├── MANIFEST.in
├── phone
    ├── __init__.py
    ├── phone.dat
    ├── test_phone.py
    ├── build_record.py
    └── phone.py
├── travis
    └── setup.py
├── .travis.yml
├── setup.py
├── .gitignore
└── README.md


/MANIFEST.in:
--------------------------------------------------------------------------------
1 | include phone/phone.dat
2 | 


--------------------------------------------------------------------------------
/phone/__init__.py:
--------------------------------------------------------------------------------
1 | from .phone import Phone
2 | 
3 | __all__ = ['Phone']
4 | 


--------------------------------------------------------------------------------
/phone/phone.dat:
--------------------------------------------------------------------------------
https://raw.githubusercontent.com/ls0f/phone/HEAD/phone/phone.dat


--------------------------------------------------------------------------------
/travis/setup.py:
--------------------------------------------------------------------------------
1 | from phone import Phone
2 | 
3 | 
4 | p = Phone()
5 | p.test()


--------------------------------------------------------------------------------
/.travis.yml:
--------------------------------------------------------------------------------
 1 | # https://travis-ci.org/lovedboy/phone
 2 | language: python
 3 | python:
 4 |     - 2.7
 5 |     - pypy
 6 |     - 3.3
 7 |     - 3.4
 8 |     - 3.5
 9 |     - nightly
10 |     - pypy3
11 |     
12 | branches:
13 |   only:
14 |     - master
15 |     
16 | script:
17 |     - python -m unittest discover
18 |     - python setup.py -q install
19 |     - python ./travis/setup.py
20 | 


--------------------------------------------------------------------------------
/setup.py:
--------------------------------------------------------------------------------
 1 | # coding:utf-8
 2 | from setuptools import setup, find_packages
 3 | 
 4 | PACKAGE = "phone"
 5 | NAME = "phone"
 6 | DESCRIPTION = "手机号码库"
 7 | AUTHOR = "ls0f"
 8 | AUTHOR_EMAIL = "admin@lovedboy.com"
 9 | URL = "https://github.com/ls0f/phone"
10 | VERSION = '0.4.5'
11 | 
12 | setup(
13 |     name=NAME,
14 |     version=VERSION,
15 |     description=DESCRIPTION,
16 |     author=AUTHOR,
17 |     author_email=AUTHOR_EMAIL,
18 |     license="BSD",
19 |     url=URL,
20 |     include_package_data=True,
21 |     packages=find_packages(),
22 |     classifiers=[
23 |         'License :: OSI Approved :: MIT License',
24 |         'Programming Language :: Python',
25 |         'Intended Audience :: Developers',
26 |         'Operating System :: OS Independent',
27 |     ],
28 |     zip_safe=False,
29 | )
30 | 


--------------------------------------------------------------------------------
/.gitignore:
--------------------------------------------------------------------------------
 1 | # Byte-compiled / optimized / DLL files
 2 | __pycache__/
 3 | *.py[cod]
 4 | 
 5 | # C extensions
 6 | *.so
 7 | 
 8 | # Distribution / packaging
 9 | .Python
10 | env/
11 | bin/
12 | build/
13 | develop-eggs/
14 | dist/
15 | eggs/
16 | lib/
17 | lib64/
18 | parts/
19 | sdist/
20 | var/
21 | *.egg-info/
22 | .installed.cfg
23 | *.egg
24 | 
25 | # Installer logs
26 | pip-log.txt
27 | pip-delete-this-directory.txt
28 | 
29 | # Unit test / coverage reports
30 | htmlcov/
31 | .tox/
32 | .coverage
33 | .cache
34 | nosetests.xml
35 | coverage.xml
36 | 
37 | # Translations
38 | *.mo
39 | 
40 | # Mr Developer
41 | .mr.developer.cfg
42 | .project
43 | .pydevproject
44 | 
45 | # Rope
46 | .ropeproject
47 | 
48 | # Django stuff:
49 | *.log
50 | *.pot
51 | 
52 | # Sphinx documentation
53 | docs/_build/
54 | .idea/*
55 | 


