def set_graph(self, image_obj, graph_file=NA):
self.__cur_image_obj = image_obj
digest = None
if NA == graph_file:
graph_file, digest = GraphDirHandler(image_obj.location).get_graph() if image_obj.location else \
GraphFetcher(size=image_obj.size, option=image_obj.option).fetch(image_obj.pattern)
if NA == graph_file:
return False
show(graph_file)
with open(graph_file, 'rb') as f:
try:
image = GraphViewer.get_image(f)
except IOError as e:
f.close() # close f here for we are going to delete the file below
# some image cannot be opened (maybe it's not image format?), err msg is 'cannot identify image file'
info(get_msg(Msg.fail_to_open_image), str(e))
GraphFetcher().handle_image(graph_file, DELETE)
return False
# we met "Decompressed Data Too Large" for ~/Inside Out/Image_124.jpg...
except ValueError as e:
info(get_msg(Msg.fail_to_open_image), str(e))
return False
self.__cur_graph_file = graph_file
self.__graph_history.append([self.__cur_image_obj, self.__cur_graph_file])
if digest:
digest_str = digest + "\n"
else:
digest_str = "%s：%s\n" % (get_msg(Msg.path), graph_file)
self.__cur_digest = digest_str + "%s：%sx%s" % (get_msg(Msg.size), image.size[0], image.size[1])
self.select_phrase(image_obj.pattern)
return self.set_graph_content(graph_file, image)
def handle_image(location, graph_file, action):
assert action in [INC_RANK, DEC_RANK]
handler = GraphDirHandler(location)
assert handler.__valid
base_file = graph_file.replace(location + get_delim(), "")
for image in handler.__status_cache:
if image == base_file:
has_change = True
status = handler.__status_cache[image]
if INC_RANK == action:
status.rank += 1
msg = get_msg(Msg.change_rank_to) + str(status.rank)
else:
if 1 == status.rank:
msg = get_msg(Msg.cannot_lower_down_rank_as_it_is_already_the_lowest)
has_change = False
else:
status.rank -= 1
msg = get_msg(Msg.change_rank_to) + str(status.rank)
if has_change:
handler.__status_cache[image] = status
cache_file = location + get_delim() + GraphDirHandler.CACHE_FILE
# TODO: it shows that the timestamp not change...
timestamp = time.ctime(os.path.getmtime(location))
save(cache_file, [timestamp, handler.__status_cache])
return msg
assert False
def handle_image(graph_file, action):
assert action in [DELETE, DISCARD, INC_RANK, DEC_RANK]
key_str = get_delim() + "image_"
end_pos = graph_file.find(key_str)
assert -1 != end_pos
begin_pos = graph_file[:end_pos].rfind(get_delim())
assert -1 != begin_pos
pattern = graph_file[begin_pos + 1:end_pos]
has_cache, cached_objs = load(GraphFetcher.get_cache_file(pattern))
assert has_cache
file_encoding = graph_file[graph_file.find(key_str) + len(key_str):graph_file.find(".jpg")]
for url in cached_objs:
image_slot = cached_objs[url]
if image_slot.encoding == file_encoding:
new_encoding = NA if DELETE == action else \
None if DISCARD == action else \
image_slot.encoding # no change
new_rank = image_slot.rank + 1 if INC_RANK == action else \
image_slot.rank - 1 if DEC_RANK == action and image_slot.rank is not 1 else \
image_slot.rank # no change
updated_slot = ImageSlot(timestamp=image_slot.timestamp,
encoding=new_encoding,
rank=new_rank)
cached_objs[url] = updated_slot
save(GraphFetcher.get_cache_file(pattern), cached_objs)
if action in [DELETE, DISCARD]:
os.remove(graph_file)
msg = "" if action in [DELETE, DISCARD] else \
get_msg(Msg.change_rank_to) + str(new_rank) + "！" if new_rank is not image_slot.rank else \
get_msg(Msg.cannot_lower_down_rank_as_it_is_already_the_lowest) if image_slot.rank is 1 else \
None
assert msg is not None
return msg
assert False
def get_graph_digest(self, graph_file):
if NA is graph_file:
return "NA"
full_graph_file = self.__location + get_delim() + graph_file
timestamp = time.ctime(os.path.getmtime(full_graph_file))
