assert len(od) == 10
od.clear()
assert len(od) == 0
od._check()
def test_equality():
od = OrderedDict.fromkeys('abc')
assert od == {'a': None, 'b': None, 'c': None}
assert od != {}
assert od != OrderedDict()
od._check()
def test_pop_error():
od = OrderedDict()
with pytest.raises(KeyError):
od.pop(0)
def test_pop_error():
od = OrderedDict()
od.pop(0)
def test_items():
od = OrderedDict(enumerate(range(10)))
assert od.items() == list(enumerate(range(10)))
od._check()
def test_viewkeys():
od = OrderedDict(enumerate(range(10)))
view = od.viewkeys()
assert list(reversed(view)) == list(reversed(range(10)))
od._check()
def test_itervalues():
od = OrderedDict(enumerate(range(10)))
assert list(od.itervalues()) == list(range(10))
od._check()
def test_copy():
od = OrderedDict(enumerate(range(10)))
copy = od.copy()
assert od == copy
def test_copy():
od = OrderedDict(enumerate(range(10)))
copy = od.copy()
assert od == copy
def test_items():
items = list(enumerate(range(10)))
od = OrderedDict(enumerate(range(10)))
assert list(reversed(od.items())) == list(reversed(items))
assert od.items()[:3] == [(0, 0), (1, 1), (2, 2)]
od._check()
def test_fromkeys():
od = OrderedDict.fromkeys('abc')
assert od == {'a': None, 'b': None, 'c': None}
od._check()
def test_init():
od = OrderedDict()
assert len(od) == 0
od._check()
od = OrderedDict(enumerate(range(10)))
assert len(od) == 10
od._check()
od = OrderedDict(a=0, b=1, c=2)
assert len(od) == 3
od._check()
od = OrderedDict(pairs)
assert len(od) == 10
od._check()
def test_values():
od = OrderedDict(enumerate(range(10)))
assert od.values() == list(range(10))
od._check()
def test_repr():
od = OrderedDict()
assert repr(od) == 'OrderedDict([])'
assert str(od) == 'OrderedDict([])'
def test_reduce():
od = OrderedDict(enumerate(range(10)))
data = pickle.dumps(od)
copy = pickle.loads(data)
assert od == copy
class Indexer(object):
"""
Inverted index data structure
"""
def __init__(self, corpus_folder_path, stop_words_file_path, to_stem,
path_for_posting_and_dictionary):
self.stop_words_container = StopWordsContainer(stop_words_file_path)
self.stemmer = Stemmer() if to_stem else None
self.parser = Parse()
self.doc_posting_fd = None
self.doc_counter = 0 # Will be used to count the number of iterations
self.doc_repr_list = []
self.term_posting_fd = None
self.global_term_dict = {}
self.number_of_docs_processed = 0
self.corpus_folder = corpus_folder_path
self.stop_words_file_path = stop_words_file_path
self.path_for_posting_and_dictionary = path_for_posting_and_dictionary
self.stem_doc_posting_folder = os.path.join(
path_for_posting_and_dictionary,
IR_CONFIG["storage"]["stem_doc_posting_folder"])
self.doc_posting_folder = os.path.join(
path_for_posting_and_dictionary,
IR_CONFIG["storage"]["doc_posting_folder"])
self.stem_term_posting_folder = os.path.join(
path_for_posting_and_dictionary,
IR_CONFIG["storage"]["stem_term_posting_folder"])
self.term_posting_folder = os.path.join(
path_for_posting_and_dictionary,
IR_CONFIG["storage"]["term_posting_folder"])
if self.stemmer is not None:
self.doc_posting_file_path = self.stem_doc_posting_folder
self.term_posting_file_target = os.path.join(
self.stem_term_posting_folder,
IR_CONFIG["storage"]["stem_term_posting_file_name"])
self.doc_posting_file_target = os.path.join(
self.stem_doc_posting_folder,
IR_CONFIG["storage"]["stem_doc_posting_file_name"])
self.cache_file_path = os.path.join(
path_for_posting_and_dictionary,
IR_CONFIG["storage"]["cache_file_name_stem"])
self.dictionary_file_path = os.path.join(
path_for_posting_and_dictionary,
IR_CONFIG["storage"]["dictionary_file_name_stem"])
else:
self.doc_posting_file_path = self.doc_posting_folder
self.term_posting_file_target = os.path.join(
self.term_posting_folder,
IR_CONFIG["storage"]["term_posting_file_name"])
self.doc_posting_file_target = os.path.join(
self.doc_posting_folder,
IR_CONFIG["storage"]["doc_posting_file_name"])
self.cache_file_path = os.path.join(
path_for_posting_and_dictionary,
IR_CONFIG["storage"]["cache_file_name"])
self.dictionary_file_path = os.path.join(
path_for_posting_and_dictionary,
IR_CONFIG["storage"]["dictionary_file_name"])
self.dictionary = MainDictionary(self.term_posting_file_target)
def build_index(self):
"""
The main function for the index building.
