def send_emails(modeladmin, request, queryset):
messages = Queue()
for user in queryset:
process = Process(target=send_email, args=(user, messages))
process.start()
messages.get().send()
process.join()
def _get_system_information_threaded(host):
system_information_functions = [
collect_win_application_stats, collect_win_bios_stats,
collect_win_disk_stats, collect_win_local_account_stats,
collect_win_local_group_stats, collect_win_mem_stats,
collect_win_network_stats, collect_os_stats,
collect_win_processes_stats, collect_win_cpu_stats,
collect_win_services_stats
]
system_information = {}
queue = Queue()
list_of_processes = []
for hardware in system_information_functions:
process = _Process(target=hardware, args=(
host,
1,
queue,
))
list_of_processes.append(process)
process.start()
for process in list_of_processes:
process.join()
system_information.update(queue.get())
return system_information
def main():
'''load video, process frames, display to user'''
tque = Queue() #(maxsize=120)
framequeue = Queue() #(maxsize=120)
cthread = threading.Thread(target=cvworker, args=(tque, ))
cthread.daemon = True
cthread.start()
tthread = threading.Thread(target=tfworker, args=(tque, framequeue))
tthread.daemon = True #terminate testloop when user closes window
tthread.start()
start = time.time()
frame = 0
videoend = False
while True:
cvw = cv2.waitKey(1)
if cvw & 0xFF == ord('q'): break
if not videoend:
print('got', frame, time.time())
frame += 1
print('frame:', frame)
f = framequeue.get()
if type(f) == type(None):
videoend = True
pass #whats this do
else:
#time.sleep(1/30) #limit to realtime
cv2.imshow('frame', f)
print('new took:', time.time() - start)
cv2.destroyAllWindows()
class Actor(object):
def __init__(self):
# Actor内部的消息缓存队列
self.__mailbox = Queue()
def send(self, msg):
self.__mailbox.put(msg)
def recv(self):
return self.__mailbox.get()
def find_words(start_words,
center_words=None,
neg_words=None,
min_sim=0.6,
max_sim=1.,
alpha=0.25):
if center_words == None and neg_words == None:
min_sim = max(min_sim, 0.6)
center_vec, neg_vec = np.zeros([word_size]), np.zeros([word_size])
if center_words: # 中心向量是所有中心种子词向量的平均
_ = 0
for w in center_words:
if w in word2vec.wv.vocab:
center_vec += word2vec[w]
_ += 1
if _ > 0:
center_vec /= _
if neg_words: # 负向量是左右负种子词向量的平均（此处无用）
_ = 0
for w in neg_words:
if w in word2vec.wv.vocab:
neg_vec += word2vec[w]
_ += 1
if _ > 0:
neg_vec /= _
queue_count = 1
task_count = 0
cluster = []
queue = Queue() # 建立队列
for w in start_words:
queue.put((0, w))
if w not in cluster:
cluster.append(w)
while not queue.empty():
idx, word = queue.get()
queue_count -= 1
task_count += 1
sims = most_similar(word, center_vec, neg_vec)
min_sim_ = min_sim + (max_sim - min_sim) * (1 - np.exp(-alpha * idx))
if task_count % 10 == 0:
log = '%s in cluster, %s in queue, %s tasks done, %s min_sim' % (
len(cluster), queue_count, task_count, min_sim_)
print(log)
for i, j in sims:
if j >= min_sim_:
if i not in cluster and is_good(i): # is_good是人工写的过滤规则
queue.put((idx + 1), i)
if i not in cluster and is_good(i):
cluster.append(i)
queue_count += 1
return cluster
def _add_doi(metadata, identifier, citekey):
"""Add an entry from a DOI."""
info_messages = []
with StatusMessage('Querying DOI metadata...') as message:
if metadata.doi_exists(identifier):
raise ZoiaAddException(f'DOI {identifier} already exists.')
# Query Semantic Scholar to get the corresponding arxiv ID (if there is
# one) in a separate thread.
arxiv_queue = ThreadQueue()
arxiv_process = ThreadProcess(
target=lambda q, x: q.put(requests.get(x)),
args=(
arxiv_queue,
f'https://api.semanticscholar.org/v1/paper/{identifier}',
),
)
arxiv_process.start()
doi_metadata = _get_doi_metadata(identifier)
metadatum = zoia.backend.metadata.Metadatum.from_dict(doi_metadata)
if citekey is None:
citekey = zoia.parse.citekey.create_citekey(metadata, metadatum)
paper_dir = os.path.join(metadata.config.library_root, citekey)
os.mkdir(paper_dir)
message.update(
'Querying Semantic Scholar for corresponding arXiv ID...')
arxiv_metadata_response = arxiv_queue.get()
arxiv_process.join()
arxiv_metadata = json.loads(arxiv_metadata_response.text)
if (arxiv_id := arxiv_metadata.get('arxivId')) is not None:
doi_metadata['arxiv_id'] = arxiv_id
message.update('Downloading PDF from arXiv...')
pdf_response = requests.get(
f'https://arxiv.org/pdf/{arxiv_id}.pdf')
if pdf_response.status_code == 200:
with open(os.path.join(paper_dir, 'document.pdf'), 'wb') as fp:
fp.write(pdf_response.content)
doi_metadata['pdf_md5'] = hashlib.md5(
pdf_response.content).hexdigest()
else:
info_messages.append('Was unable to fetch a PDF')
metadata[citekey] = doi_metadata
def run_tasks(tasks, context=None, nb_threads=1, watchdog=None):
got_keyboard_interrupt = False
watchdogs = [
lambda _: _KEYBOARD_INTERRUPT_ERROR_MESSAGE
if got_keyboard_interrupt else None
]
if watchdog:
watchdogs.append(watchdog)
for task in tasks:
check_task_dependencies(task)
remaining_tasks = list(tasks)
completed_tasks = list()
pool = Pool(nb_threads)
completed_tasks_queue = Queue()
try:
schedule_tasks_to_be_run(
pop_runnable_tasks(remaining_tasks, completed_tasks, nb_threads),
watchdogs, context, pool, completed_tasks_queue)
while len(completed_tasks) != len(tasks):
# wait for one task to complete
completed_task = completed_tasks_queue.get()
completed_tasks.append(completed_task)
# schedule tasks to be run waiting for task success or simple completion
tasks_to_be_run = pop_runnable_tasks(remaining_tasks,
completed_tasks, nb_threads)
schedule_tasks_to_be_run(tasks_to_be_run, watchdogs, context, pool,
completed_tasks_queue)
except KeyboardInterrupt:
got_keyboard_interrupt = True
skip_all_tasks(tasks, remaining_tasks, completed_tasks, context, pool,
completed_tasks_queue,
_KEYBOARD_INTERRUPT_ERROR_MESSAGE)
finally:
pool.close()
exceptions = [
task.result.stacktrace for task in tasks
if isinstance(task.result, TaskResultException)
]
if exceptions:
raise TasksExecutionFailure("Caught exceptions:\n%s" %
"\n".join(exceptions))
class Recon:
def __init__(self, model_dir, path_group_dict: Dict[str, int]):
self.feed = Queue()
self.mid = Queue()
self.out = Queue()
t = MtCNN(self.feed, self.mid)
t.start()
t1 = EmbeddingCmp(self.feed, self.mid, self.out, path_group_dict,
model_dir)
t1.start()
def face_check(self, _im: np.ndarray, group: int):
self.feed.put((_im, group))
return self.out.get()
class BaseActor(object):
def __init__(self):
"""queue：Actor内部的邮箱队列"""
self.__mailbox = Queue()
def recv(self):
"""Actor接受消息"""
msg = self.__mailbox.get()
if msg is ActorExit:
# 抛出异常（模版方法会处理）
raise ActorExit
return msg
def send(self, msg):
"""Actor发送消息"""
self.__mailbox.put(msg)
def close(self):
"""发送结束标识"""
self.send(ActorExit)
def start(self):
self.__terminated_event = Event() # 为Join服务
t = threading.Thread(target=self.__templet)
t.setDaemon(True) # 设置为守护线程
t.start()
def __templet(self):
"""模版方法（run会被子类重写）"""
try:
self.run() # 执行Run代码
except ActorExit:
pass # 防止线程挂掉
finally:
# 设置Event标识
self.__terminated_event.set()
def join(self):
# Event在set之后便结束等待
self.__terminated_event.wait()
def run(self):
"""
由子类实现即可，eg：
while True:
msg = self.recv()
print(msg)
"""
pass
def get_stats(): print 'Fetching NBA player stats...' stats_outfile = RUNDAY+'_nba_stats.csv' csvout = open(stats_outfile, 'wb') NUM_THREADS = 8 in_queue = Queue() out_queue = Queue() queue_players(in_queue) while not in_queue.empty(): jobs = [] for i in range(NUM_THREADS): if not in_queue.empty(): thread = Process(target=get_stats_helper, args=(in_queue, out_queue)) jobs.append(thread) thread.start() for thread in jobs: thread.join() while not out_queue.empty(): player = out_queue.get() del player['SUCCESS'] try: name = player['NAME'] except KeyError as e: continue player['TIME'] = RUNDAY fieldnames = [ 'TIME', 'NAME', 'JERSEY', 'SPORT', 'TEAM', 'POSITION', 'PTS', 'REB', 'AST', 'URL' ] csvwriter = csv.DictWriter(csvout, delimiter='|', fieldnames=fieldnames) csvwriter.writerow(player) csvout.close() print 'Finished fetching NBA player stats.' print 'Ouput saved in %s' % stats_outfile
def _add_arxiv_id(metadata, identifier, citekey=None):
info_messages = []
with StatusMessage('Querying arXiv...') as message:
if metadata.arxiv_id_exists(identifier):
raise ZoiaAddException(f'arXiv paper {identifier} already exists.')