return "%s：%s\n%s：%s\n%s：%s" % (
get_msg(Msg.location), full_graph_file,
get_msg(Msg.timestamp), timestamp,
get_msg(Msg.rank), self.__status_cache[graph_file].rank)
def get_graph_digest(graph_file, url_obj):
if NA is graph_file:
return "NA"
relative_pos = graph_file.find(pic_home())
assert 0 == relative_pos
relative_graph_file = graph_file[len(pic_home()):]
return "%s：%s\n%s：%s\n%s：%s" % (
get_msg(Msg.location), relative_graph_file,
get_msg(Msg.timestamp), url_obj.timestamp.strftime("%B %d, %Y"),
get_msg(Msg.rank), url_obj.rank)
def check_access_status():
print(get_msg(Msg.check_network_connection))
try:
urllib2.urlopen('http://google.com', timeout=3)
print(get_msg(Msg.network_status_succeed))
return True
except urllib2.URLError:
pass
print(get_msg(Msg.network_status_fail))
return False
def load(pickle_file):
"""output: is_exist, value"""
try:
pickle_fd = open(pickle_file, "r")
except IOError as err:
if errno.ENOENT == err.errno:
show(get_msg(Msg.cache_file_does_not_exist), pickle_file)
return False, None
assert False
try:
value = cPickle.load(pickle_fd)
return True, value
except (ValueError, UnpicklingError, EOFError):
error(get_msg(Msg.cannot_read_pickle_file), pickle_file, get_msg(Msg.suggest_re_fetch_pickle_file))
assert False
def set_graph_content(self, graph_file, image=None):
if image is None:
try:
image = GraphViewer.get_image(graph_file)
except IOError as e:
error(str(e))
assert False
self.__root.geometry(self.__full_geom if self.__fullscreen_mode else
'%dx%d+0+0' % (image.size[0], image.size[1]))
if self.__fullscreen_mode:
resize_width, resize_height, x_pos, y_pos = self.get_adjusted_geom(image.size[0], image.size[1])
try:
resized = image.resize((resize_width, resize_height), Image.ANTIALIAS)
except IOError as e:
# 'incomplete downloaded image' may go here
info(get_msg(Msg.fail_to_convert_image_to_fullscreen), str(e))
GraphFetcher().handle_image(graph_file, DISCARD)
return False
image = resized
self.__root.title(self.__cur_image_obj.group_name)
tk_image_obj = ImageTk.PhotoImage(image)
self.__tk_obj_ref = tk_image_obj
self.__canvas.delete('all')
self.__canvas.create_image(x_pos if self.__fullscreen_mode else 0, y_pos if self.__fullscreen_mode else 0,
image=tk_image_obj, anchor=Tkinter.NW)
self.show_onscreen_help()
self.show_onscreen_info()
self.show_onscreen_phrase()
return True
def save(pickle_file, value):
pickle_fd = open(pickle_file, "w")
try:
cPickle.dump(value, pickle_fd)
except AttributeError as msg:
error(get_msg(Msg.fail_to_write_cache), str(msg))
pickle_fd.close()
def __init__(self, in_file):
self.__fd = open(in_file)
try:
self.__json_data = json.load(self.__fd)
except Exception, e:
error(get_msg(Msg.fail_read_file) + "\"", in_file, "\"")
error(str(e))
assert False
def delete_image(self, *unused):
if self.__cur_image_obj.location:
return # spec.: not support remove image that user 'specified'
info(get_msg(Msg.remove_image), self.__cur_graph_file)
self.__graph_history.remove([self.__cur_image_obj, self.__cur_graph_file])
GraphFetcher.handle_image(self.__cur_graph_file, DELETE)
self.cancel_pending_jobs()
self.timer_action(True)
def decrement_rank(self, *unused):
info(get_msg(Msg.decrease_rank), self.__cur_graph_file)
if self.__cur_image_obj.location:
msg = GraphDirHandler.handle_image(self.__cur_image_obj.location, self.__cur_graph_file, DEC_RANK)
else:
msg = GraphFetcher.handle_image(self.__cur_graph_file, DEC_RANK)
self.__cur_digest += "\n%s" % msg
self.show_onscreen_info()
def get_recent_result(self, key):
"""output: urls, size_ratio"""
if key not in self.__url_map:
return None, None
[retrieved_date, new_result, urls, size_ratio] = self.__url_map[key]
if not self.__network_reachable or Crawler.__STOP_SEARCH:
show(get_msg(Msg.use_previous_search_result))
