一直想寫一篇關于字體反爬的文章，但是由于時間問題一直拖到現在，字體反爬的網站有很多比如貓眼電影專業版、汽車之家、58同城、大眾點評等，今天我就拿58同城試刀，當然58同城不止求職簡歷才有字體反爬，比如租房模塊也有，大家有空可以自己去研究。
目標
抓取58同城深圳地區的簡歷信息，由于此次主要是為了破解字體反爬，而且簡歷信息只有姓名、性別、年齡、工作經驗、學歷才有字體反爬，所有我們只抓這幾個信息，并且時間問題我們只抓取第一頁
目標分析
首先用瀏覽器對頁面進行抓包分析，找到信息所在的文件：
通過抓包發現信息就在當前頁面，并沒有JS加載渲染以及Ajax異步加載，既然找到了信息的地方，那我們先檢查一下我們要抓取的信息：
一檢查就傻眼了，發現我們要抓取的信息全部都亂碼了，我們再查看一下網頁源代碼：
我們發現有一些字被編碼了，這到底是怎么回事呢？其實這就是我們今天要講的字體反爬，網站采用了自定義的字體文件，通過字體映射然后在瀏覽器上正常顯示，但是爬蟲抓取下來的數據要么就是亂碼，要么就是變成其他字符，那么知道了字體反爬的原理，我們就要找到字體文件，字體文件在哪里呢？字體文件一般抓包的時候可以抓到或者定義到了頁面里：
可以看到兩個地方都找到了字體文件，而且我們發現字體文件進行了base64編碼，我們直接把字體文件下載下來，然后用百度的FontEditor工具來查看字體文件：
查看字體文件字體對應的編碼然后把它們帶入到網頁源代碼中的編碼，你會發現居然還原了內容，看到這里你也許會說這還不簡單，我把這些字和編碼組成對應的映射集合，然后每次抓取的數據按照這個映射集合來替換不就是了，確實可以這樣，但是如果你刷新頁面后，你再看網頁源代碼中的編碼和字體文件，你會發現字體文件對應的編碼變了：
我們發現字體文件對應的編碼變了，但是所有的字沒有變，那我們該怎么辦呢？這是就要用到Python里面的字體庫Font-Tools了。
字體庫Font-Tools介紹
Font-Tools?是一套以?ttx?為核心的工具集，用于處理與字體編輯有關的各種問題，程序用?Python?編寫完成，代碼開源，具有良好的跨平臺性。Font-Tools?由以下?4?個程序組成：
- ttx?可將字體文件與?xml?文件進行雙向轉換
- pyft-merge?可將數個字體文件合并成為一個字體文件
- pyft-sub-set?可產生一個由字體的指定字符組成的子集
- pyftin-spect?可顯示字體文件的二進制組成信息
Font-Tools?原本是托管在?Source-forge?上的項目，由于原項目長期停滯，Be-hdad?在?Github?上?fork?并繼續進行開發。由于?Font-Tools?基于?Python?寫成，在安裝?Font-Tools?之前需要首先安裝?Python。
字體庫Font-Tools基本使用
-
TTFont()用于打開本地字體文件
from fontTools.ttLib import TTFont
# 可以是.ttf類型的字體文件也可以是.woff類型的字體文件
# font=TTFont('58.ttf')
font=TTFont('58.woff')
-
coordinates用于獲取字體坐標
from fontTools.ttLib import TTFont
font=TTFont('58.woff')
# 獲取編碼為uniE0AC的字體的坐標
x_y = font['glyf']['uniE0AC'].coordinates
-
saveXML()將ttf文件或woff文件轉化成xml格式并保存到本地，主要是方便查看內部數據結構
from fontTools.ttLib import TTFont
font=TTFont('58.woff')
font.saveXML('58.xml')
把字體文件轉化成xml格式，以便打開查看里面的數據結構。打開xml文件可以看到類似html標簽的結構：
而對我們有用的是
<GlyphOrder>標簽對象和glyf標簽對象：
點開標簽內部，
<GlyphOrder...>內包含著所有編碼信息，注意前兩個是不是0-9的編碼，需要去除。
<glyf...> 內包含著每一個字符對象<TTGlyph>，同樣第一個和最后一個不是0-9的字符，需要去除。點開<TTGlyph>對象，里面的信息如下，是一些坐標點的信息，可以想到這些點應該是描繪字體形狀的，而且我們發現不同的字體文件雖然編碼不一樣，但是只要它們對應的文字一樣，所以我們可以在<TTGlyph>對象里找出坐標規律，這就是我們破解字體反爬的關鍵所在。
破解思路
先在本地保存一份字體文件58.woff，并通過FontEditor工具確認編碼和數字的對應關系，保存到字典中。然后重新訪問網頁的時候，把網頁中新的字體文件也下載保存到本地58tc.woff。先獲取58tc.woff中的<GlyphID...>里的編碼name的值(uni編碼)，再通過uni的對象獲取其對應的TTGlyph對象的坐標，然后取前2個計算差值，與58.woff中的每一個TTGlyph對象的坐標差值相減注逐一判斷是否等于0，再根據TTGlyph對象對應的編碼，在字典中找到對應的數字。