# Downloading the PDF can take a while, so start it early in a separate
# thread.
pdf_queue = ThreadQueue()
pdf_process = ThreadProcess(
target=lambda q, x: q.put(requests.get(x)),
args=(pdf_queue, f'https://arxiv.org/pdf/{identifier}.pdf'),
)
pdf_process.start()
arxiv_metadata = _get_arxiv_metadata(identifier)
if 'doi' in arxiv_metadata:
message.update('Querying DOI information...')
arxiv_metadata.update(_get_doi_metadata(arxiv_metadata['doi']))
if citekey is None:
metadatum = zoia.backend.metadata.Metadatum.from_dict(
arxiv_metadata)
citekey = zoia.parse.citekey.create_citekey(metadata, metadatum)
paper_dir = os.path.join(metadata.config.library_root, citekey)
os.mkdir(paper_dir)
message.update(text='Downloading PDF...')
pdf = pdf_queue.get()
pdf_process.join()
if pdf.status_code == 200:
with open(os.path.join(paper_dir, 'document.pdf'), 'wb') as fp:
fp.write(pdf.content)
md5_hash = hashlib.md5(pdf.content).hexdigest()
arxiv_metadata['pdf_md5'] = md5_hash
if metadata.pdf_md5_hash_exists(md5_hash):
raise ZoiaAddException(
f'arXiv paper {identifier} already exists.')
else:
info_messages.append('Was unable to fetch a PDF')
metadata[citekey] = arxiv_metadata
return citekey, metadatum, info_messages
def main(test_cases):
procs = []
queue = Queue()
case_number = 0
for test_case in test_cases:
proc = Process(target=handle_test_case,
args=(test_case, case_number, queue))
procs.append(proc)
case_number += 1
for proc in procs:
proc.start()
for proc in procs:
proc.join()
results = [queue.get() for proc in procs]
results.sort()
for r in results:
print(r[1])
def stat_files():
all_files = []
for root, dirs, files in os.walk('/home/gzguoyubo/mf/tw2/res/entities/custom_type'):
ignore = False
for ig_path in ignore_paths:
if ig_path in root:
ignore = True
if ignore:
continue
for fname in files:
if not fname.endswith('.py'):
continue
abs_file_path = join(root, fname)
all_files.append(abs_file_path)
file_sections = []
file_total_nums = len(all_files)
for i in xrange(P_NUM):
start = i * file_total_nums / P_NUM
stop = start + file_total_nums / P_NUM
if i == P_NUM - 1:
stop = -1
file_sections.append(all_files[start : stop])
res_queue = Queue()
processes = []
for section in file_sections:
p = Process(target=stat_file, args=(section, res_queue))
p.start()
processes.append(p)
for p in processes:
p.join()
total_stats = defaultdict(int)
while not res_queue.empty():
stat = res_queue.get()
for author, cnt in stat.iteritems():
total_stats[author] += cnt
print total_stats
class TreeCrawler(object):
def __init__(self, base_path, callback=None):
if not os.path.isdir(base_path):
raise IOError("Base path not found: " + base_path)
self.base_path = base_path
self.unsearched = Manager().Queue()
self.dirpath_queue = Queue()
self.cpu_count = multiprocessing.cpu_count()
self.pool = Pool(self.cpu_count)
self.first_level_dirs = ""
self.callback = callback
def __explore_path(self):
directories = []
dirpath = self.dirpath_queue.get()
for filename in os.walk(dirpath).next()[1]:
fullname = os.path.join(dirpath, filename)
directories.append(fullname)
return directories
def run_crawler(self):
# acquire the list of all paths inside base path
self.first_level_dirs = next(os.walk(self.base_path))[1]
for path in self.first_level_dirs:
self.unsearched.put(self.base_path + "/" + path)
self.pool.map_async(self.parallel_worker, range(self.cpu_count))
self.pool.close()
self.unsearched.join()
def parallel_worker(self, task_num):
while True:
dirpath = self.unsearched.get()
print "Task: " + str(task_num) + " >>> Explored path: " + dirpath
self.dirpath_queue.put(dirpath)
dirs = self.__explore_path()
for newdir in dirs:
self.unsearched.put(newdir)
self.unsearched.task_done()
class ExThread(Thread):
def __init__(self, *args, **kwargs):
super().__init__(*args, **kwargs)
self.__status_queue = Queue()
def run(self, *args, **kwargs):
try:
super().run(*args, **kwargs)
except Exception as ex:
#except_type, except_class, tb = sys.exc_info()
self.__status_queue.put(ex)
self.__status_queue.put(None)
def wait_for_exc_info(self):
return self.__status_queue.get()
def join(self):
ex = self.wait_for_exc_info()
if ex is None:
return
else:
raise ex
class ExThread(Thread):
def __init__(self, *args, **kwargs):
super().__init__(*args, **kwargs)
self.__status_queue = Queue()
def run(self, *args, **kwargs):
try:
super().run(*args, **kwargs)
except Exception as ex:
#except_type, except_class, tb = sys.exc_info()
self.__status_queue.put(ex)
self.__status_queue.put(None)
def wait_for_exc_info(self):
return self.__status_queue.get()
def join(self):
ex = self.wait_for_exc_info()
if ex is None:
return
else:
raise ex
def downloadFiles(downloadFiles, n_downloadFiles, size_downloadFiles):
global q, writeDict
downloadChunks = partitionDownload(downloadFiles, n_downloadFiles, size_downloadFiles)
n_downloadChunks = len(downloadChunks)
n_threads = min(n_concurrentConnections, n_downloadChunks)
if (n_threads < n_concurrentConnections):
print('Number of connection threads was limited by download size.')
q = Queue()
completedDict = {}
writeDict = {}
delList = []
for df in downloadFiles:
writeDict[df['file_name']] = [0, int(df['file_size']/size_chunk)]
print('Starting {} connection threads.'.format(n_threads))
with progress.Bar(expected_size=size_downloadFiles) as bar:
tp = ThreadPool(n_threads)
tp.imap_unordered(downloadChunk, downloadChunks)
current_size = 0
while current_size < size_downloadFiles:
if q.qsize() > 0:
f, chunk_id, content = q.get()
completedDict[(f['file_name'], chunk_id)] = (content, f['path'])
for writer in writeDict:
chunk = completedDict.get((writer, writeDict[writer][0]))
if chunk != None:
current_size += writeChunk(chunk, writer, writeDict[writer][0])
del completedDict[(writer, writeDict[writer][0])]
writeDict[writer][0] += 1
if writeDict[writer][0] > writeDict[writer][1]:
delList.append(writer)
if len(delList) > 0:
for item in delList:
del writeDict[item]
delList.clear()
bar.show(current_size)
tp.close()
tp.join()
print('Download complete.')
def _get_hardware_threaded(host):
hardware_functions = [
collect_win_bios_stats, collect_win_disk_stats, collect_win_mem_stats,
collect_win_network_stats, collect_win_cpu_stats
]
hardware_info = {}
queue = Queue()
list_of_processes = []
for hardware in hardware_functions:
process = _Process(target=hardware, args=(
host,
1,
queue,
))
list_of_processes.append(process)
process.start()
for process in list_of_processes:
process.join()
hardware_info.update(queue.get())
return hardware_info
def proxy_thr(fun):
"""
多线程获取代理地址
:param fun: 处理url函数，用于获取页面内的proxy
:return:
"""
q = Queue()
pool = Pool(40)
proxys = []
i = 1
for u in proxy_url_list():
pool.apply_async(fun, (q, u))
pool.close()
pool.join()
while 1:
if q.empty():
break
s = q.get().split("\n")
proxys += s
proxys = set(proxys)
return list(proxys)
class mmonly:
def __init__(self):
self.ua = UserAgent()
self.headers = {}
self.q1 = Queue(300)
self.q2 = Queue(1000)
self.lock = Lock()
# self.path = 'D:/IMG/'
self.main_page_urls = []
self.subpageurls = []
conn = sqlite3.connect('mmonly.db')
conn.isolation_level = None
try:
conn.execute(
'''create table subpageurl(url text primary key not null)''')
conn.execute(
'''create table imgurl(url text primary key not null)''')
except (Exception) as e:
print('创建表：{}'.format(e).decode('utf-8').encode(type))
finally:
conn.close()
self.rootpath = os.getcwd().replace('\\', '/')
self.path = os.path.join(self.rootpath, 'imges/')
if not os.path.exists(self.path):
os.mkdir(self.path)
def get_mainpage_urls(self, inurl):
# 得到所有主页url
self.headers['User-Agent'] = self.ua.random
try:
req = requests.get(inurl, headers=self.headers, timeout=10)
req.encoding = 'gbk'
cont = req.text
content = pq(cont)
elem = list(content('div #pageNum').children('a').items())
for ele in elem:
if ele.text() == '末页':
pgnum = int(ele.attr('href').split('_')[-1].split('.')[0])
spurl = inurl.split('_')
for i in range(1, pgnum + 1):
self.main_page_urls.append('{}_{}_{}.html'.format(
spurl[0], spurl[1], str(i)))
print('主页计算完毕！！'.decode('utf-8').encode(type))
except (Exception) as e:
self.lock.acquire()
print('主页读取错误：{}'.format(e).decode('utf-8').encode(type))
self.lock.release()
return
def get_subpage_urls(self, inurl):
# 得到所有子页面url
self.headers['User-Agent'] = self.ua.random
try:
req = requests.get(inurl, headers=self.headers, timeout=10)
req.encoding = 'gbk'
cont = req.text
content = pq(cont)
elems = list(content('div .ABox').children('a').items())
for ele in elems:
url = ele.attr('href')
self.q1.put(url)
print('取得子页面地址：{}'.format(url).decode('utf-8').encode(type))
except (Exception) as e:
self.lock.acquire()
print('遍历主页面读取错误：{}'.format(e).decode('utf-8').encode(type))
self.lock.release()
return
def savesuburl(self):
# 将子页面url存入数据库subpageurl表中
while 1:
try:
suburl = self.q1.get(timeout=20)
self.subpageurls.append(suburl)
print('列表存入子页面：{}'.format(suburl).decode('utf-8').encode(type))
except (Exception) as e:
print('读取子页面url：{}'.format(e).decode('utf-8').encode(type))
time.sleep(2)
if self.q1.empty():
time.sleep(2)
if self.q1.empty():
break
conn = sqlite3.connect('mmonly.db')
cur = conn.cursor()