--------------------------------------------------------------------------------
/phone/test_phone.py:
--------------------------------------------------------------------------------
 1 | # coding:utf-8
 2 | 
 3 | from unittest import TestCase
 4 | from phone import Phone
 5 | 
 6 | 
 7 | class TestPhone(TestCase):
 8 | 
 9 |     def setUp(self):
10 |         self.p = Phone()
11 | 
12 |     def tearDown(self):
13 |         pass
14 | 
15 |     def test_get_phone_no_type(self):
16 | 
17 |         self.assertEqual(self.p.get_phone_no_type(1), '移动')
18 |         self.assertEqual(self.p.get_phone_no_type(2), '联通')
19 |         self.assertEqual(self.p.get_phone_no_type(3), '电信')
20 |         self.assertEqual(self.p.get_phone_no_type(6), '移动虚拟运营商')
21 |         self.assertEqual(self.p.get_phone_no_type(5), '联通虚拟运营商')
22 |         self.assertEqual(self.p.get_phone_no_type(4), '电信虚拟运营商')
23 |         self.assertEqual(self.p.get_phone_no_type(7), '广电')
24 |         self.assertEqual(self.p.get_phone_no_type(8), '广电虚拟运营商')
25 | 
26 |     def test_find_on_assert_error(self):
27 | 
28 |         try:
29 |             self.p.find(123)
30 |             self.assertTrue(False)
31 |         except AssertionError:
32 |             pass
33 | 
34 |     def test_find_on_ok(self):
35 | 
36 |         res = self.p.find(1521147)
37 |         self.assertEqual(res['zip_code'], '421000')
38 |         self.assertEqual(res['area_code'], '0734')
39 |         self.assertEqual(res['city'], '衡阳')
40 |         self.assertEqual(res['province'], '湖南')
41 |         self.assertEqual(res['phone_type'], '移动')
42 | 
43 |     def test_find_on_no_res(self):
44 | 
45 |         res = self.p.find(1991147)
46 |         self.assertEqual(res, None)
47 | 
48 | 
49 | 


--------------------------------------------------------------------------------
/README.md:
--------------------------------------------------------------------------------
 1 | [![Build Status](https://travis-ci.org/lovedboy/phone.svg?branch=master)](https://travis-ci.org/lovedboy/phone)
 2 | 
 3 | ## 手机号码库
 4 | 
 5 | #### 安装
 6 | 
 7 | 使用pip安装:
 8 | 
 9 | ```
10 | pip install phone
11 | 
12 | ```
13 | 或者直接clone源码。
14 | 
15 | 使用：
16 | 
17 | ```
18 | from phone import Phone
19 | p  = Phone()
20 | p.find(1888888)
21 | ```
22 | 
23 | ### 支持号段
24 | 13\*,15\*,18\*,14[5,7],17[0,6,7,8], 19[2,3,7,9]
25 | 
26 | #### 记录条数
27 | 
28 | 499527 (updated:2023年12月)
29 | 
30 | #### 其他语言支持
31 | 
32 | 下载[phone.dat](https://github.com/lovedboy/phone/raw/master/phone/phone.dat)文件，用其他语言解析即可。
33 | 
34 | * [lua解析](https://gist.github.com/lovedboy/bbff19c91e3d98388d52)，如果不支持bit32，用[这个](https://gist.github.com/lovedboy/fe7750e202572712615a)
35 | * [go解析](https://github.com/xluohome/phonedata)
36 | * [Node解析](https://github.com/conzi/phone)
37 | * [PHP解析](https://github.com/shitoudev/phone-location)
38 | 
39 | #### phone.dat文件格式
40 | 
41 | ```
42 | 
43 |         | 4 bytes |                     <- phone.dat 版本号
44 |         ------------
45 |         | 4 bytes |                     <-  第一个索引的偏移
46 |         -----------------------
47 |         |  offset - 8            |      <-  记录区
48 |         -----------------------
49 |         |  index                 |      <-  索引区
50 |         -----------------------
51 | 
52 | ```
53 | 
54 | * `头部` 头部为8个字节，版本号为4个字节，第一个索引的偏移为4个字节(<4si)。      
55 | * `记录区` 中每条记录的格式为"\<省份\>|\<城市\>|\<邮编\>|\<长途区号\>\0"。 每条记录以'\0'结束。  
56 | * `索引区` 中每条记录的格式为"<手机号前七位><记录区的偏移><卡类型>"，每个索引的长度为9个字节(`<iiB`)。
57 | 
58 | 解析步骤:
59 | 
60 |  * 解析头部8个字节，得到索引区的第一条索引的偏移。
61 |  * 在索引区用二分查找得出手机号在记录区的记录偏移。
62 |  * 在记录区从上一步得到的记录偏移处取数据，直到遇到'\0'。
63 |  
64 | 我定义的卡类型为:
65 | 
66 | * 1 移动
67 | * 2 联通
68 | * 3 电信
69 | * 4 电信虚拟运营商
70 | * 5 联通虚拟运营商
71 | * 6 移动虚拟运营商
72 | 
73 | ## License
74 | #### [MIT](https://opensource.org/licenses/mit-license.php)
75 | 


