def primary_hardware_type_to_precise(path,
path_precise,
path_save,
brand="Vivo"):
"""
提取primaryHardwareType字符串作为model
品牌：huawei
输入：Vivo^V5 (1601)^V5^Android^Mobile Phone^2016^0^vivo vivo x20plus a
输出：koobee^koobee A3^2^koobee A3
Tablet Mobile Phone
:return:
"""
utils.delete_existed_file(path_save)
data_set = []
for line in utils.read_txt(path):
lines = line.split("^")
model = lines[2].title().strip()
if lines[4] == "Mobile Phone":
type_str = "2"
else:
type_str = "3"
model = clean_space(model.replace("-", " ").replace("-", " "))
keywords = lines[7].upper().strip()
data_set.append(brand + "^" + model + "^" + type_str + "^" + keywords)
for data in utils.read_txt(path_precise):
datas = data.split("^")
datas_keywords = datas[3].split("|")
for k in datas_keywords:
data_set.append(brand + "^" + datas[1].title() + "^" + datas[2] +
"^" + k)
utils.save_to_datas_file(data_set, path_save)
utils.deduplication_data(path_save)
def combine(self, path_save):
self.load_atlas()
utils.delete_existed_file(path_save)
for line in utils.read_txt(setting.UA2DEVICE_PRECISE_DATE):
model_dict = {}
line = utils.format_device_ua(line)
lines = line.split("^")
model_dict["brand"] = lines[0]
model_dict["model"] = lines[1]
model_dict["type"] = lines[2]
model_dict["keywords"] = lines[3]
keywords = lines[3].split("|")
for k in keywords:
if k in self.atlas:
# 判断品牌 相同品牌以atlas库为准 不同品牌以rtb库为准
if lines[0] != self.atlas[k]["brand"]:
self.atlas[k] = model_dict
else:
self.rtb_ua[k] = model_dict
dataset = []
for k, v in self.atlas.items():
dataset.append(v["brand"] + "^" + v["model"] + "^" + v["type"] +
"^" + v["keywords"])
for k, v in self.rtb_ua.items():
dataset.append(v["brand"] + "^" + v["model"] + "^" + v["type"] +
"^" + v["keywords"])
utils.save_to_datas_file(dataset, path_save)
utils.deduplication_data(path_save)
def check_ua_tv():
# /sas/tk_history/box_ua_0327.log zzy19
# path_input = "/Users/wangchun/PycharmProjects/Analysis/mobile_identify/data/tv_ua_text.txt"
path_input = "/sas/tk_history/box_ua_0327.log"
# print(path_input)
identify_ua = identify.UaIdentify()
identify_ua.load_data()
count_brand = {}
count = 0
for line in utils.read_txt(path_input):
# print(line)
count += 1
data_result = identify_ua.identify_tv(line)
if data_result is None:
continue
if data_result["brand"] in count_brand:
count_brand[data_result["brand"]] += 1
else:
count_brand[data_result["brand"]] = 1
count_brand = sorted(count_brand.items(), key=lambda e: e[1], reverse=True)
data_set = []
sum_brand = 0
for k, v in count_brand:
sum_brand += v
data_set.append(k + "\t" + str(v) + "\t" + str(v / count))
data_set.append("总数:" + str(count) + "\t" + "识别数:" + str(sum_brand) +
"\t" + "识别率:" + str(sum_brand / count))
utils.save_to_datas_file(data_set, "count_result.txt")
def dropout_repeat_keywords(path, path_save_data, path_save_repeat):
# 删除过重复值
# Zte^yuanhang 4^2^ZTE BA610C|ZTE BA610T|ZTE BLADE A610C UA2DEVICE_PRECISE_DATE setting.device_ua_output
utils.delete_existed_file(path_save_data)
utils.delete_existed_file(path_save_repeat)
keywords_repeat = utils.get_repeat_keywords(path)
data_set = []
data_repeat = []
keywords_list = []
for line in utils.read_txt(path):
lines = utils.format_device_ua(line).split("^")
keywords = lines[3].upper().split("|")
for k in keywords:
if k in keywords_repeat:
lines_fix, keywords_list = ua_fix.fix_repeat_data(lines, keywords_list)
if lines_fix:
# print(keywords_list)
data_set.append(lines[0] + "^" + lines[1] + "^" + lines[2] + "^" + k)