time.sleep(4)
print('开始将子页面url写入数据库'.decode('utf-8').encode(type))
for date in self.subpageurls:
try:
cur.execute('insert into subpageurl values(?)', (date, ))
print('写入：{}'.format(date).decode('utf-8').encode(type))
except (Exception) as er:
print('写入数据库错误：{}'.format(er).decode('utf-8').encode(type))
conn.commit()
conn.close()
print('写入完毕！！'.decode('utf-8').encode(type))
def get_img_url(self, inurl):
# get图片地址
self.headers['User-Agent'] = self.ua.random
try:
req = requests.get(inurl, headers=self.headers, timeout=10)
time.sleep(0.2)
req.encoding = 'gbk'
cont = req.text
content = pq(cont)
imgnum = int(content('.totalpage').text())
urlsp = '.'.join(inurl.split('.')[:-1])
for n in range(1, imgnum + 1):
imgpage = '{}_{}.html'.format(urlsp, n)
self.headers['User-Agent'] = self.ua.random
try:
req = requests.get(imgpage,
headers=self.headers,
timeout=10)
time.sleep(0.3)
req.encoding = 'gbk'
cont = req.text
content = pq(cont)
imgurl = content('.down-btn').attr('href')
self.q2.put(imgurl)
except (Exception) as ee:
print('get图片url错误：{}'.format(ee).decode('utf-8').encode(
type))
print(
'get图片url：{}'.format(imgurl).decode('utf-8').encode(type))
except (Exception) as e:
print('get图片页面地址错误：{}'.format(e).decode('utf-8').encode(type))
return
def download(self, inurl):
# 下载图片
# inurl = q.get(timeout=10)
na = inurl.split('/')
imgname = '{}{}'.format(na[-2], na[-1])
imgpath = '{}{}'.format(self.path, imgname)
statu = os.path.exists(imgpath)
if not statu:
self.headers['User-Agent'] = self.ua.random
try:
req = requests.get(inurl, headers=self.headers,
timeout=8).content
with open(imgpath, 'wb') as f:
f.write(req)
self.lock.acquire()
print('下载图片：{}'.format(imgname).decode('utf-8').encode(type))
self.lock.release()
except (Exception) as e:
self.lock.acquire()
print('下载错误：{}'.format(e).decode('utf-8').encode(type))
self.lock.release()
else:
self.lock.acquire()
print('重复图片：{}'.format(imgname).decode('utf-8').encode(type))
self.lock.release()
def run(self, inurl):
ch = eval(
input('输入1表示采集页面\n输入2表示下载图片\n输入3退出程序\n输入：'.decode('utf-8').encode(
type)))
if ch == 1:
self.get_mainpage_urls(inurl)
time.sleep(4)
pool1 = Pool(20)
for mainurl in self.main_page_urls:
pool1.apply_async(self.get_subpage_urls, (mainurl, ))
time.sleep(1)
self.savesuburl()
pool1.close()
pool1.join()
print('子页面采集完毕！！！'.decode('utf-8').encode(type))
self.run('http://www.mmonly.cc/mmtp/list_9_2.html')
elif ch == 2:
conn = sqlite3.connect('mmonly.db')
cur = conn.cursor()
pool2 = Pool(10)
pool3 = Pool(30)
cur.execute('select * from subpageurl')
suburls = cur.fetchall()
while 1:
for nn in range(200):
try:
for i in suburls:
pool2.apply_async(self.get_img_url, i)
cur.execute('delete from subpageurl where url=?',
i)
while 1:
img = self.q2.get(timeout=20)
pool3.apply_async(self.download, (img, ))
except (Exception) as e:
print('数据库读取子页面url：{}'.format(e).decode(
'utf-8').encode(type))
time.sleep(2)
if self.q2.empty():
time.sleep(2)
if self.q2.empty():
break
conn.commit()
conn.close()
conn = sqlite3.connect('mmonly.db')
cur = conn.cursor()
cur.execute('select * from subpageurl')
suburls = cur.fetchall()
time.sleep(2)
if self.q2.empty():
time.sleep(2)
if self.q2.empty():
break
pool3.close()
pool2.close()
pool3.join()
pool2.join()
else:
print('结束程序！'.decode('utf-8').encode(type))
class ParallelDownloader(URL_Fetcher):
'Parallel threaded web page downloader'
def __init__(self,
db_name,
proc_count,
site_base_url,
fUseCache=True,
fCacheSearchPages=True,
fUseCookies=False,
timeout=secHTTP_WAIT_TIMEOUT,
search_proc_count=2,
proxies=None):
self.proxies = proxies
self.queue = Queue()
self.fSaveSearchPages = fCacheSearchPages
self.site_base_url = site_base_url
self.pool = Pool(processes=proc_count)
self.search_queue = Queue()
self.url_extract_pool = Pool(processes=search_proc_count)
URL_Fetcher.__init__(self,
db_name,
fUseCache,
fUseCookies,
timeout=timeout,
proxies=proxies)
def process_urls_from_search_queue(self):
while not self.search_queue.empty():
search_page_url = self.search_queue.get()
# logOut('search pages queue size: %d'%self.search_queue.qsize())
logDbg('search page: %s' % search_page_url)
search_page = self.get_page(search_page_url,
fUseCache=self.fSaveSearchPages)
rel_urls = extract_data_xpath(search_page, self.url_extract_xpath)
#rel_urls = self.extract_page_xpath(self.url_extract_xpath, search_page_url)
# logOut('URLs from %s extracted'%search_page_url)
logOut('%d urls extracted from [%s]. Queuing...' %
(len(rel_urls), search_page_url))
logDbg('Extracted urls: %s. Queuing to download...' % rel_urls)
list(map(self.queue.put, self.prefix_site_base_url(rel_urls)))
self.queue.put(None)
self.postprocess_search_page_list(rel_urls, search_page)
def queue_pages(self, url_list):
list(map(self.queue.put, url_list))
# признак завершения очереди заданий
self.queue.put(None)
def postprocess_search_page_list(self, url, page):
pass
def prefix_site_base_url(self, rel_urls):
return [self.site_base_url + url for url in rel_urls]
def process_pages(self, page_processor, *add_processor_args):
self.page_processor = page_processor
self.add_pprocessor_args = add_processor_args
self.pool.apply(self.process_page)
def process_page(self):
while True:
url = self.queue.get()
logDbg('Url got from queue: %s' % url)
if not url:
break
page = self.get_page(url) #, proxies=self.proxies
#logOut('pp_arg_list: [%s]'%pp_arg_list)
if page:
self.page_processor(url, page, *self.add_pprocessor_args)
class tizyoutubeproxy(object):
"""A class that accesses YouTube, retrieves stream URLs and creates and manages
a playback queue.
"""
def __init__(self):
self.queue = list()
self.queue_index = -1
self.play_queue_order = list()
self.play_modes = TizEnumeration(["NORMAL", "SHUFFLE"])
self.current_play_mode = self.play_modes.NORMAL
self.now_playing_stream = None
# Create multiprocess queues
self.task_queue = Queue()
self.done_queue = Queue()
# Workers
self.workers = list()
def set_play_mode(self, mode):
""" Set the playback mode.
:param mode: current valid values are "NORMAL" and "SHUFFLE"
"""
self.current_play_mode = getattr(self.play_modes, mode)
self.__update_play_queue_order()
def enqueue_audio_stream(self, arg):
"""Add the audio stream of a YouTube video to the
playback queue.
:param arg: a search string
"""
logging.info('arg : %s', arg)
try:
yt_video = pafy.new(arg)
yt_audio = yt_video.getbestaudio(preftype="webm")
if not yt_audio:
raise ValueError(str("No WebM audio stream for : %s" % arg))
yt_info = VideoInfo(ytid=arg, title=yt_audio.title)
self.add_to_playback_queue(audio=yt_audio,
video=yt_video,
info=yt_info)
self.__update_play_queue_order()
except ValueError:
raise ValueError(str("Video not found : %s" % arg))
def enqueue_audio_playlist(self, arg):
"""Add all audio streams in a YouTube playlist to the playback queue.
:param arg: a YouTube playlist id
"""
logging.info('arg : %s', arg)
try:
count = len(self.queue)
playlist = pafy.get_playlist2(arg)
if len(playlist) > 0:
for yt_video in playlist:
self.add_to_playback_queue(video=yt_video, \
info=VideoInfo(ytid=yt_video.videoid, \
title=yt_video.title))
if count == len(self.queue):
raise ValueError
self.__update_play_queue_order()
except ValueError:
raise ValueError(str("Playlist not found : %s" % arg))
def enqueue_audio_search(self, arg):
"""Search YouTube and add the audio streams to the
playback queue.
:param arg: a search string
"""
logging.info('arg : %s', arg)
try:
query = generate_search_query(arg)
wdata = pafy.call_gdata('search', query)
wdata2 = wdata
count = 0
while True:
for track_info in get_tracks_from_json(wdata2):
self.add_to_playback_queue(info=track_info)
count += 1
if count > 100:
break
if not wdata2.get('nextPageToken'):
break
query['pageToken'] = wdata2['nextPageToken']
wdata2 = pafy.call_gdata('search', query)
self.__update_play_queue_order()
except ValueError:
raise ValueError(str("Could not find any mixes : %s" % arg))
def enqueue_audio_mix(self, arg, feelinglucky=True):
"""Obtain a YouTube mix associated to a given video id or url and add all audio
streams in the mix playlist to the playback queue.
:param arg: a YouTube video id
:param feelinglucky: If True, it will perform another YouTube search to find
alternatives if the original mix cannot be found.
"""
logging.info('arg : %s', arg)
yt_video = None
try:
count = len(self.queue)
yt_video = pafy.new(arg)
playlist = yt_video.mix
if len(playlist) > 0:
for yt_video in playlist:
video_id = yt_video.videoid
video_title = yt_video.title
yt_info = VideoInfo(ytid=video_id, title=video_title)
self.add_to_playback_queue(video=yt_video, info=yt_info)
if count == len(self.queue):
raise ValueError
self.__update_play_queue_order()
except IndexError:
if not feelinglucky:
raise ValueError
else:
print_wrn("[YouTube] Could not find a mix for '{0}'. "\
"Searching YouTube instead. Feeling lucky?." \
.format(arg.encode('utf-8')))
if yt_video.title:
self.enqueue_audio_search(yt_video.title)
else:
self.enqueue_audio_stream(arg)
def enqueue_audio_mix_search(self, arg):
"""Obtain a YouTube mix associated to a given textual search and add all the
audio streams in the mix playlist to the playback queue.