# though size_ratio can be valid, we do not return it for caller usage is not expected
return urls, None
# spec.: we will execute a new search when there is enough new result on previous search
# => if previous new result is n, all result is m, we will have a new search after m/n days
# => if all previous result is new, then after 1 day we will have a search
# => if no previous result is new, then we will have a search after 'TARGET_SEARCH_RESULT_SIZE' days
valid_day_size = len(urls) / new_result if new_result > 0 else \
1 if NA is new_result else \
TARGET_SEARCH_RESULT_SIZE # new_result = 0 => no new result before
from util.global_def import get_search_latency
valid_day_size *= get_search_latency()
current_date = datetime.today()
date_diff = current_date - retrieved_date
if date_diff > timedelta(days=valid_day_size): # 'valid_day_size' is the valid duration of search result
return None, size_ratio
to_next_query = timedelta(days=valid_day_size) - date_diff
hours, remainder = divmod(to_next_query.seconds, 3600)
minutes, seconds = divmod(remainder, 60)
show(get_msg(Msg.to_next_search),
to_next_query.days, get_msg(Msg.day),
hours, get_msg(Msg.hour),
minutes, get_msg(Msg.minute),
seconds, (get_msg(Msg.second) + ","), get_msg(Msg.current_url_count), len(urls))
# though size_ratio can be valid, we do not return it for caller usage is not expected
return urls, None
def timer_action(self, user_next_image=False):
if not user_next_image and self.__pause_slideshow:
self.prepare_for_next_view(get_slideshow_frequency() * 1000)
return
success = self.set_graph(self.select_pattern())
if not success:
self.prepare_for_next_view(1, get_msg(Msg.try_fetch_image_again))
return
self.prepare_for_next_view(get_slideshow_frequency() * 1000)
def __load_or_create_status(self):
status_cache = {} # key: image_file, value: status
cache_file = self.__location + get_delim() + GraphDirHandler.CACHE_FILE
cache_existed = os.path.exists(cache_file)
if cache_existed:
success, cache_data = load(cache_file)
assert success
[timestamp, status_cache] = cache_data
if not self.dir_changed(timestamp):
return status_cache
else:
info(get_msg(Msg.directory), self.__location, get_msg(Msg.has_changed_update_cache_file))
else:
info("%s%s" % (get_msg(Msg.create_new_cache_file_for_directory), self.__location))
image_files = []
for root, _, files in os.walk(self.__location):
assert len(root) >= 1
if root[-1] != get_delim():
root += get_delim()
for base_file in files:
basename, ext = os.path.splitext(base_file)
if ext.replace(".", "") in GraphDirHandler.RECOGNIZED_IMAGE_EXT:
image_files.append((root + base_file).replace(self.__location, ""))
if not image_files:
if cache_existed:
os.remove(cache_file)
self.__valid = False
return None
existed_image = {}
for image in image_files:
existed_image[image] = 1 # 1 is just a dummy value
if image not in status_cache:
status_cache[image] = Status()
to_be_deleted = []
for image in status_cache: # this check works when some image is deleted
if image not in existed_image:
to_be_deleted.append(image)
for image in to_be_deleted:
status_cache.pop(image)
# TODO: this makes an 'always' has-changed 2nd time image
timestamp = time.ctime(os.path.getmtime(self.__location))
save(cache_file, [timestamp, status_cache])
return status_cache
def get_graph_file(self, pattern, url, cached_encoding):
"""output: graph_file, encoding"""
if NA == cached_encoding: # mean this url is not retrievable
return NA, NA
file_encoding = cached_encoding
if not file_encoding:
file_encoding = GraphFetcher.get_file_encoding(pattern)
graph_dir = GraphFetcher.get_graph_dir(pattern)
if not os.path.exists(graph_dir):
try:
os.makedirs(graph_dir)
except OSError as e:
error(get_msg(Msg.cannot_create_directory), str(e))