else:
data_repeat.append(lines[0] + "^" + lines[1] + "^" + lines[2] + "^" + k)
else:
data_set.append(lines[0] + "^" + lines[1] + "^" + lines[2] + "^" + k)
utils.save_to_datas_file(data_set, path_save_data)
utils.deduplication_data(path_save_data)
utils.save_to_datas_file(data_repeat, path_save_repeat)
utils.deduplication_data(path_save_repeat)
def get_raw_data(path_data):
"""
原始数据格式：
brand +"^"+ vendor + "^" + model + "^" +marketingName + "^" + osName + "^" + primaryHardwareType + "^" + yearReleased + "^" + isTablet + "^" + identifierUa;
保持原数据格式，提取identifierUa。
Vivo^Y11i T^Y11^Android^Mobile Phone^2013^0^vivo Y11i T Build/KTU84P
:return:
"""
data_set = []
uf = ua_fix.UaFix()
for line in utils.read_txt(path_data):
lines = line.split("^")
# 过滤异常
if uf.filter_main(lines[0], lines[8]):
continue
if "build/" in lines[8].lower():
build = lines[8].lower().split("build/")[0].strip()
if len(build) == 0:
continue
lines[8] = build
data_set.append("^".join(lines))
elif "miui/" in lines[8].lower():
build = lines[8].lower().split("miui/")[0].strip()
if len(build) == 0:
continue
lines[8] = build
data_set.append("^".join(lines))
else:
pass
return data_set
def format_ua_device_pattern(path=setting.UA2DEVICE_PATTERN, path_save=setting.UA2DEVICE_PATTERN_DATE):
"""格式化 ua2device_pattern"""
utils.delete_existed_file(path_save)
data_set = []
for line in utils.read_txt(path):
line = utils.format_device_ua(line)
data_set.append(line)
utils.save_to_datas_file(data_set, path_save)
utils.deduplication_data(path_save)
def load_old(self, path):
# Zte^yuanhang 4^2^ZTE BA610C|ZTE BA610T|ZTE BLADE A610C
for line in utils.read_txt(path):
lines = line.split("^")
brand = utils.clean_space(lines[0]).lower()
model = utils.clean_space(lines[1]).lower()
if not (brand in self.brand_old):
self.brand_old[brand] = lines[0]
if not (model in self.model_old):
self.model_old[model] = lines[1]
def get_info_form_ua(path, path_save, keyword):
# 提取原有库数据
data_set = []
utils.delete_existed_file(path_save)
for line in utils.read_txt(path):
lines = line.split("^")
if lines[0].upper() != keyword.upper():
continue
data_set.append(line)
utils.save_to_datas_file(data_set, path_save)
def load_atlas(self):
# Zte^yuanhang 4^2^ZTE BA610C|ZTE BA610T|ZTE BLADE A610C
for line in utils.read_txt(setting.device_ua_output):
model_dict = {}
lines = utils.format_device_ua(line).split("^")
model_dict["brand"] = lines[0]
model_dict["model"] = lines[1]
model_dict["type"] = lines[2]
model_dict["keywords"] = lines[3]
keywords = lines[3].split("|")
for k in keywords:
if k not in self.atlas:
self.atlas[k] = model_dict
def data_to_dict(self, data_dic, name):
# Zte^yuanhang 4^2^ZTE BA610C|ZTE BA610T|ZTE BLADE A610C
path = self.RESOURCE_PATH + name
for line in utils.read_txt(path):
model_dict = {}
lines = utils.format_device_ua(line).split("^")
model_dict["brand"] = lines[0]
model_dict["model"] = lines[1]
model_dict["type"] = lines[2]
model_dict["keywords"] = lines[3]
keywords = lines[3].split("|")
for k in keywords:
if k not in data_dic:
data_dic[k] = model_dict
def check_tong_not_match_ua_tv_():
data_path = "/home/galen/ua_keywords/"
log_name_list = [
"baidu_model_0328.log", "miaozhen_model_0328.log",
"youku_model_0328.log"
]
identify_ua = identify.UaIdentify()
identify_ua.load_data()
data_set = []
count = 0
count_type = {}
path_result = data_path + "count_result.txt"
utils.delete_existed_file(path_result)
for name in log_name_list:
path_input = data_path + name