:param arg: a search string
"""
logging.info('arg : %s', arg)
try:
query = generate_search_query(arg)
wdata = pafy.call_gdata('search', query)
wdata2 = wdata
count = len(self.queue)
for track_info in get_tracks_from_json(wdata2):
if track_info and track_info.ytid:
try:
self.enqueue_audio_mix(track_info.ytid,
feelinglucky=False)
break
except ValueError:
logging.info(
'Could not find a mix. Trying another video')
if count == len(self.queue):
raise ValueError
except ValueError:
raise ValueError(str("Could not find any mixes : %s" % arg))
def current_audio_stream_title(self):
""" Retrieve the current stream's title.
"""
stream = self.now_playing_stream
title = ''
if stream:
title = to_ascii(stream['a'].title).encode("utf-8")
return title
def current_audio_stream_author(self):
""" Retrieve the current stream's author.
"""
stream = self.now_playing_stream
author = ''
if stream:
author = to_ascii(stream['v'].author).encode("utf-8")
return author
def current_audio_stream_file_size(self):
""" Retrieve the current stream's file size.
"""
stream = self.now_playing_stream
size = 0
if stream:
size = stream['a'].get_filesize()
return size
def current_audio_stream_duration(self):
""" Retrieve the current stream's duration.
"""
stream = self.now_playing_stream
duration = ''
if stream:
duration = to_ascii(stream['v'].duration).encode("utf-8")
return duration
def current_audio_stream_bitrate(self):
""" Retrieve the current stream's bitrate.
"""
stream = self.now_playing_stream
bitrate = ''
if stream:
bitrate = stream['a'].bitrate
return bitrate
def current_audio_stream_view_count(self):
""" Retrieve the current stream's view count.
"""
stream = self.now_playing_stream
viewcount = 0
if stream:
viewcount = stream['v'].viewcount
return viewcount
def current_audio_stream_description(self):
""" Retrieve the current stream's description.
"""
stream = self.now_playing_stream
description = ''
if stream:
description = to_ascii(stream['v'].description).encode("utf-8")
return description
def current_audio_stream_file_extension(self):
""" Retrieve the current stream's file extension.
"""
stream = self.now_playing_stream
file_extension = ''
if stream:
file_extension = to_ascii(stream['a'].extension).encode("utf-8")
return file_extension
def current_audio_stream_video_id(self):
""" Retrieve the current stream's video id.
"""
stream = self.now_playing_stream
video_id = ''
if stream:
video_id = to_ascii(stream['i'].ytid).encode("utf-8")
return video_id
def current_audio_stream_published(self):
""" Retrieve the current stream's upload date and time.
"""
stream = self.now_playing_stream
if stream:
published = to_ascii(stream['v'].published).encode("utf-8")
return published
def current_audio_stream_queue_index_and_queue_length(self):
""" Retrieve index in the queue (starting from 1) of the current stream and the
length of the playback queue.
"""
return self.queue_index + 1, len(self.queue)
def clear_queue(self):
""" Clears the playback queue.
"""
self.queue = list()
self.queue_index = -1
def remove_current_url(self):
"""Remove the currently active url from the playback queue.
"""
logging.info("")
if len(self.queue) and self.queue_index:
stream = self.queue[self.queue_index]
print_nfo("[YouTube] [Stream] '{0}' removed." \
.format(to_ascii(stream['i'].title).encode("utf-8")))
del self.queue[self.queue_index]
self.queue_index -= 1
if self.queue_index < 0:
self.queue_index = 0
self.__update_play_queue_order()
def next_url(self):
""" Retrieve the url of the next stream in the playback queue.
"""
logging.info("")
try:
if len(self.queue):
self.queue_index += 1
if (self.queue_index < len(self.queue)) \
and (self.queue_index >= 0):
next_stream = self.queue[self.play_queue_order \
[self.queue_index]]
return self.__retrieve_stream_url(
next_stream, self.queue_index).rstrip()
else:
self.queue_index = -1
return self.next_url()
else:
return ''
except (KeyError, AttributeError):
# TODO: We don't remove this for now
# del self.queue[self.queue_index]
logging.info("exception")
return self.next_url()
def prev_url(self):
""" Retrieve the url of the previous stream in the playback queue.
"""
logging.info("")
try:
if len(self.queue):
self.queue_index -= 1
if (self.queue_index < len(self.queue)) \
and (self.queue_index >= 0):
prev_stream = self.queue[self.play_queue_order \
[self.queue_index]]
return self.__retrieve_stream_url(
prev_stream, self.queue_index).rstrip()
else:
self.queue_index = len(self.queue)
return self.prev_url()
else:
return ''
except (KeyError, AttributeError):
# TODO: We don't remove this for now
# del self.queue[self.queue_index]
logging.info("exception")
return self.prev_url()
def __update_play_queue_order(self):
""" Update the queue playback order.
A sequential order is applied if the current play mode is "NORMAL" or a
random order if current play mode is "SHUFFLE"
"""
total_streams = len(self.queue)
if total_streams:
if not len(self.play_queue_order):
# Create a sequential play order, if empty
self.play_queue_order = range(total_streams)
if self.current_play_mode == self.play_modes.SHUFFLE:
random.shuffle(self.play_queue_order)
print_nfo("[YouTube] [Streams in queue] '{0}'." \
.format(total_streams))
def __retrieve_stream_url(self, stream, queue_index):
""" Retrieve a stream url
"""
try:
if not len(self.workers):
for _ in range(WORKER_PROCESSES):
proc = Process(target=obtain_stream, \
args=(self.task_queue, \
self.done_queue)).start()
self.workers.append(proc)
while not self.done_queue.empty():
stream = self.done_queue.get()
self.queue[stream['q']] = stream
stream = self.queue[queue_index]
if not stream.get('v') or not stream.get('a'):
logging.info("ytid : %s", stream['i'].ytid)
video = stream.get('v')
if not video:
video = pafy.new(stream['i'].ytid)
audio = video.getbestaudio(preftype="webm")
if not audio:
logging.info("no suitable audio found")
raise AttributeError()
stream.update({'a': audio, 'v': video})
# streams = stream.get('v').audiostreams[::-1]
# pprint.pprint(streams)
# dump_stream_info(streams)
self.now_playing_stream = stream
return stream['a'].url.encode("utf-8")
except AttributeError:
logging.info("Could not retrieve the stream url!")
raise
def add_to_playback_queue(self, audio=None, video=None, info=None):
""" Add to the playback queue. """
if audio:
print_nfo("[YouTube] [Stream] '{0}' [{1}]." \
.format(to_ascii(audio.title).encode("utf-8"), \
to_ascii(audio.extension)))
if info:
print_nfo("[YouTube] [Stream] '{0}'." \
.format(to_ascii(info.title).encode("utf-8")))
queue_index = len(self.queue)
self.task_queue.put(dict(a=audio, v=video, i=info, q=queue_index))
self.queue.append(dict(a=audio, v=video, i=info, q=queue_index))
class tizyoutubeproxy(object):
"""A class that accesses YouTube, retrieves stream URLs and creates and manages
a playback queue.
"""
def __init__(self):
self.queue = list()
self.queue_index = -1
self.play_queue_order = list()
self.play_modes = TizEnumeration(["NORMAL", "SHUFFLE"])
self.current_play_mode = self.play_modes.NORMAL
self.now_playing_stream = None
# Create multiprocess queues
self.task_queue = Queue()
self.done_queue = Queue()
# Workers
self.workers = list()
def set_play_mode(self, mode):
""" Set the playback mode.
:param mode: current valid values are "NORMAL" and "SHUFFLE"
"""
self.current_play_mode = getattr(self.play_modes, mode)
self.__update_play_queue_order()
def enqueue_audio_stream(self, arg):
"""Add the audio stream of a YouTube video to the
playback queue.
:param arg: a search string
"""
logging.info('arg : %s', arg)
try:
yt_video = pafy.new(arg)
yt_audio = yt_video.getbestaudio(preftype="webm")
if not yt_audio:
raise ValueError(str("No WebM audio stream for : %s" % arg))
yt_info = VideoInfo(ytid=arg, title=yt_audio.title)
self.add_to_playback_queue(audio=yt_audio, video=yt_video, info=yt_info)
self.__update_play_queue_order()
except ValueError:
raise ValueError(str("Video not found : %s" % arg))
def enqueue_audio_playlist(self, arg):
"""Add all audio streams in a YouTube playlist to the playback queue.
:param arg: a YouTube playlist id
"""
logging.info('arg : %s', arg)
try:
count = len(self.queue)
playlist = pafy.get_playlist2(arg)
if len(playlist) > 0:
for yt_video in playlist:
self.add_to_playback_queue(video=yt_video, \
info=VideoInfo(ytid=yt_video.videoid, \
title=yt_video.title))
if count == len(self.queue):
raise ValueError
self.__update_play_queue_order()
except ValueError:
raise ValueError(str("Playlist not found : %s" % arg))
def enqueue_audio_search(self, arg):
"""Search YouTube and add the audio streams to the
playback queue.
:param arg: a search string
"""
logging.info('arg : %s', arg)
try:
query = generate_search_query(arg)
wdata = pafy.call_gdata('search', query)
wdata2 = wdata
count = 0
while True:
for track_info in get_tracks_from_json(wdata2):
self.add_to_playback_queue(info=track_info)
count += 1
if count > 100:
break
if not wdata2.get('nextPageToken'):
break
query['pageToken'] = wdata2['nextPageToken']
wdata2 = pafy.call_gdata('search', query)
self.__update_play_queue_order()
except ValueError:
raise ValueError(str("Could not find any mixes : %s" % arg))
def enqueue_audio_mix(self, arg, feelinglucky=True):
"""Obtain a YouTube mix associated to a given video id or url and add all audio
streams in the mix playlist to the playback queue.