import sys
sys.exit()
abs_graph_file = graph_dir + "image_" + file_encoding + ".jpg"
if os.path.exists(abs_graph_file):
return abs_graph_file, file_encoding
if not self.__network_reachable:
info(get_msg(Msg.give_up_fetch_image))
return NA, None
self.__has_write = True
try:
info(get_msg(Msg.fetch_image), url)
try:
web_content = urllib2.urlopen(url, timeout=10)
except httplib.BadStatusLine:
info(get_msg(Msg.obtain_unrecognized_status_code), url)
return NA, NA
fd = open(abs_graph_file, 'wb')
fd.write(web_content.read())
fd.close()
assert os.path.exists(abs_graph_file)
if os.stat(abs_graph_file).st_size <= 10240:
info(get_msg(Msg.give_up_acquired_image_with_size), os.stat(abs_graph_file).st_size, "Bytes")
info(get_msg(Msg.remove_image), abs_graph_file)
os.remove(abs_graph_file)
return NA, NA
info(get_msg(Msg.fetch_succeed))
return abs_graph_file, file_encoding
except (IOError, httplib.IncompleteRead, ssl.CertificateError) as e:
info(get_msg(Msg.failed_url), url)
info(get_msg(Msg.error_message), str(e))
if os.path.exists(abs_graph_file):
fd.close()
os.remove(abs_graph_file)
return NA, NA
def crawl_by_asking_google_search(pattern, start, size, option=""):
assert type(pattern) in [str, unicode]
from util.global_def import get_api_key, get_cx
api_key = get_api_key()
cx = get_cx()
if not api_key or not cx:
if not Crawler._HAS_SHOW_NO_SEARCH_MSG:
Crawler._HAS_SHOW_NO_SEARCH_MSG = True
info(get_msg(Msg.no_search_due_to_no_api_key_and_cx))
return [], False
size_option = "&imgSize=" + size if size else ""
full_option = size_option + (option if option else "")
base_url = 'https://www.googleapis.com/customsearch/v1?key=%s&cx=%s&searchType=image&num=%d' \
'&q=' + pattern + '&start=%d' + full_option
request_str = base_url % (api_key, cx, G_SEARCH_PER_REQ_SIZE, start)
urls = []
success = True
try:
r = requests.get(request_str)
res = json.loads(r.text)
if "error" in res:
Crawler.print_error(res["error"])
if "This API requires billing to be enabled on the project" in res["error"]["message"]:
# this is the 'out of quota' message
Crawler.__STOP_SEARCH = True
return urls, False
if 'items' not in res:
info(get_msg(Msg.cannot_fetch_image_url), "empty query")
return urls, True # return 'True' is okay?
for image_info in res['items']:
assert 'link' in image_info
url = image_info['link']
urls.append(url)
except TypeError as e: # for unhandled error...
info(get_msg(Msg.cannot_fetch_image_url), str(e))
success = False
except requests.ConnectionError as e:
info(get_msg(Msg.cannot_fetch_image_url), str(e))
success = False
return urls, success
def view(self, image_obj_list, phrase_obj_list):
if not phrase_obj_list:
# the WTF 'mutable default argument' property makes us not have [] firstly
phrase_obj_list = []
if not image_obj_list:
info(get_msg(Msg.not_any_image_specified_program_exit))
sys.exit()
self.setup_image_stuff(image_obj_list)
self.setup_phrase_stuff(image_obj_list, phrase_obj_list)
while True:
self.timer_action(True)
self.__root.mainloop()
self.cancel_pending_jobs()
def crawl(self, pattern, size_list, option="", print_url=False):
"""output: urls, is_new_result"""
show(get_msg(Msg.search_target), "\"" + pattern + "\"")
key = Crawler.get_search_key(pattern, option)
urls, size_ratio = self.get_recent_result(key)
if urls:
return urls, False
if not self.__network_reachable or Crawler.__STOP_SEARCH:
return None, False
assert size_list and (not size_ratio or isinstance(size_ratio, dict))
dice = Crawler.get_dice(size_list, size_ratio)
urls = []
next_size_ratio = {size: 0 for size in size_list} # key: size, value: number of new result (initial with 0)
start = {size: 1 for size in size_list} # key: size, value: next search start offset (start from 1 by google)
tried_size = 0
