用于处理彩票的大数据算法
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

138 lines
5.1 KiB

from db import *
from util import *
# 获取全部数据
def get_all_data():
return db_get_all_data()
# 获取所有篮球的统计数据
def get_all_blue(_all_data):
all_blue = dict()
if _all_data is not False:
for data in _all_data:
if data['blue'] in all_blue.keys():
all_blue[data['blue']] += 1
else:
all_blue[data['blue']] = 1
return dict_sort(all_blue)
else:
return False
# 获取蓝球统计次数与平均值的额偏差
def get_delta_blue(_blue_data):
return dict_sort(get_analysis_data(_blue_data), 'val', True)
# 获取全部向前追溯的蓝球数据列表
def get_all_prv_blue_list(_all_data, _last_blue_list, _num):
all_prv_blue_list = []
min_id, max_id = 1, _all_data[-1]['id'] # 获取最大和最小id
for n in range(len(_last_blue_list)):
# print(_last_blue_list[n])
for _data in _all_data:
if _data['blue'] == _last_blue_list[n]:
_id = _data['id'] + _num - n
if min_id <= _id <= max_id:
# print(f"{_data['id']} - {_id}")
all_prv_blue_list.append(_id)
return sorted(set(all_prv_blue_list))
# 获取全部向后预测的历史id列表
def get_all_post_id_list(_all_data, _blue_id_list, _num):
post_id_list = []
min_id, max_id = 1, _all_data[-1]['id'] # 获取最大和最小id
for _id in _blue_id_list:
for n in range(1, _num + 1):
_post_id = _id + n
if min_id <= _post_id <= max_id:
post_id_list.append(_post_id)
return post_id_list
# 获取蓝球统计数据
def get_blue_statistics(_prv_blue_id_list):
_blue_data = db_get_data_in_ids(_prv_blue_id_list)
_blue_dict = dict()
for _data in _blue_data:
if _data['blue'] in _blue_dict.keys():
_blue_dict[_data['blue']] += 1
else:
_blue_dict[_data['blue']] = 1
return _blue_dict
# 获取分析后的差异数据
def get_analysis_data(_raw_blue_data):
total = 0
for _num in _raw_blue_data.values():
total += _num
average = total / len(_raw_blue_data.keys())
_delta_blue_dict = dict()
for _key, _val in _raw_blue_data.items():
_delta_blue_dict[_key] = round(_raw_blue_data[_key] - average, 2)
return _delta_blue_dict
# 根据蓝球获取全部历史中出现过此蓝球的id列表
def get_blue_id_list_by_blue(_all_data, _last_blue):
blue_id_list = []
for _data in _all_data:
if _data['blue'] == _last_blue:
blue_id_list.append(_data['id'])
return blue_id_list
# 获取预测的数据
def get_blue_forecast(_all_data, _method, _num):
# last_blue = db_get_last_data()['blue']
# if last_blue is False:
# return False
# blue_id_list = []
# for _data in _all_data:
# if _data['blue'] == last_blue:
# blue_id_list.append(_data['id'])
# # print(blue_id_list)
# min_id, max_id = 1, _all_data[-1]['id'] # 获取最大和最小id
"""
当_num=3
先取当前一期的蓝球号码然后计算历史上相同蓝球出现的后一期蓝球
然后往前推一期取这一期蓝球号码统计这个号码后两期的蓝球不包含后一期
然后再往前推一期取这一期蓝球号码统计这个号码后三期的蓝球不包含后一期和后两期
"""
if _method == 'prv': # 当前一期往前推num期,包含此期数据
recent_blue_id_list = [d['blue'] for d in _all_data[-_num:]]
prv_blue_id_list = get_all_prv_blue_list(_all_data, recent_blue_id_list, _num) # 根据蓝球出现的历史记录查询出向前推算的蓝球对应数据id号
_raw_blue_data = get_blue_statistics(prv_blue_id_list) # 根据上步查出的id列表,统计出蓝球出现的次数
_result = get_analysis_data(_raw_blue_data) # 根据上步蓝球的出现次数统计,计算偏差值
return dict_sort(_result, 'val', True)
elif _method == 'post': # 当前一期往后num期,不包含此期的数据
last_blue = _all_data[-1]['blue'] # 获取最后一期蓝球号码
blue_id_list = get_blue_id_list_by_blue(_all_data, last_blue) # 获取所以蓝球等于最后一期蓝球的期数id
post_blue_id_list = get_all_post_id_list(_all_data, blue_id_list, _num)
_raw_blue_data = get_blue_statistics(post_blue_id_list) # 根据上步查出的id列表,统计出蓝球出现的次数
_result = get_analysis_data(_raw_blue_data) # 根据上步蓝球的出现次数统计,计算偏差值
return dict_sort(_result, 'val', True)
else:
return False
if __name__ == "__main__":
all_data = get_all_data()
history_total = get_delta_blue(get_all_blue(all_data))
prv_five = get_blue_forecast(get_all_data(), 'prv', 5)
prv_three = get_blue_forecast(get_all_data(), 'prv', 3)
post_five = get_blue_forecast(get_all_data(), 'post', 5)
post_three = get_blue_forecast(get_all_data(), 'post', 3)
print(history_total)
print(prv_five)
print(prv_three)
print(post_five)
print(post_three)