:param arg: a YouTube video id
:param feelinglucky: If True, it will perform another YouTube search to find
alternatives if the original mix cannot be found.
"""
logging.info('arg : %s', arg)
yt_video = None
try:
count = len(self.queue)
yt_video = pafy.new(arg)
playlist = yt_video.mix
if len(playlist) > 0:
for yt_video in playlist:
video_id = yt_video.videoid
video_title = yt_video.title
yt_info = VideoInfo(ytid=video_id, title=video_title)
self.add_to_playback_queue(video=yt_video, info=yt_info)
if count == len(self.queue):
raise ValueError
self.__update_play_queue_order()
except IndexError:
if not feelinglucky:
raise ValueError
else:
print_wrn("[YouTube] Could not find a mix for '{0}'. "\
"Searching YouTube instead. Feeling lucky?." \
.format(arg.encode('utf-8')))
if yt_video.title:
self.enqueue_audio_search(yt_video.title)
else:
self.enqueue_audio_stream(arg)
def enqueue_audio_mix_search(self, arg):
"""Obtain a YouTube mix associated to a given textual search and add all the
audio streams in the mix playlist to the playback queue.
:param arg: a search string
"""
logging.info('arg : %s', arg)
try:
query = generate_search_query(arg)
wdata = pafy.call_gdata('search', query)
wdata2 = wdata
count = len(self.queue)
for track_info in get_tracks_from_json(wdata2):
if track_info and track_info.ytid:
try:
self.enqueue_audio_mix(track_info.ytid, feelinglucky=False)
break
except ValueError:
logging.info('Could not find a mix. Trying another video')
if count == len(self.queue):
raise ValueError
except ValueError:
raise ValueError(str("Could not find any mixes : %s" % arg))
def enqueue_audio_channel_uploads(self, arg):
"""Add all audio streams in a YouTube channel to the playback queue.
:param arg: a YouTube channel url
"""
logging.info('arg : %s', arg)
try:
count = len(self.queue)
channel = pafy.get_channel(arg)
if channel:
for yt_video in channel.uploads:
self.add_to_playback_queue(video=yt_video, \
info=VideoInfo(ytid=yt_video.videoid, \
title=yt_video.title))
if count == len(self.queue):
raise ValueError
self.__update_play_queue_order()
except ValueError:
raise ValueError(str("Channel not found : %s" % arg))
def enqueue_audio_channel_playlist(self, channel_name, playlist_name):
"""Search a playlist within a channel and if found, adds all the audio streams
to the playback queue.
:param arg: a YouTube playlist id
"""
logging.info('args : %s - %s', channel_name, playlist_name)
try:
count = len(self.queue)
channel = pafy.get_channel(channel_name)
if channel:
pl_dict = dict()
pl_titles = list()
pl_name = ''
playlist = None
for pl in channel.playlists:
print_nfo("[YouTube] [Playlist] '{0}'." \
.format(to_ascii(pl.title)))
if fuzz.partial_ratio(playlist_name, pl.title) > 50:
pl_dict[pl.title] = pl
pl_titles.append(pl.title)
if len(pl_titles) > 1:
pl_name = process.extractOne(playlist_name, pl_titles)[0]
playlist = pl_dict[pl_name]
elif len(pl_titles) == 1:
pl_name = pl_titles[0]
playlist = pl_dict[pl_name]
if pl_name:
if pl_name.lower() != playlist_name.lower():
print_wrn("[YouTube] Playlist '{0}' not found. " \
"Playing '{1}' instead." \
.format(to_ascii(playlist_name), \
to_ascii(pl_name)))
for yt_video in playlist:
self.add_to_playback_queue(video=yt_video, \
info=VideoInfo(ytid=yt_video.videoid, \
title=yt_video.title))
if count == len(self.queue):
raise ValueError
self.__update_play_queue_order()
except ValueError:
raise ValueError(str("Channel not found : %s" % channel_name))
def current_audio_stream_title(self):
""" Retrieve the current stream's title.
"""
stream = self.now_playing_stream
title = ''
if stream:
title = to_ascii(stream['a'].title).encode("utf-8")
return title
def current_audio_stream_author(self):
""" Retrieve the current stream's author.
"""
stream = self.now_playing_stream
author = ''
if stream:
author = to_ascii(stream['v'].author).encode("utf-8")
return author
def current_audio_stream_file_size(self):
""" Retrieve the current stream's file size.
"""
stream = self.now_playing_stream
size = 0
if stream:
size = stream['a'].get_filesize()
return size
def current_audio_stream_duration(self):
""" Retrieve the current stream's duration.
"""
stream = self.now_playing_stream
duration = ''
if stream:
duration = to_ascii(stream['v'].duration).encode("utf-8")
return duration
def current_audio_stream_bitrate(self):
""" Retrieve the current stream's bitrate.
"""
stream = self.now_playing_stream
bitrate = ''
if stream:
bitrate = stream['a'].bitrate
return bitrate
def current_audio_stream_view_count(self):
""" Retrieve the current stream's view count.
"""
stream = self.now_playing_stream
viewcount = 0
if stream:
viewcount = stream['v'].viewcount
return viewcount
def current_audio_stream_description(self):
""" Retrieve the current stream's description.
"""
stream = self.now_playing_stream
description = ''
if stream:
description = to_ascii(stream['v'].description).encode("utf-8")
return description
def current_audio_stream_file_extension(self):
""" Retrieve the current stream's file extension.
"""
stream = self.now_playing_stream
file_extension = ''
if stream:
file_extension = to_ascii(stream['a'].extension).encode("utf-8")
return file_extension
def current_audio_stream_video_id(self):
""" Retrieve the current stream's video id.
"""
stream = self.now_playing_stream
video_id = ''
if stream:
video_id = to_ascii(stream['i'].ytid).encode("utf-8")
return video_id
def current_audio_stream_published(self):
""" Retrieve the current stream's upload date and time.
"""
stream = self.now_playing_stream
if stream:
published = to_ascii(stream['v'].published).encode("utf-8")
return published
def current_audio_stream_queue_index_and_queue_length(self):
""" Retrieve index in the queue (starting from 1) of the current stream and the
length of the playback queue.
"""
return self.queue_index + 1, len(self.queue)
def clear_queue(self):
""" Clears the playback queue.
"""
self.queue = list()
self.queue_index = -1
def remove_current_url(self):
"""Remove the currently active url from the playback queue.
"""
logging.info("")
if len(self.queue) and self.queue_index:
stream = self.queue[self.queue_index]
print_nfo("[YouTube] [Stream] '{0}' removed." \
.format(to_ascii(stream['i'].title).encode("utf-8")))
del self.queue[self.queue_index]
self.queue_index -= 1
if self.queue_index < 0:
self.queue_index = 0
self.__update_play_queue_order()
def next_url(self):
""" Retrieve the url of the next stream in the playback queue.
"""
logging.info("")
try:
if len(self.queue):
self.queue_index += 1
if (self.queue_index < len(self.queue)) \
and (self.queue_index >= 0):
next_stream = self.queue[self.play_queue_order \
[self.queue_index]]
return self.__retrieve_stream_url(next_stream, self.queue_index).rstrip()
else:
self.queue_index = -1
return self.next_url()
else:
return ''
except (KeyError, AttributeError):
# TODO: We don't remove this for now
# del self.queue[self.queue_index]
logging.info("KeyError, or AttributeError exception")
return self.next_url()
except (IOError):
# Remove this video
del self.queue[self.queue_index]
logging.info("IOError exception")
return self.next_url()
def prev_url(self):
""" Retrieve the url of the previous stream in the playback queue.
"""
logging.info("")
try:
if len(self.queue):
self.queue_index -= 1
if (self.queue_index < len(self.queue)) \
and (self.queue_index >= 0):
prev_stream = self.queue[self.play_queue_order \
[self.queue_index]]
return self.__retrieve_stream_url(prev_stream, self.queue_index).rstrip()
else:
self.queue_index = len(self.queue)
return self.prev_url()
else:
return ''
except (KeyError, AttributeError):
# TODO: We don't remove this for now
# del self.queue[self.queue_index]
logging.info("exception")
return self.prev_url()
except (IOError):
# Remove this video
del self.queue[self.queue_index]
logging.info("IOError exception")
return self.next_url()
def __update_play_queue_order(self):
""" Update the queue playback order.
A sequential order is applied if the current play mode is "NORMAL" or a
random order if current play mode is "SHUFFLE"
"""
total_streams = len(self.queue)
if total_streams:
if not len(self.play_queue_order):
# Create a sequential play order, if empty
self.play_queue_order = range(total_streams)
if self.current_play_mode == self.play_modes.SHUFFLE:
random.shuffle(self.play_queue_order)
print_nfo("[YouTube] [Streams in queue] '{0}'." \
.format(total_streams))
def __retrieve_stream_url(self, stream, queue_index):
""" Retrieve a stream url
"""
try:
if not len(self.workers):
for _ in range(WORKER_PROCESSES):
proc = Process(target=obtain_stream, \
args=(self.task_queue, \
self.done_queue)).start()
self.workers.append(proc)
while not self.done_queue.empty():
stream = self.done_queue.get()
self.queue[stream['q']] = stream
stream = self.queue[queue_index]
if not stream.get('v') or not stream.get('a'):
logging.info("ytid : %s", stream['i'].ytid)
video = stream.get('v')
if not video:
video = pafy.new(stream['i'].ytid)
audio = video.getbestaudio(preftype="webm")
if not audio:
logging.info("no suitable audio found")
raise AttributeError()
stream.update({'a': audio, 'v': video})
# streams = stream.get('v').audiostreams[::-1]
# pprint.pprint(streams)
# dump_stream_info(streams)
self.now_playing_stream = stream
return stream['a'].url.encode("utf-8")
except AttributeError:
logging.info("Could not retrieve the stream url!")