--------------------------------------------------------------------------------
/phone/build_record.py:
--------------------------------------------------------------------------------
  1 | # -*- coding: utf-8 -*-
  2 | 
  3 | import struct
  4 | import sys
  5 | from phone import Phone
  6 | 
  7 | 
  8 | # build txt record from phone.dat
  9 | def build_record():
 10 |     self = Phone()
 11 |     start_offset = 8
 12 |     offset_map = {}
 13 |     records = []
 14 |     while start_offset < self.first_phone_record_offset:
 15 |         end_offset = self.buf.find(b'\x00', start_offset)
 16 |         record = self.buf[start_offset:end_offset]
 17 |         records.append(record)
 18 |         offset_map[start_offset] = record
 19 |         start_offset = end_offset + 1
 20 |     records.sort()
 21 |     offset_index = {}
 22 |     for k, v in offset_map.items():
 23 |         offset_index[k] = records.index(v)
 24 |     with open("record.txt", "wb") as f:
 25 |         for i, record in enumerate(records):
 26 |             f.write(str(i).encode())
 27 |             f.write(b',')
 28 |             f.write(record)
 29 |             f.write(b'\n')
 30 |     with open("phone.index", "wb") as f:
 31 |         current_offset = self.first_phone_record_offset
 32 |         while current_offset < len(self.buf):
 33 |             buffer = self.buf[current_offset: current_offset +
 34 |                               self.phone_fmt_length]
 35 |             cur_phone, record_offset, phone_type = struct.unpack(self.phone_fmt,
 36 |                                                                  buffer)
 37 |             f.write(str(cur_phone).encode())
 38 |             f.write(b',')
 39 |             f.write(str(offset_index[record_offset]).encode())
 40 |             f.write(b',')
 41 |             f.write(str(phone_type).encode())
 42 |             f.write(b'\n')
 43 |             current_offset += self.phone_fmt_length
 44 | 
 45 | 
 46 | def build_dat(version=b"2312"):
 47 |     self = Phone()
 48 |     index_offset = dict()
 49 |     init_offset = 8
 50 |     records = []
 51 |     record_length = 0
 52 |     with open("record.txt", "rb") as f:
 53 |         for line in f:
 54 |             index, record = line.strip().split(b",")
 55 |             index_offset[int(index.decode())] = init_offset
 56 |             init_offset += len(record) + 1
 57 |             record_length += len(record) + 1
 58 |             records.append(record)
 59 | 
 60 |     phones = []
 61 |     uniq_no = dict()
 62 |     with open("phone.index", "rb") as f:
 63 |         for line in f:
 64 |             phone_no, index, phone_type = [
 65 |                 int(i.decode()) for i in line.split(b",")]
 66 |             if phone_no not in uniq_no:
 67 |                 uniq_no[phone_no] = 1
 68 |                 phones.append((phone_no, index_offset[index], phone_type))
 69 | 
 70 |     phones.sort(key=lambda x: x[0])
 71 | 
 72 |     with open("phone.dat2", "wb") as f:
 73 |         f.write(struct.pack(self.head_fmt, version, record_length+8))
 74 |         for record in records:
 75 |             f.write(record)
 76 |             f.write(b'\x00')
 77 |         for phone in phones:
 78 |             f.write(struct.pack(self.phone_fmt, *phone))
 79 | 
 80 | 
 81 | def test():
 82 |     phone1 = Phone()
 83 |     phone2 = Phone(dat_file="phone.dat2")
 84 |     phone1.get_phone_dat_msg()
 85 |     phone2.get_phone_dat_msg()
 86 |     for n in (131, 132, 133, 134, 135, 136, 137, 138, 139, 152, 153, 154, 155, 186, 188, 192, 193, 195, 199):
 87 |         for i in range(n*10000, n*10000+10000):
 88 |             assert phone1.find(i) == phone2.find(i)
 89 | 
 90 | 
 91 | def get_phone_type(phone_no):
 92 |     if phone_no.startswith("192"):
 93 |         return 7
 94 | 
 95 | 
 96 | def add_phone_no():
 97 | 
 98 |     city_code_index = {}
 99 |     with open("record.txt", "rb") as f:
100 |         for line in f:
101 |             index, record = line.strip().split(b",")
102 |             city_code = record.decode().split("|")[3]
103 |             city_code_index[city_code] = int(index.decode())
104 |     for line in sys.stdin:
105 |         line = line.strip().split()
106 |         if not line:
107 |             continue
108 |         phone_no, city_code = line
109 |         phone_type = get_phone_type(phone_no)
110 |         index = city_code_index[city_code]
111 |         print(f"{phone_no},{index},{phone_type}".format())
112 | 
113 | 
114 | if __name__ == "__main__":
115 |     if len(sys.argv) > 1:
116 |         locals()[sys.argv[1]]()
117 |     else:
118 |         build_record()
119 | 