while tried_size < TARGET_SEARCH_RESULT_SIZE:
chosen_size = get_weighted_random_dict_key(dice)
this_urls, success = Crawler.crawl_by_asking_google_search(pattern, start[chosen_size], chosen_size, option)
if not success:
break
urls += this_urls
new_result = self.get_this_time_new_result_num(key, this_urls)
next_size_ratio[chosen_size] += (new_result if NA != new_result else len(this_urls))
start[chosen_size] += G_SEARCH_PER_REQ_SIZE
tried_size += G_SEARCH_PER_REQ_SIZE
# 'set' to filter out duplicated item (though not expected, but we found g-search may give duplicated result)
urls = list(set(urls))
if not Crawler._HAS_SHOW_NO_SEARCH_MSG:
info("%s：%s, %s：%i" % (
get_msg(Msg.target), pattern,
get_msg(Msg.acquired_url_count), len(urls)))
if print_url:
for url in urls:
show(url)
if success:
next_size_ratio = {size: 1 if 0 == next_size_ratio[size] else next_size_ratio[size]
for size in next_size_ratio}
self.cache_url(key, urls, next_size_ratio)
return urls, success
def select_pattern(self):
if self.__arbitrator.is_active():
choice_pattern = None
while not choice_pattern:
choice_pattern = self.__arbitrator.arbitrate()
if not choice_pattern:
show(get_msg(Msg.no_available_image_wait_10_minutes))
self.__root.withdraw()
import time
time.sleep(600)
self.__root.deiconify()
return self.__cur_image_obj_dict[choice_pattern]
image_obj_size = len(self.__cur_image_obj_list)
return self.__cur_image_obj_list[random.randrange(0, image_obj_size)]
def __parse_config(self, config_file):
from util.config import Config
config = Config(config_file)
config.set_general_setting()
image_target = config.get_setting("image", "target")
if not image_target:
print(get_msg(Msg.not_any_image_specified_program_exit))
sys.exit()
phrase_target = config.get_setting("phrase", "target")
import glob
self.__image_setting += glob.glob(image_target)
self.__phrase_setting += glob.glob(phrase_target) if phrase_target else []
def fetch(self, pattern):
self.__has_write = False
new_objs, old_objs = self.get_updated_url(pattern)
show(get_msg(Msg.total_data_count), len(new_objs) + len(old_objs))
url = self.choose_url(new_objs, old_objs)
if NA == url:
return NA, NA
image_objs = old_objs
image_objs.update(new_objs)
image_slot = image_objs[url]
graph_file, new_encoding = self.get_graph_file(pattern, url, image_slot.encoding)
new_slot = ImageSlot(image_slot.timestamp, new_encoding, image_slot.rank)
image_objs[url] = new_slot
if self.__has_write:
save(GraphFetcher.get_cache_file(pattern), image_objs)
return graph_file, GraphFetcher.get_graph_digest(graph_file, image_objs[url])
@staticmethod
def get_file_encoding(pattern):
# TODO: add a file to keep last largest number to avoid possible long glob time...
file_list = glob.glob(GraphFetcher.get_graph_dir(pattern) + "image_*.jpg")
largest_idx = 0
# noinspection PyShadowingNames
for graph_file in file_list:
begin_pos = graph_file.find("image_")
end_pos = graph_file.find(".jpg")
assert -1 != begin_pos and -1 != end_pos
begin_pos += len("image_")
iter_idx = int(graph_file[begin_pos:end_pos])
assert iter_idx > 0
largest_idx = iter_idx if iter_idx > largest_idx else largest_idx
largest_idx += 1
return str(largest_idx)
if __name__ == '__main__':
from util.global_def import config_action
config_action()
# name '_' before the 'obj' to let python not free imported module before __del__ is called
# (or we will have 'NoneType' object has no attribute 'dump' for cPickle.dump)
_obj = GraphFetcher()
graph_file, digest = _obj.fetch("Inside Out")
if NA == graph_file:
print(get_msg(Msg.fetch_image_fail))
else:
print(graph_file)
def help_str():
return get_msg(Msg.help_message)
def print_error(data):
assert isinstance(data, dict) and "message" in data
error(get_msg(Msg.search_engine_err_msg), data["message"])