raise
def add_to_playback_queue(self, audio=None, video=None, info=None):
""" Add to the playback queue. """
if audio:
print_nfo("[YouTube] [Stream] '{0}' [{1}]." \
.format(to_ascii(audio.title).encode("utf-8"), \
to_ascii(audio.extension)))
if info:
print_nfo("[YouTube] [Stream] '{0}'." \
.format(to_ascii(info.title).encode("utf-8")))
queue_index = len(self.queue)
self.task_queue.put(dict(a=audio, v=video, i=info, q=queue_index))
self.queue.append(
dict(a=audio, v=video, i=info, q=queue_index))
class EventEngine():
'''事件处理引擎'''
def __init__(self):
'''初始化'''
# 创建队列
self._queue = Queue()
# 是否触发时间
self._active = False
# 线程池
self._thread = [Process(target=self._run) for _ in range(5)]
self._workers_n = 5
# 执行
self._handlers = {}
def _run(self):
logger.info('Worker启动...')
while self._active:
try:
event = self._queue.get(block=True, timeout=0.3)
logger.debug('准备处理时间: %s' % event)
self._process(event)
except Empty:
pass
logger.info('Worker停止...')
return ''
def _process(self, event):
"""处理事件"""
# 检查是否存在对该事件进行监听的处理函数
handlers = self._handlers.get(event.type, [])
for handler in handlers:
try:
if callable(handler):
ret_events = handler(event)
if isinstance(ret_events, list) == False:
ret_events = [ret_events]
for ret_event in ret_events:
if isinstance(ret_event, Event):
self.trigger(ret_event)
except Exception as err:
logger.warning('事件: %s 运行错误: %s' % (event.type, err))
def start(self):
'''开始执行'''
# 后台运行
logger.info('启动后台线程')
self._active = True
for thread in self._thread:
thread.start()
def stop(self):
logger.info('关闭后台线程')
self._active = False
# 等待后台线程结束
for thread in self._thread:
thread.join(timeout=1)
def register(self, event_type, handler):
'''注册监听事件'''
logger.info('register event type: %s, handler: %s' % (event_type, handler))
handlers = self._handlers.pop(event_type, [])
if handler not in handlers:
handlers.append(handler)
self._handlers[event_type] = handlers
def unregister(self, event_type, handler):
'''取消注册监听事件'''
logger.debug('unregister event type: %s, handler: %s' % (event_type, handler))
handlers = self._handlers.pop(event_type, [])
if handler in handlers:
handlers.remove(handler)
if handlers:
self._handlers[event_type] = handlers
def trigger(self, event):
'''触发一个消息'''
logger.debug('trigger event: %s' % event)
self._queue.put(event)
return
def handle(self, event_type, context=None):
def deco(func):
@wraps(func)
def event_handler(event):
# TODO 检查参数，然后自动增加相应的数据进去
return func(event=event, context=context)
self.register(event_type, event_handler)
return func
return deco
def on_tick(self, context=None):
return self.handle('on_tick', context=context)
def on_open(self, context=None):
return self.handle('on_open', context=context)
def before_trade(self, context=None):
return self.handle('before_trade', context=context)
def pre_close(self, context=None):
return self.handle('pre_close', context=context)
def on_close(self, context=None):
return self.handle('on_close', context=context)
def after_close(self, context=None):
return self.handle('alfter_close', context=context)
class mzitu:
def __init__(self):
self.starturl = 'http://www.mzitu.com/all/'
self.ua = UserAgent()
self.qu = Queue(1000)
self.qu2 = Queue(1000)
self.suburls = []
self.oldmurls = []
conno = sqlite3.connect('mzituoldu.db')
try:
conno.execute('create table oldmurls(url text primary key)')
except (Exception) as e:
print('创建表ou:{}'.format(e))
conno.close()
conns = sqlite3.connect('mzitusubu.db')
try:
conns.execute(
'create table suburls(url text primary key,status int default 1)'
)
except (Exception) as e:
print('创建表su:{}'.format(e))
conns.close()
def gethtml(self, url):
headers = {}
headers['User-Agent'] = self.ua.random
headers['Referer'] = url
try:
resp = requests.get(url, headers=headers, timeout=8)
if resp.encoding == 'ISO-8859-1':
resp.encoding = requests.utils.get_encodings_from_content(
resp.content)[0]
html = pq(resp.text)
cont = resp.content
return html, cont
except (Exception) as e:
print('gethtml错误：{}'.format(e))
logging.warning('gethtml错误：{}'.format(e))
return None, None
def getmurl(self):
resp = self.gethtml(self.starturl)[0]
if resp != None:
elem = resp('.all')('a').items()
murls = {}
for ele in elem:
murl = ele.attr('href')
title = ele.text()
murls[murl] = title
return murls
else:
print('murl未get到网页内容！')
logging.warning('murl未get到网页内容！')
return None
def getsuburl(self, url):
resp = self.gethtml(url)[0]
if resp != None:
elem = resp('.pagenavi').children('a')
num = int(elem.eq(-2).children('span').text())
for i in range(1, num + 1):
suburl = '{}/{}'.format(url, i)
try:
self.qu.put(suburl, timeout=60)
except (Exception) as e:
print('qu.put Error:{}'.format(e))
self.qu2.put(url)
print('已get完：{}'.format(url))
else:
print('suburl未get到网页内容！')
logging.warning('suburl未get到网页内容！')
def getimgurl(self, url):
try:
resp = self.gethtml(url)[0]
if resp != None:
elem = resp('.main-image')('img').attr('src')
title = resp('.currentpath').next('h2').text()
return elem, title
else:
print('imgurl未get到网页内容！')
logging.warning('imgurl未get到网页内容！')
return None, None
except (Exception) as e:
print('getimgurlE:{}'.format(e))
logging.warning('getimgurlE:{}'.format(e))
def mkdir(self, rootpath=os.path.abspath('.'), dir='mzitu'):
global path
path = os.path.join(rootpath, dir)
if not os.path.exists(path):
os.mkdir(path)
def download(
self,
url,
):
imgurl = self.getimgurl(url)
global path
imgname = '{}{}{}'.format(imgurl[0].split('/')[-3],
imgurl[0].split('/')[-2],
imgurl[0].split('/')[-1])
imgpath = os.path.join(path, imgname)
statu = os.path.exists(imgpath)
headers = {}
if not statu:
for t in range(2):
headers['User-Agent'] = self.ua.random
headers['Referer'] = url
headers['Host'] = 'i.meizitu.net'
try:
resp = requests.get(imgurl[0], headers=headers, timeout=8)
with open(imgpath, 'wb') as f:
f.write(resp.content)
print('下载：{}'.format(imgurl[0]))
break
except (Exception) as e:
print('下载图片错误：{}'.format(e))
logging.warning('下载图片错误：{}'.format(e))
time.sleep(5)
else:
print('已下载：{}'.format(imgname))
def savesuburl(self):
i = 0
conn = sqlite3.connect('mzitusubu.db')
cur = conn.cursor()
suburls = []
while 1:
try:
suburl = (self.qu.get(timeout=20), )
suburls.append(suburl)
i += 1
if i >= 2000:
cur.executemany(
'insert or ignore into suburls(url) values(?)',
suburls)
conn.commit()
suburls = []
i = 0
except (Exception) as e:
print('suburls Error:{}'.format(e))
cur.executemany('insert or ignore into suburls(url) values(?)',
suburls)
conn.commit()
cur.close()
conn.close()
break
print('保存suburl')
def saveoldmurl(self):
conn = sqlite3.connect('mzituoldu.db')
cur = conn.cursor()
oldmurls = []
while 1:
try:
oldmurl = (self.qu2.get(timeout=20), )
oldmurls.append(oldmurl)
except (Exception) as e:
print('oldmurl Error:{}'.format(e))
cur.executemany(
'insert or ignore into oldmurls(url) values(?)', oldmurls)
conn.commit()
oldmurls = []
cur.close()
conn.close()
break
print('保存oldmurl')
def run(self):
self.mkdir(rootpath='H:/')
murls = self.getmurl()
pool1 = Pool(50)
pool1.apply_async(self.savesuburl)
pool1.apply_async(self.saveoldmurl)
conno = sqlite3.connect('mzituoldu.db')
oldurls = conno.execute('select url from oldmurls').fetchall()
conno.close()
for murl in murls:
omurl = (murl, )
if omurl not in oldurls:
pool1.apply_async(self.getsuburl(murl, ))
# oldurls.append(omurl)
else:
print('{}已经下过了'.format(murl))
pool1.close()
pool1.join()
pool2 = Pool(20)
conns = sqlite3.connect('mzitusubu.db')
cur = conns.cursor()
while 1:
surls = cur.execute(
'select url from suburls where status=1').fetchmany(300)
if surls == []:
break
oldsurls = []
for surl in surls:
pool2.apply_async(self.download, (surl[0], ))
oldsurls.append(surl)
time.sleep(8)
cur.executemany('update suburls set status=2 where url=?',
oldsurls)
conns.commit()
cur.close()
conns.close()
pool2.close()
pool2.join()
print('下载完成！！！！')
class DouYu:
"""Dou Yu dan mu spider."""
@staticmethod
def find_room():
"""Find the most hot 50 rooms of douyu."""
result = []
for page in range(0, 400, 30):
datas = json.loads(
requests.get("http://capi.douyucdn.cn/api/v1/live?"