"""
read_file_obj = ReadFile(self.corpus_folder)
for data in read_file_obj.extract_from_tags():
# Get handler
corpus_handler = self.get_handler(data)
# Get doc object to work on
doc_obj = self.get_doc_obj(corpus_handler)
# Append the doc obj to the list
self.doc_repr_list.append(doc_obj.__repr__())
# Update counter
self.doc_counter += 1
# In case we got to counter limit
if IR_CONFIG["indexer"]["doc_to_process"] == self.doc_counter:
self.start_posting_procedure()
# Still having data to process
if 0 < len(self.global_term_dict.values()):
self.start_posting_procedure()
# Merge the results to one single file
self.merge_terms_posting()
if self.doc_posting_fd is not None:
self.doc_posting_fd.close()
self.dictionary.update_term_data()
# self.dictionary.initialize_terms_pointer()
# self.dictionary.update_term_tf()
self.dictionary.filter_low_frequency_term()
self.dictionary.init_cache()
self.dictionary.initialize_document_dictionary(
self.doc_posting_file_target)
self.add_document_data()
try:
self.save_dict_and_cache(self.path_for_posting_and_dictionary)
except:
pass
def update_term_data(self):
self.dictionary.update_term_data()
def add_document_data(self):
if self.stemmer is not None:
path = os.path.join(
(os.path.join(
self.path_for_posting_and_dictionary,
IR_CONFIG["storage"]["stem_doc_posting_folder"])),
IR_CONFIG["storage"]["stem_doc_posting_file_name"])
else:
path = os.path.join(
(os.path.join(self.path_for_posting_and_dictionary,
IR_CONFIG["storage"]["doc_posting_folder"])),
IR_CONFIG["storage"]["doc_posting_file_name"])
with open(path + "2", 'w') as out_file:
with open(path, 'r') as in_file:
for line in in_file:
out_file.write(
self.calc_document_wight(line.rstrip('\n')) + '\n')
# os.remove(path)
# os.rename(path + "2", path.replace("2", ""))
def calc_document_wight(self, line):
splitted_line = line.split(",")
# doc id, doc file name, header, date, document length, term with max tf, *term#tf*....
document_weight = 0
max_tf_freq = int(splitted_line[5])
for data in splitted_line[6].split("*"):
try:
splitted_data = data.split("#")
term_tf = int(splitted_data[1])
term = splitted_data[1]
term_df = self.dictionary.data_dict[term][0]
num_of_docs = len(self.dictionary.documents_dictionary)
term_idf = math.log(num_of_docs / term_df, 2)
document_weight += ((term_tf / max_tf_freq) * term_idf)**2
except Exception:
pass # couldn't find the term in the dictionary
return ",".join([
splitted_line[0], splitted_line[1], splitted_line[2],
splitted_line[3], splitted_line[4], splitted_line[5],
str(document_weight**0.5)
])
def merge_terms_posting(self):
if self.stemmer is None:
merge_sorter = MergeFiles(
TermsMergerStrategy(),
(os.path.join(self.path_for_posting_and_dictionary,
IR_CONFIG["storage"]["term_posting_folder"])),
IR_CONFIG["storage"]["term_posting_file_name"])
else:
merge_sorter = MergeFiles(
TermsMergerStrategy(), (os.path.join(
self.path_for_posting_and_dictionary,
IR_CONFIG["storage"]["stem_term_posting_folder"])),
IR_CONFIG["storage"]["stem_term_posting_file_name"])
merge_sorter.merge_sort()
def start_posting_procedure(self):
# Get new file descriptors for writing the buffer
if self.term_posting_fd is None:
self.get_new_term_posting_file()
if self.doc_posting_fd is None:
self.get_new_doc_posting_file()
if self.stemmer is not None:
self.stemmer.reset_dictionary()
docs_thread = Thread(name='post_docs', target=self.post_docs())
docs_thread.start()
main_dictionary_thread = Thread(name='store_terms',
target=self.dictionary.add_terms(
self.global_term_dict.values()))
main_dictionary_thread.start()
# Sort the term buffer before writing to storage
self.global_term_dict = OrderedDict(
sorted(self.global_term_dict.items()))
# Post the data to storage
self.post_terms_and_reset_posting(docs_thread, main_dictionary_thread)
def reset_posting(self):
"""
Resets the data for new group of documents to process
:return:
"""
if self.term_posting_fd is not None:
self.term_posting_fd.close()
# self.doc_posting_fd = None
self.doc_counter = 0 # Will be used to count the number of iterations
self.doc_repr_list = []
self.term_posting_fd = None
self.global_term_dict = {}
def post_terms_and_reset_posting(self, docs_thread_in,
main_dictionary_thread_in):
"""
:return:
"""
term_thread = Thread(name='post_term', target=self.post_terms())
term_thread.start()
main_dictionary_thread_in.join()
term_thread.join()
docs_thread_in.join()
self.reset_posting()
def post_terms(self):
for term_str, term_obj in self.global_term_dict.iteritems():
self.term_posting_fd.write(term_obj.__repr__() + "\n")
def post_docs(self):
index = 0
for doc_repr in self.doc_repr_list:
self.number_of_docs_processed += 1
self.doc_posting_fd.write(doc_repr + "\n")
index += 1
def get_doc_obj(self, corpus_handler):
"""
Creates a new Document object from given handler.
:param corpus_handler: An handler of the current document
:return: A new Document to process
"""
return Document.Document(self.global_term_dict,
self.stop_words_container, self.stemmer,
corpus_handler)
def get_handler(self, data):
"""
Gets the proper handler for the given document
:param data: a document
:return: An handler for specific type of documents
"""
try:
prefix = " ".join(data[1].split())[:2]
if "FB" == prefix:
corpus_handler = FBCorpusHandler(data, self.parser)
elif "FT" == prefix:
corpus_handler = FTCorpusHandler(data, self.parser)
elif "LA" == prefix:
corpus_handler = LACorpusHandler(data, self.parser)
else:
raise ValueError("Prefix '%s' is not supported" % (prefix, ))
return corpus_handler
except Exception, e:
print traceback.print_exc(e.message)
def test_fromkeys():
od = OrderedDict.fromkeys('abc')
assert od == {'a': None, 'b': None, 'c': None}
od._check()