--------------------------------------------------------------------------------
/phone/phone.py:
--------------------------------------------------------------------------------
  1 | # -*- coding: utf-8 -*-
  2 | 
  3 | import os
  4 | import struct
  5 | import sys
  6 | 
  7 | __author__ = 'lovedboy'
  8 | 
  9 | if sys.version_info > (3, 0):
 10 |     def get_record_content(buf, start_offset):
 11 |         end_offset = buf.find(b'\x00', start_offset)
 12 |         return buf[start_offset:end_offset].decode()
 13 | else:
 14 |     def get_record_content(buf, start_offset):
 15 |         end_offset = buf.find('\x00', start_offset)
 16 |         return buf[start_offset:end_offset]
 17 | 
 18 | 
 19 | class Phone(object):
 20 |     def __init__(self, dat_file=None):
 21 | 
 22 |         if dat_file is None:
 23 |             dat_file = os.path.join(os.path.dirname(__file__), "phone.dat")
 24 |         self.dat_file = dat_file
 25 |         with open(self.dat_file, 'rb') as f:
 26 |             self.buf = f.read()
 27 | 
 28 |         self.head_fmt = "<4si"
 29 |         self.phone_fmt = "<iiB"
 30 |         self.head_fmt_length = struct.calcsize(self.head_fmt)
 31 |         self.phone_fmt_length = struct.calcsize(self.phone_fmt)
 32 |         self.version, self.first_phone_record_offset = struct.unpack(
 33 |             self.head_fmt, self.buf[:self.head_fmt_length])
 34 |         self.phone_record_count = (len(
 35 |             self.buf) - self.first_phone_record_offset) // self.phone_fmt_length
 36 | 
 37 |     def get_phone_dat_msg(self):
 38 |         print("版本号:{}".format(self.version))
 39 |         print("总记录条数:{}".format(self.phone_record_count))
 40 |         print("dat_file:{}".format(self.dat_file))
 41 | 
 42 |     @staticmethod
 43 |     def get_phone_no_type(no):
 44 |         if no == 4:
 45 |             return "电信虚拟运营商"
 46 |         if no == 5:
 47 |             return "联通虚拟运营商"
 48 |         if no == 6:
 49 |             return "移动虚拟运营商"
 50 |         if no == 3:
 51 |             return "电信"
 52 |         if no == 2:
 53 |             return "联通"
 54 |         if no == 1:
 55 |             return "移动"
 56 |         if no == 7:
 57 |             return "广电"
 58 |         if no == 8:
 59 |             return "广电虚拟运营商"
 60 | 
 61 |     @staticmethod
 62 |     def _format_phone_content(phone_num, record_content, phone_type):
 63 | 
 64 |         province, city, zip_code, area_code = record_content.split('|')
 65 |         return {
 66 |             "phone": phone_num,
 67 |             "province": province,
 68 |             "city": city,
 69 |             "zip_code": zip_code,
 70 |             "area_code": area_code,
 71 |             "phone_type": Phone.get_phone_no_type(phone_type)
 72 |         }
 73 | 
 74 |     def _lookup_phone(self, phone_num):
 75 | 
 76 |         phone_num = str(phone_num)
 77 |         assert 7 <= len(phone_num) <= 11
 78 |         int_phone = int(str(phone_num)[0:7])
 79 | 
 80 |         left = 0
 81 |         right = self.phone_record_count
 82 |         buflen = len(self.buf)
 83 |         while left <= right:
 84 |             middle = (left + right) // 2
 85 |             current_offset = (self.first_phone_record_offset +
 86 |                               middle * self.phone_fmt_length)
 87 |             if current_offset >= buflen:
 88 |                 return
 89 | 
 90 |             buffer = self.buf[current_offset: current_offset +
 91 |                               self.phone_fmt_length]
 92 |             cur_phone, record_offset, phone_type = struct.unpack(self.phone_fmt,
 93 |                                                                  buffer)
 94 | 
 95 |             if cur_phone > int_phone:
 96 |                 right = middle - 1
 97 |             elif cur_phone < int_phone:
 98 |                 left = middle + 1
 99 |             else:
100 |                 record_content = get_record_content(self.buf, record_offset)
101 |                 return Phone._format_phone_content(phone_num, record_content,
102 |                                                    phone_type)
103 | 
104 |     def find(self, phone_num):
105 |         return self._lookup_phone(phone_num)
106 | 
107 |     @staticmethod
108 |     def human_phone_info(phone_info):
109 |         if not phone_info:
110 |             return ''
111 | 
112 |         return "{}|{}|{}|{}|{}|{}".format(phone_info['phone'],
113 |                                           phone_info['province'],
114 |                                           phone_info['city'],
115 |                                           phone_info['zip_code'],
116 |                                           phone_info['area_code'],
117 |                                           phone_info['phone_type'])
118 | 
119 |     def test(self):
120 |         self.get_phone_dat_msg()
121 |         for i in range(1529900, 1529999):
122 |             print(self.human_phone_info(self.find(i)))
123 | 
124 | 
125 | if __name__ == "__main__":
126 |     phone = Phone()
127 |     phone.test()
128 | 


--------------------------------------------------------------------------------