"limit={}&offset={}".format(page + 30,
page)).text)["data"]
for data in datas:
result.append((data["game_name"], data["room_id"]))
return list(set(result))
@staticmethod
def log(url, err):
"""Log when error happened."""
with open("dy_log", "a") as f:
f.write(time.time + "\t" + str(url) + "\t" + str(err))
def __init__(self):
"""Initailization spider."""
self.fname = time.strftime("%Y%m%d_%H%M%S")
self.rooms = self.find_room()
self.pool_queue = Queue()
self.msg_queue = Queue()
def pool_join(self):
"""Processes pool join function."""
while 1:
p = self.pool_queue.get()
if p is None:
return
else:
try:
p.join()
except KeyboardInterrupt:
return
except BaseException:
pass
def record(self):
"""Record dan mu."""
with open("record/" + self.fname + ".danmu", "w") as f:
while 1:
msg = self.msg_queue.get()
if msg is None:
return
else:
try:
f.write(msg)
except KeyboardInterrupt:
return
except BaseException:
pass
def _run(self, room, game):
"""Run for multiprocessing."""
dmc = DanMuClient('http://www.douyu.com/{}'.format(room))
if not dmc.isValid():
self.log('http://www.douyu.com/{}'.format(room), "Url not valid")
@dmc.default
def danmu_fn(msg):
"""Dan mu function."""
res = "\t".join(
(str(room), str(game), str(time.time()), msg["MsgType"],
msg["NickName"], msg["Content"], "\n"))
self.msg_queue.put(res)
dmc.start(blockThread=False)
time.sleep(3600)
dmc.stop()
def run(self):
"""Run and get datas."""
print(self.fname, len(self.rooms))
p_re = Process(target=self.record)
pool = Process(target=self.pool_join)
pool.start()
p_re.start()
for (game, room) in self.rooms:
p = Process(target=self._run, args=(room, game))
p.start()
self.pool_queue.put(p)
self.pool_queue.put(None)
pool.join()
self.msg_queue.put(None)
return p_re, self.fname
class tizyoutubeproxy(object):
"""A class that accesses YouTube, retrieves stream URLs and creates and manages
a playback queue.
"""
def __init__(self, api_key=API_KEY):
self.queue = list()
self.queue_index = -1
self.play_queue_order = list()
self.play_modes = TizEnumeration(["NORMAL", "SHUFFLE"])
self.current_play_mode = self.play_modes.NORMAL
self.now_playing_stream = None
# Create multiprocess queues
self.task_queue = Queue()
self.done_queue = Queue()
# Workers
self.workers = list()
self.api_key = api_key if api_key != "" else API_KEY
pafy.set_api_key(self.api_key)
def set_play_mode(self, mode):
""" Set the playback mode.
:param mode: current valid values are "NORMAL" and "SHUFFLE"
"""
self.current_play_mode = getattr(self.play_modes, mode)
self._update_play_queue_order()
def enqueue_audio_stream(self, arg):
"""Add the audio stream of a YouTube video to the
playback queue.
:param arg: a search string
"""
logging.info("arg : %s", arg)
try:
print_msg("[YouTube] [Audio strean] : '{0}'. ".format(arg))
yt_search = MEMORY.cache(run_youtube_search)
yt_video = yt_search(arg)
yt_audio = yt_video.getbestaudio(preftype="webm")
if not yt_audio:
raise ValueError(str("No WebM audio stream for : %s" % arg))
yt_info = VideoInfo(ytid=arg, title=yt_audio.title)
self._add_to_playback_queue(audio=yt_audio,
video=yt_video,
info=yt_info)
self._update_play_queue_order()
except ValueError:
raise ValueError(str("Video not found : %s" % arg))
def enqueue_audio_playlist(self, arg):
"""Add all audio streams in a YouTube playlist to the playback queue.
:param arg: a YouTube playlist id
"""
logging.info("arg : %s", arg)
try:
print_msg("[YouTube] [Audio playlist] : '{0}'. ".format(arg))
count = len(self.queue)
yt_pl_search = MEMORY.cache(run_youtube_playlist_search)
playlist = yt_pl_search(arg)
if len(playlist) > 0:
for yt_video in playlist:
self._add_to_playback_queue(
video=yt_video,
info=VideoInfo(ytid=yt_video.videoid,
title=yt_video.title),
)
if count == len(self.queue):
raise ValueError
self._update_play_queue_order()
except ValueError:
raise ValueError(str("Playlist not found : %s" % arg))
def enqueue_audio_search(self, arg):
"""Search YouTube and add the audio streams to the
playback queue.
:param arg: a search string
"""
logging.info("arg : %s", arg)
try:
print_msg("[YouTube] [Audio search] : '{0}'. ".format(arg))
yt_dt_search = MEMORY.cache(run_youtube_data_search)
query = generate_search_query(arg, self.api_key)
wdata = yt_dt_search("search", query)
wdata2 = wdata
count = 0
while True:
for track_info in get_tracks_from_json(wdata2):
self._add_to_playback_queue(info=track_info)
count += 1
if count > 100:
break
if not wdata2.get("nextPageToken"):
break
query["pageToken"] = wdata2["nextPageToken"]
wdata2 = yt_dt_search("search", query)
self._update_play_queue_order()
except ValueError:
raise ValueError(str("Could not find any mixes : %s" % arg))
def enqueue_audio_mix(self, arg, feelinglucky=True):
"""Obtain a YouTube mix associated to a given video id or url and add all audio
streams in the mix playlist to the playback queue.
:param arg: a YouTube video id
:param feelinglucky: If True, it will perform another YouTube search to find
alternatives if the original mix cannot be found.
"""
logging.info("arg : %s", arg)
yt_video = None
try:
print_msg("[YouTube] [Audio mix] : '{0}'. ".format(arg))
count = len(self.queue)
yt_search = MEMORY.cache(run_youtube_search)
yt_video = yt_search(arg)
playlist = yt_video.mix
if len(playlist) > 0:
for yt_video in playlist:
video_id = yt_video.videoid
video_title = yt_video.title
yt_info = VideoInfo(ytid=video_id, title=video_title)
self._add_to_playback_queue(video=yt_video, info=yt_info)
if count == len(self.queue):
raise ValueError
self._update_play_queue_order()
except IndexError:
if not feelinglucky:
raise ValueError
else:
print_adv("[YouTube] Could not find a mix for '{0}'. "
"Searching YouTube instead. Feeling lucky?.".format(
arg.encode("utf-8")))
if yt_video.title:
self.enqueue_audio_search(yt_video.title)
else:
self.enqueue_audio_stream(arg)
def enqueue_audio_mix_search(self, arg):
"""Obtain a YouTube mix associated to a given textual search and add all the
audio streams in the mix playlist to the playback queue.
:param arg: a search string
"""
logging.info("arg : %s", arg)
try:
print_msg("[YouTube] [Audio mix search] : '{0}'. ".format(arg))
yt_dt_search = MEMORY.cache(run_youtube_data_search)
wdata = yt_dt_search("search",
generate_search_query(arg, self.api_key))
wdata2 = wdata
count = len(self.queue)
for track_info in get_tracks_from_json(wdata2):
if track_info and track_info.ytid:
try:
self.enqueue_audio_mix(track_info.ytid,
feelinglucky=False)
break
except ValueError:
logging.info(
"Could not find a mix. Trying another video")
if count == len(self.queue):
raise ValueError
except ValueError:
raise ValueError(str("Could not find any mixes : %s" % arg))
def enqueue_audio_channel_uploads(self, arg):
"""Add all audio streams in a YouTube channel to the playback queue.
:param arg: a YouTube channel url
"""
logging.info("arg : %s", arg)
try:
print_msg(
"[YouTube] [Audio channel uploads] : '{0}'. ".format(arg))
count = len(self.queue)
yt_ch_search = MEMORY.cache(run_youtube_channel_search)
channel = yt_ch_search(arg)
if channel:
for yt_video in channel.uploads:
self._add_to_playback_queue(
video=yt_video,
info=VideoInfo(ytid=yt_video.videoid,
title=yt_video.title),
)
if count == len(self.queue):
raise ValueError
self._update_play_queue_order()
except ValueError:
raise ValueError(str("Channel not found : %s" % arg))
def enqueue_audio_channel_playlist(self, channel_name, playlist_name):
"""Search a playlist within a channel and if found, adds all the audio streams
to the playback queue.
:param arg: a YouTube playlist id
"""
logging.info("args : %s - %s", channel_name, playlist_name)
try:
print_msg(
"[YouTube] [Audio channel playlist] : '{0} - {1}'. ".format(
channel_name, playlist_name))
count = len(self.queue)
yt_ch_search = MEMORY.cache(run_youtube_channel_search)
channel = yt_ch_search(channel_name)
if channel:
pl_dict = dict()
pl_titles = list()
pl_name = ""
playlist = None
for pl in channel.playlists:
print_nfo("[YouTube] [Playlist] '{0}'.".format(
to_ascii(pl.title)))
if fuzz.partial_ratio(playlist_name, pl.title) > 50:
pl_dict[pl.title] = pl
pl_titles.append(pl.title)
if len(pl_titles) > 1:
pl_name = process.extractOne(playlist_name, pl_titles)[0]
playlist = pl_dict[pl_name]
elif len(pl_titles) == 1:
pl_name = pl_titles[0]
playlist = pl_dict[pl_name]
if pl_name:
if pl_name.lower() != playlist_name.lower():
print_adv("[YouTube] Playlist '{0}' not found. "
"Playing '{1}' instead.".format(
to_ascii(playlist_name),
to_ascii(pl_name)))
for yt_video in playlist:
self._add_to_playback_queue(
video=yt_video,
info=VideoInfo(ytid=yt_video.videoid,
title=yt_video.title),
)
if count == len(self.queue):
raise ValueError
self._update_play_queue_order()
except ValueError:
raise ValueError(str("Channel not found : %s" % channel_name))
def current_audio_stream_title(self):
""" Retrieve the current stream's title.
"""
stream = self.now_playing_stream
title = ""
if stream:
title = to_ascii(stream["a"].title)
return title
def current_audio_stream_author(self):
""" Retrieve the current stream's author.
"""
stream = self.now_playing_stream
author = ""
if stream:
author = to_ascii(stream["v"].author)
return author
def current_audio_stream_file_size(self):
""" Retrieve the current stream's file size.
"""
stream = self.now_playing_stream
size = 0
if stream:
size = stream["a"].get_filesize()
return size
def current_audio_stream_duration(self):
""" Retrieve the current stream's duration.
"""
stream = self.now_playing_stream
duration = ""
if stream:
duration = to_ascii(stream["v"].duration)
return duration
def current_audio_stream_bitrate(self):
""" Retrieve the current stream's bitrate.
"""
stream = self.now_playing_stream
bitrate = ""
if stream:
bitrate = stream["a"].bitrate
return bitrate
def current_audio_stream_view_count(self):
""" Retrieve the current stream's view count.