ua_not_in_path = data_path + name.replace(".log", "_output.log")
utils.delete_existed_file(ua_not_in_path)
print(path_input)
for line in utils.read_txt(path_input):
# print(line)
count += 1
data_result = identify_ua.detect_model(clean_character(line))
if data_result is None:
data_set.append(line)
else:
if data_result["type"] in count_type:
count_type[data_result["type"]] += 1
else:
count_type[data_result["type"]] = 1
if len(data_set) >= 500:
utils.save_to_datas_file(data_set, ua_not_in_path)
data_set = []
if count % 100000 == 0:
print(count)
utils.save_to_datas_file(data_set, ua_not_in_path)
sum_type = 0
data_set_count = []
tv_unkown_sum = count
for k, v in count_type.items():
sum_type += v
if k != "4":
tv_unkown_sum = tv_unkown_sum - v
data_set_count.append(k + "\t" + str(v) + "\t" + str(v / count))
data_set_count.append("总数:" + str(tv_unkown_sum) + "\t" + "识别数:" +
str(sum_type) + "\t" + "识别率:" + str())
utils.save_to_datas_file(data_set_count, path_result)
def format_precise(path, path_save, path_brand_save):
"""
输入：koobee^koobee A3^2^koobee A3
输出：koobee^koobee A3^2^koobee A3|XXXX\
过滤后拼接
:return:
"""
utils.delete_existed_file(path_save)
data_set = []
# models_length : length brand model type keywords
models_length = {}
for line in utils.read_txt(path):
model_dict = {}
lines = line.split("^")
model = lines[1]
keyword = lines[3]
model_l = utils.delete_space(model)
# 过滤 AA B与AAB 格式数据
if model_l in models_length:
old_model_len = models_length[model_l]["length"]
# print(models_length[model_l])
if old_model_len >= len(model):
models_length[model_l]["model"] = model
if not utils.whether_repeat(models_length[model_l]["keywords"], keyword):
models_length[model_l]["keywords"] = models_length[model_l]["keywords"] + "|" + keyword
else:
models_length[model_l]["keywords"] = keyword
else:
model_dict["model"] = model
model_dict["type"] = lines[2]
model_dict["brand"] = lines[0]
model_dict["length"] = len(model_l)
model_dict["keywords"] = keyword
models_length[model_l] = model_dict
brand_set = set()
for k, v in models_length.items():
# print(v)
brand_set.add(v["brand"])
data_set.append(v["brand"] + "^" + v["model"] + "^" + v["type"] + "^" + v["keywords"])
utils.save_to_datas_file(data_set, path_save)
utils.save_to_datas_file(list(brand_set), path_brand_save)
utils.deduplication_data(path_save)
def get_not_match_ua_tv():
# /sas/tk_history/box_ua_0327.log zzy19
# path_input = "/Users/wangchun/PycharmProjects/Analysis/mobile_identify/data/tv_ua_text.txt"
path_input = "/sas/tk_history/box_ua_0327.log"
# print(path_input)
identify_ua = identify.UaIdentify()
identify_ua.load_data()
data_set = []
count = 0
for line in utils.read_txt(path_input):
# print(line)
count += 1
data_result = identify_ua.identify_tv(line)
if data_result is None:
data_set.append(line)
if len(data_set) >= 500:
print(count)
utils.save_to_datas_file(data_set, "not_in_list_ua_3.txt")
data_set = []
utils.save_to_datas_file(data_set, "not_in_list_ua_3.txt")
def combine(self, path_save, path_update):
self.load_old(setting.UA2DEVICE_PRECISE)
self.load_old(setting.UA2DEVICE_PATTERN)
utils.delete_existed_file(path_save)
utils.delete_existed_file(path_update)
dataset = []
dataset_update_list = []
# setting.UA2DEVICE_PRECISE_DATE 是device_ua_combine.txt复制过来的
for line in utils.read_txt(setting.UA2DEVICE_PRECISE_DATE):
lines = line.split("^")
brand = utils.clean_space(lines[0]).lower()
model = utils.clean_space(lines[1]).lower()
if brand in self.brand_old:
dataset_update_list.append("brand:" + lines[0] + "<=" +
self.brand_old[brand])
lines[0] = self.brand_old[brand]
if model in self.model_old:
dataset_update_list.append("model:" + lines[1] + "<=" +
self.model_old[model])
lines[1] = self.model_old[model]
else:
if model.startswith(brand):
model = model.replace(brand, "").strip()
if model in self.model_old:
print("model:" + lines[1] + "<=" +
self.model_old[model])
dataset_update_list.append("model:" + lines[1] + "<=" +
self.model_old[model])
lines[1] = self.model_old[model]
# for k, v in self.model_old.items():
# if model.endswith(k):
# dataset_update_list.append(lines[1] + "<=" + self.model_old[k])
# print(lines[1] + "<=" + self.model_old[k])
# lines[1] = self.model_old[k]
# break
dataset.append("^".join(lines))
utils.save_to_datas_file(dataset, path_save)
utils.deduplication_data(path_save)
utils.save_to_datas_file(dataset_update_list, path_update)
utils.deduplication_data(path_update)
def extract_band(brand, keyword, path_data, path_raw, path_build,
path_not_build):
"""
原始数据格式
vendor + "^" + model + "^" +marketingName + "^" + osName + "^" + primaryHardwareType + "^" + yearReleased + "^" + isTablet + "^" + identifierUa;
Vivo^Y11i T^Y11^Android^Mobile Phone^2013^0^vivo Y11i T Build/KTU84P
:return:
"""
utils.delete_existed_file(path_raw)
utils.delete_existed_file(path_build)
utils.delete_existed_file(path_not_build)
for line in utils.read_txt(path_data):
if brand.lower() == "xiaomi":
# XIAOCAISHEN XIAOMANYAO XING_FOURTEEN_V3 XINDAN XINGMI
if filter_xiaomi(line):
continue
lines = line.split("^")
if brand.lower() != lines[0].lower():
continue
if '(' in lines[7] or ')' in lines[7]:
if keyword.lower() not in lines[7].lower():
continue
utils.save_to_data_file(line + "\n", path_raw)
else:
if "build/" in lines[7].lower():
build = lines[7].lower().split("build")[0].strip()
if len(build) == 0:
continue
lines[7] = build
utils.save_to_data_file("^".join(lines) + "\n", path_build)
elif "miui/" in lines[7].lower():
build = lines[7].lower().split("miui")[0].strip()
if len(build) == 0:
continue
lines[7] = build
utils.save_to_data_file("^".join(lines) + "\n", path_build)
else:
utils.save_to_data_file(line + "\n", path_not_build)
utils.deduplication_data(path_build)
def combine_precise_format(path, path_save, brand="Vivo"):
"""
输入：Vivo^V5 (1601)^V5^Android^Mobile Phone^2016^0^vivo vivo x20plus a
输出：koobee^koobee A3^2^koobee A3
:return:
"""
utils.delete_existed_file(path_save)
data_set = []
models = {}
models_length = {}
for line in utils.read_txt(path):
lines = line.split("^")
model = lines[1].lower()
if brand not in ["huawei", "xiaomi", "samsung", "gionee"]:
if brand.lower() in model:
model = brand + " " + model.replace(brand.lower(), "").strip()
else:
model = brand.lower() + " " + model
model_l = delete_space(model.replace("-", " "))
if model_l in models_length:
old_model_len = int(models_length[model_l].split("^")[0])
old_model = models_length[model_l].split("^")[-1]
if old_model_len > len(model):
model = old_model
else:
models_length[model_l] = str(len(model)) + "^" + model
# model+type
model = clean_space(model).strip().title() + "^" + lines[2]
keywords = lines[3].upper()
if model in models:
if not whether_repeat(models[model], keywords):
models[model] = models[model] + "|" + keywords
else:
models[model] = keywords
for k, v in models.items():
data_set.append(brand.title() + "^" + k + "^" + v)
utils.save_to_datas_file(data_set, path_save)
utils.deduplication_data(path_save)