"""
stream = self.now_playing_stream
viewcount = 0
if stream:
viewcount = stream["v"].viewcount
return viewcount
def current_audio_stream_description(self):
""" Retrieve the current stream's description.
"""
stream = self.now_playing_stream
description = ""
if stream:
description = to_ascii(stream["v"].description)
return description
def current_audio_stream_file_extension(self):
""" Retrieve the current stream's file extension.
"""
stream = self.now_playing_stream
file_extension = ""
if stream:
file_extension = to_ascii(stream["a"].extension)
return file_extension
def current_audio_stream_video_id(self):
""" Retrieve the current stream's video id.
"""
stream = self.now_playing_stream
video_id = ""
if stream:
video_id = to_ascii(stream["i"].ytid)
return video_id
def current_audio_stream_published(self):
""" Retrieve the current stream's upload date and time.
"""
stream = self.now_playing_stream
if stream:
published = to_ascii(stream["v"].published)
return published
def current_audio_stream_queue_index_and_queue_length(self):
""" Retrieve index in the queue (starting from 1) of the current stream and the
length of the playback queue.
"""
return self.play_queue_order[self.queue_index] + 1, len(self.queue)
def clear_queue(self):
""" Clears the playback queue.
"""
self.queue = list()
self.queue_index = -1
def remove_current_url(self):
"""Remove the currently active url from the playback queue.
"""
logging.info("")
if len(self.queue) and self.queue_index:
stream = self.queue[self.queue_index]
print_nfo("[YouTube] [Stream] '{0}' removed.".format(
to_ascii(stream["i"].title)))
del self.queue[self.queue_index]
self.queue_index -= 1
if self.queue_index < 0:
self.queue_index = 0
self._update_play_queue_order()
def next_url(self):
""" Retrieve the url of the next stream in the playback queue.
"""
logging.info("")
try:
if len(self.queue):
self.queue_index += 1
if (self.queue_index < len(self.queue)) and (self.queue_index
>= 0):
next_stream = self.queue[self.play_queue_order[
self.queue_index]]
return self._retrieve_stream_url(
next_stream,
self.play_queue_order[self.queue_index]).rstrip()
else:
self.queue_index = -1
return self.next_url()
else:
return ""
except (KeyError, AttributeError):
# TODO: We don't remove this for now
# del self.queue[self.queue_index]
logging.info("KeyError, or AttributeError exception")
return self.next_url()
except (IOError):
# Remove this video
del self.queue[self.queue_index]
logging.info("IOError exception")
return self.next_url()
def prev_url(self):
""" Retrieve the url of the previous stream in the playback queue.
"""
logging.info("")
try:
if len(self.queue):
self.queue_index -= 1
if (self.queue_index < len(self.queue)) and (self.queue_index
>= 0):
prev_stream = self.queue[self.play_queue_order[
self.queue_index]]
return self._retrieve_stream_url(
prev_stream,
self.play_queue_order[self.queue_index]).rstrip()
else:
self.queue_index = len(self.queue)
return self.prev_url()
else:
return ""
except (KeyError, AttributeError):
# TODO: We don't remove this for now
# del self.queue[self.queue_index]
logging.info("exception")
return self.prev_url()
except (IOError):
# Remove this video
del self.queue[self.queue_index]
logging.info("IOError exception")
return self.next_url()
def _update_play_queue_order(self):
""" Update the queue playback order.
A sequential order is applied if the current play mode is "NORMAL" or a
random order if current play mode is "SHUFFLE"
"""
total_streams = len(self.queue)
if total_streams:
if not len(self.play_queue_order):
# Create a sequential play order, if empty
self.play_queue_order = list(range(total_streams))
if self.current_play_mode == self.play_modes.SHUFFLE:
random.shuffle(self.play_queue_order)
print_nfo(
"[YouTube] [Streams in queue] '{0}'.".format(total_streams))
def _retrieve_stream_url(self, stream, queue_index):
""" Retrieve a stream url
"""
try:
if not len(self.workers):
for _ in range(WORKER_PROCESSES):
proc = Process(target=obtain_stream,
args=(self.task_queue,
self.done_queue)).start()
self.workers.append(proc)
while not self.done_queue.empty():
stream = self.done_queue.get()
self.queue[stream["q"]] = stream
stream = self.queue[queue_index]
if not stream.get("v") or not stream.get("a"):
logging.info("ytid : %s", stream["i"].ytid)
video = stream.get("v")
if not video:
yt_search = MEMORY.cache(run_youtube_search)
video = yt_search(stream["i"].ytid)
audio = video.getbestaudio(preftype="webm")
if not audio:
logging.info("no suitable audio found")
raise AttributeError()
stream.update({"a": audio, "v": video})
# streams = stream.get('v').audiostreams[::-1]
# pprint.pprint(streams)
# dump_stream_info(streams)
self.now_playing_stream = stream
return stream["a"].url
except AttributeError:
logging.info("Could not retrieve the stream url!")
raise
def _add_to_playback_queue(self, audio=None, video=None, info=None):
""" Add to the playback queue. """
if audio:
print_nfo("[YouTube] [Stream] '{0}' [{1}].".format(
to_ascii(audio.title), to_ascii(audio.extension)))
if info:
print_nfo("[YouTube] [Stream] '{0}'.".format(to_ascii(info.title)))
queue_index = len(self.queue)
self.task_queue.put(dict(a=audio, v=video, i=info, q=queue_index))
self.queue.append(dict(a=audio, v=video, i=info, q=queue_index))
def pipe_dream(layer, logger, args, backward_event, targets_queue, e,
data_size, trainloader):
criterion = nn.CrossEntropyLoss()
optimizer = optim.SGD(layer.parameters(),
lr=0.01,
momentum=0.9,
weight_decay=5e-4)
layer.train()
if dist.get_rank() == 0:
criterion.cuda(0)
output_queue = ThreadQueue(2)
data_iter = iter(trainloader)
batch_idx = 0
while True:
try:
if output_queue.qsize() == 2:
backward_event.wait()
optimizer.zero_grad()
grad = torch.zeros([args.batch_size, 128, 16, 16])
dist.recv(tensor=grad, src=1)
outputs = output_queue.get()
outputs.backward(grad.cuda(0))
optimizer.step()
backward_event.clear()
continue
else:
inputs, targets = next(data_iter)
inputs = inputs.cuda(0)
targets_queue.put(targets.numpy(), block=False)
outputs = layer(inputs)
send_opt = dist.isend(tensor=outputs.cpu(), dst=1)
send_opt.wait()
output_queue.put(outputs)
batch_idx += 1
except StopIteration as stop_e:
send_opt = dist.isend(tensor=torch.zeros(0), dst=1)
send_opt.wait()
while output_queue.qsize() > 0:
#backward_event.wait()
optimizer.zero_grad()
grad = torch.zeros([args.batch_size, 128, 16, 16])
dist.recv(tensor=grad, src=1)
outputs = output_queue.get()
outputs.backward(grad.cuda(0))
optimizer.step()
#backward_event.clear()
break
elif dist.get_rank() == 1:
batch_idx = 0
train_loss = 0
correct = 0
total = 0
criterion.cuda(1)
while True:
print("while........................")
try:
rec_val = torch.zeros([args.batch_size, 128, 16, 16])
dist.recv(tensor=rec_val, src=0)
print("recv.......")
except RuntimeError as error:
print("runtime........................")
#e.wait()
break
rec_val = rec_val.cuda(1)
rec_val.requires_grad_()
optimizer.zero_grad()
outputs = layer(rec_val)
targets = targets_queue.get(block=True, timeout=2)
targets = torch.from_numpy(targets).cuda(1)
loss = criterion(outputs, targets)
loss.backward()
optimizer.step()
train_loss += loss.item()
_, predicted = outputs.max(1)
total += targets.size(0)
correct += predicted.eq(targets).sum().item()
progress_bar(
batch_idx, data_size, 'Loss: %.3f | Acc: %.3f%% (%d/%d)' %
(train_loss /
(batch_idx + 1), 100. * correct / total, correct, total))
if not backward_event.is_set():
print("set.....")
backward_event.set()
send_opt = dist.isend(tensor=rec_val.grad.cpu(), dst=0)
print("send.....")
if batch_idx % 10 == 0:
logger.error("train:" + str(train_loss / (batch_idx + 1)))
batch_idx += 1
class Handler(Process):
"""Multiprocessing adapted handler. """
def __init__(self, handler, args=(), **kwargs):
assert issubclass(handler, logging.Handler)
self._handler = handler(*args, **kwargs)
# Patch for use non default encoding.
if issubclass(handler, logging.StreamHandler):
def _format(record):
def _encode(i):
if isinstance(i, text_type):
try:
return i.encode(sys.stdout.encoding, 'replace')
except: # pylint: disable=bare-except
pass
return i
def _decode(i):
if isinstance(i, binary_type):
try:
return u(i)
except: # pylint: disable=bare-except
pass
return i
record.msg = _decode(record.msg)
record.args = tuple(_decode(i) for i in record.args)
ret = handler.format(self._handler, record)
return _encode(ret)
self._handler.format = _format
self.queue = Queue(-1)
super(Handler, self).__init__(name=str(self._handler))
self.daemon = True
self.start()
def __getattr__(self, name):
return getattr(self._handler, name)
def run(self):
while True:
try:
record = self.queue.get()
self._handler.emit(record)
except (KeyboardInterrupt, SystemExit):
raise
except EOFError:
break
except: # pylint:disable=bare-except
traceback.print_exc(file=sys.stderr)
def _format_record(self, record):
# ensure that exc_info and args
# have been stringified. Removes any chance of
# unpickleable things inside and possibly reduces
# message size sent over the pipe
if record.args:
record.msg = record.msg % record.args
record.args = None
if record.exc_info:
self.format(record)
record.exc_info = None
return record
def emit(self, record):
"""(override)logging.handler.emit """
try:
msg = self._format_record(record)
self.queue.put_nowait(msg)
except (KeyboardInterrupt, SystemExit):
raise
except: # pylint:disable=bare-except
self.handleError(record)
def close(self):
"""(override)logging.handler.close """
self._handler.close()