import requests
from lxml import etree
import re
import base64
from fontTools.ttLib import TTFont
url = "https://sz.58.com/searchjob/"
headers = {
'user-agent': "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) "
"Chrome/73.0.3683.86 Safari/537.36",
}
response = requests.get(url, headers=headers)
html = etree.HTML(response.text)
font_face = html.xpath('//head/style[1]/text()')[0].strip()
# 提前字體文件
base64_code = re.findall(r"base64,(.*?)\)",font_face)
if len(base64_code)!=0:
base64_code = base64_code[0]
woff = base64.b64decode(base64_code)
# base64 寫入字體文件58tc.woff中，一定要wb方式寫入，每次運行代碼會覆蓋文件
with open("58tc.woff","wb") as f:
f.write(woff)
# 打開下載保存好的新字體文件58tc.woff
font = TTFont('58tc.woff')
# 打開本地保存的基本字體文件58.woff
base_font = TTFont("58.woff")
# getGlyphNames()[1:-1]和getGlyphOrder()[2:]結果是一樣的
# uni_list = font.getGlyphNames()[1:-1]
uni_list = font.getGlyphOrder()[2:]
# 定義一個臨時存儲新字體文件映射關系的字典temp
temp = {}
# 把本地字體文件的映射關系用base_uni和base_value兩個列表映射保存
base_uni = [
'uniE0AC', 'uniE0D6', 'uniE189', 'uniE19A', 'uniE1BC', 'uniE441', 'uniE47A', 'uniE4BE', 'uniE4F1',
'uniE587', 'uniE5B0', 'uniE5CE', 'uniE615', 'uniE632', 'uniE701', 'uniE87F', 'uniEAC1', 'uniEAF9',
'uniEB60', 'uniEB96', 'uniEBB0', 'uniEC03', 'uniEF5F', 'uniEF8B', 'uniF037', 'uniF076', 'uniF0A0',
'uniF13A', 'uniF14D', 'uniF1DB', 'uniF264', 'uniF2D1', 'uniF31A', 'uniF386', 'uniF406', 'uniF46B',
'uniF49A', 'uniF4DB', 'uniF5F0', 'uniF607', 'uniF62A', 'uniF6E6', 'uniF772', 'uniF787', 'uniF7B9'
]
base_value =[
'7', '下', '王', '周', '專', '0', '女', '博', '楊', '李', '校', '技', '屆', '8', '男', '科', '中',
'趙', '生', 'M', '9', '以', '經', '6', '陳', 'A', '驗', '黃', 'B', '5', '士', '1', '張', '碩', '4',
'高', '無', '大', '吳', 'E', '應', '3', '2', '本', '劉'
]
# 循環對比
for i in range(len(base_uni)):
# 編碼字體坐標轉化成了列表，列表里是一個個元組，元組里放的是(x,y)坐標
new_glyph = list(font['glyf'][uni_list[i]].coordinates)
# 用前兩個坐標作為取差值
new_glyph_difference = [abs(k[0] - k[1]) for k in new_glyph[:2]]
for j in range(len(base_uni)):
base_glyph = list(base_font['glyf'][base_uni[j]].coordinates)
base_glyph_difference = [abs(n[0] - n[1]) for n in base_glyph[:2]]
# 比較兩個差值是否為0
if int(abs(sum(new_glyph_difference) / len(new_glyph_difference)-sum(base_glyph_difference) / len(base_glyph_difference))) == 0:
# 把編碼去掉uni三個字符然后轉換成全小寫，再拼接成網頁源代碼一樣的編碼格式，最后把映射關系存儲到temp字典中
temp["&#x" + uni_list[i][3:].lower() + ';'] = base_value[j]
# 構造正則表達式用|匹配左右任意一個表達式，替換編碼
re_rule = '(' + '|'.join(temp.keys()) + ')'
# 把所有的編碼替換成文字
response_data = re.sub(re_rule, lambda x: temp[x.group()], response.text)
data = etree.HTML(response_data)
personal_information = data .xpath('//div[@id="infolist"]/ul/li//dl[@class="infocardMessage clearfix"]')
for info in personal_information:
# 姓名
name = info.xpath('./dd//span[@class="infocardName fl stonefont resumeName"]/text()')[0]
# 性別
gender = info.xpath('./dd//div[@ class="infocardBasic fl"]/div/em[1]/text()')[0]
# 年齡
age = info.xpath('./dd//div[@ class="infocardBasic fl"]/div/em[2]/text()')[0]
工作經驗
work_experiences = info.xpath('./dd//div[@ class="infocardBasic fl"]/div/em[3]/text()')
if work_experiences == []:
work_experience = ""
else:
work_experience = info.xpath('./dd//div[@ class="infocardBasic fl"]/div/em[3]/text()')[0]
# 學歷
educations = info.xpath('./dd//div[@ class="infocardBasic fl"]/div/em[4]/text()')
if educations == []:
education = ""
else:
education = info.xpath('./dd//div[@ class="infocardBasic fl"]/div/em[4]/text()')[0]
print(name, gender, age, work_experience, education)
結果
下面是58.woff文件的下載地址，直接復制這個地址到瀏覽器下載，下載好放到項目目錄同級下并改名為58.woff
58.woff文件：data:application/font-woff;charset=utf-8;base64,d09GRgABAAAAACFYAAsAAAAALRgAAQAAAAAAAAAAAAAAAAAAAAAAAAAAAABHU1VCAAABCAAAADMAAABCsP6z7U9TLzIAAAE8AAAARAAAAFZtBmVhY21hcAAAAYAAAAHxAAAFTgRm1uhnbHlmAAADdAAAGm0AACAYICFUu2hlYWQAAB3kAAAAMQAAADYbNsxXaGhlYQAAHhgAAAAgAAAAJBFlBgFobXR4AAAeOAAAADkAAAC8Ri3/t2xvY2EAAB50AAAAYAAAAGC8tMUGbWF4cAAAHtQAAAAfAAAAIAFCAIVuYW1lAAAe9AAAAXIAAALQd5CEoXBvc3QAACBoAAAA7wAAAe1VQzLJeJxjYGRgYOBikGPQYWB0cfMJYeBgYGGAAJAMY05meiJQDMoDyrGAaQ4gZoOIAgCKIwNPAHicY2Bk+8g4gYGVgYNVmD2FgYGxCkKzCjK0MO1kYGBiYGVmwAoC0lxTGBwYKr7v5Cj/+4LhM0c5kwRQmBEkBwDH0AwveJzN1LtO22AYxvG/kzTpgbb0ENqmB9pCT2nLwELFVC6AWygMbGwMCLGwESF2hkxcAENWJITEDSCxsGSJJdugGCl2Ymwx0ce87J0i1dEvsS3li/U9zxvgDlCUH1LS6RiOziiM6q5zc7/I/Zv7JcfT9R9+6zvvWOvsdU7dLbfpHnhz3rp36MV+w2/5x0E1mD5zzje6R92rcDHcCVsXxd5CbzuaiVaj3Xg2no/b/aX+yaA22EzKyXLSTNqXUVpJ62mQrWSNbP/6Wr8z7PWHdzjao+G98vULyqSkFMpUuMs95fOAER7yiMeM8oSnPOM5VcZ4wUteUeM1b3ir1MZ5zwc+MsEkn/jMF77yjTrflfNPprR4eah784+N+0+Okfyt8Ov2SrvC2i09YmfPKAU6pyafIHfL5FPkNk0+Xe6BUVp4c0a54a0bJYh3aJQlXmzyqfMbRvnit0z+dP6xUeYEVaP0CaaNesCZY9QIzjeMukH3yKgldK+M+kK4aNQcwh2jDhG2jNrERdGoV/QWjBpGb9uoa0QzRq0jWjXqH9GuUROJZ406STxv1E7itlFP6S8ZNZb+iVF3GdSMWsxg06jPJGWjZpMsG3WcpGnUdpK2Ue+5jIwmgLRiNAukdUP+GRjNB9mKyf8ps4bRzJDtG6b+Ak8hJJMAAAB4nFVZD1xUdba/5/e793dREZmBmUEkYhhmBhERYZhBIhbJkIglVlkl1lg0Y4klUpbINUU0JCIiJDIydJHUNTMy1/W5ZmbmukpuGblmZqbGUzSXNXNNYe7hnTvYe/v0w2UGuHfu7/y+5/vnXAkkafiqlCAFS0ySPImm4PDgaIn+6T+/JMcpNyR/esMkbjNKQVa3xWmV427gZesxOwRf0iQZFH5KeQbnaJPwOzCwffDyy8VL75wf7yeJeGm8RCda1CibYk7iHklRo8DgsKjGIB5lUR5V2xUFp/4GyxibBz3grsYezFtd+hpjMpovaNdFfGaRoryp4kAtXj/VNQB+Q2UKyCLPe4Iny4z9dK816ueCSXfTGqwGh1VNsnq41T0F6IPtisFqsqsG30u3ZzZf+DrOX8S2ca6NpguITlnGeqxlezH8FC+lNx/f/ovMwlmponSWwMmnA8dnFTCmHeqkQ3tkjRBzjfSZsiQhqqtEgRQuRUvTpSy6jUBVqA6n2+MeD1ZzkEpvnQ6nOTEhyWBXjSpYDXbF5FSNBm5NsqtJdovHmWhRrWaT6lRMVB6TlQozDgweWB4b356SXBMdIboY8FvN3X4hB0ubh8YsFjIwXoCv4pvwqtjGGO4thg1+QnQIAVuEKMI9MP1L6Bgq4gv//Mdja3g1yvv7T2anHji4Fbq439LBmnREmYnUEv96pZXOzxyLj2p9jNUy+sIH2XCPojTI8gv2efljQx4VQn+rHQZ/GTjeECO1xn3qX0SNNJXeyApEjQWP06Mjw275GVg9TmtQInePBVOUXeFUc1/ZTQJUs0WV7HAJRkEiHMKv0ILdijIfbbRCaHrxMVX7txBs5uO7Hz8J6yHdexd/wvtcufyVLA9eZo/CXO5Zg22W+uYFOBojwFG9Bp6jm8qbcWtJWCpjt0JnZcJovKlw1s6ZggMzfwH/ZqxjKf4tOP3+dvA7Q/fuA4tcI3Kk0RLtYVCgZI0wBEq2SHqpv7JG1EIgMLiK3w1eRe0ITAMX9uKHfCysg83ei7gE50MjrNReZb9jz/uuN2xUB8QxwoBd0pFtlxWHM8oeZTcG6ZVweyyCNpiKwEDmVqbKFkldUYF/xhfwC21FVAqbCXdhP/Y2lCFtAEssnL4z70H5Teoq/PVc/AaKX9Ier3gM2CDYstd++P63b2YWrnz/hechGkJLY/FG1OCJcaOWad/jl9v4OTx1quppMNESR9Eeeem+tksxUqpUIM2T5kulUrlUKS2hPUuymmxuyxRQLKqONiu1Ct2y4k7S0ejRf+qkg91icEKigf6bqFkNDo9qsLkJw0mJZg8taRxYTcG2yBF83wt0uoV+Z1X0Vdut5iS60ET6JO6hxTvA4aTv4h1sbGFcicE9jMXcPZPaqUOWoQCvsEjGWjh/FS7LstWpHWEMFr9Se7u0p+1A/eK3bONqexQGl2TWRsiFDNpi3H+tcqtfYFdNBZ7UohXgkVCNjRpCJUG6mS/BRln+JVi9DVCJzSmFLbXF1Tx5SV1aaxmcqqh9YNoiRVmYIGTGA+uiC5gsJr0cU75MiFlx9AGxBYqihTC2gbm0Yy2CyUPBnAv5ahhoD1elhs9ZYPa7eEFOzw8o7+X813HwbkNDWTgumnyfEPXemIVFJWULlQMZqSVpbokN/xMPqb8R86VJUh5V3ppEdbZQEbkgavAVz2L2uIOoBRxO6adveiuNAgJQEuHH6VbvBu52WkwOp2r4GejldBociaqBqQ7lB00mjKpwknNvD29xjM7O3JPfGWCubd1mNAb3F2v9yfHgzb1QhvmxLqhZgNdpsXBxKP3kHLwILlpXZkyVLB8q3jOl5DlrUv2S6rKqFc7k8vYwpUTfqp/NxkBtdH9+ARNz8o2y2HsFDmjlWJXrhnIWj7F4KCMNQmBJYVY9OL3nIQtqG/4auZH4ORsPEoM36EDunjC5taGkqWHoOudyPracmiON9OLwKMLom8TZ9+oKkWSPUlSdsU1EhNQxhv8HL/p9oslqclgVSRE+BSHoSnqzgdWtZvEOWIqXtUoTByZvQhfPkWXvroHqfcK4c2kVpmL0LO5o4Y5x1HgHYCu0uaI1xAMx2WHYr50KyWR+EKnlCgbPzDmgMO5Sf97ZF+p8DL0X5BklxtILsiwfGxhoEgKPFM7K2A7JWEXwgTY8czZ1VjdIy7L1NSm+vtPEYckqeaRM6efSbFqZjwo8FtrLxCA1kTqH+snkGdlIm76NYLS4gyxMNesrcwp9UR7HGLD6uka9I1sK/YHDabT7TkzqnYMDnEfepD1qFaKfULqm/KPMZnd5WnFraX0ov1q+oa14Xd26C1AY78b20xexPjoZKs9uq9rCwmKy6nPdq9gqIbRHtcsygLVxx4HUjOMsovz0oaXdpMIlUJySJ0ToP97Bp03OuCS6/l73rvy8osahUjkWuw/NgW7s270b4pJxy4F9xSnzag6caU3PW3XsAGSyLZVtKY3LUmalKMq4B+tn4HGMiQG/XYX9N0qvS5z2PdjPIk5KsdJcYibiI1qeievEYSC+1NlFF0Wn6iZFZMpIUeSRonAqhEHvCv1PVQt32ywkqPQ6mPyEhwf/n+AK6jIqK58CZF6ooW+Stoy2h84nbWARpPznBOcbQsHxIhJsc3Jc+N4FvI2dMenwmHb2Se8rr76IN3VhGv3LLM7YN6kzz/r/7gm5jzrmSN4i/4Dq1DnPPS8WaSsLe6/yrDpFGfoBYuxPQAYH+YPS7+IXfOR6g7EluOjWu48UZNS9f/iF9NyqPdehoW1SDLyTP3Yi1FEnyljLeSP5n5X2qOLUB7BxFPafIBnIn+NfyOInT4EUPHT6nCSpVLO71CfFY+TSbJJTipNcUjIxu84oHosutEan6iTPQ9hw0sItXH9jMBKdGA3UUFafLJuTdDfkUciHELUYrCyBZc78OUvAI5BDmI+s9PoR/T3hxH6RpygJCwbPyDjoTjur7Bs8o0ShG3bSneYq8DbmCpEGVV4/2Im5+hd/mArTR0v3Ppx4hPOq+3gc52pWqLZdliuwixijOyMdolkYu94/FB135cY12PT4O3S1YghpgBAdE6TLz4jZUgitL5Z6xuX2xEGSKw0SwsEULNRwCAyAiDjwBNPuupNcjggRFGxOcLsckaK5s+ax6sVHf+hZ9LuSmg01/fjNxdoD215sfOdd/H5HctuJ117+eg2cqb8QP21nacXu0t/sfrLkTylTL+CPn1VV9TY1bnnzudp33mK5pa+sPd76kkQeB6+rvyePMMZXaTex0wydn3TG0aXTmETG0pNIus5UZvVRVITH7bEyxRkVZNU5fixYFOuIwlLxo+4Gj08Wzep3wpUx1T7obRd1rHfMqK4xo/lrwUYH9h7o8TOfgTJzoL938AFtYilrm63dD2bmxvEwBrco7ChzE/Q+x/WcJWo9MPfhennwFnMTsHuELHK8mdSh+fkw7S2tNrf40QcR23Cf34zykJ2VjXj/+y2vwUNM1lniWNixnFLO/8DGXiSSzs9Xzue84vOyw2F+qjhCLsZNuMoYYS0Cie5no+xkYgx2C9HvHdaiShi5k/hoMplVC/1KsXDdJKhK0Ije+4739VXhIcha1xAQCnVQ49rVBaOS8GBBy9JFtZW5dht57+quSLaVBy6eCKn+uisoN903pA2dnXe2sbQZ9xX3VZUWQQYkwJJmcAT6Z2eHpbaFGJH8QfOFvGXVlS1l3lhaFGwGw3E8dnLWwUyZ2mcoJbwlvBAFSQHLNE2VKxduq61uxZ6SnWVLF+hYw6/VWWI9Ye0uKUrf2SBnosFK7sbktNLm2iiYkNCMI1o06CbGw51K0DrOhNjOR28i07JOlrdTo2hdUY5CEayt+xxeHmxhRULcwLVcvgxsJ+Z4F+UHOHDrRu1T1obdsKWcTAvzkvNmD8MbBPx3df3DQXVIvC2lSA9JksUYRJAyUzFVt0pO2jDCh7a7QdUrb9AtlX53PkPNzRbZLlTZHsV1+fAYabcsjBuDVMYZp6YPYlM7/Ntk45xN2NGmC1V9GraQq1p0z11LKSnkjsatsE2PPZwDZRvcynl1GyALiApGifx1cHdyRTTvXbJ3Z0fl2ubde3Y0dHbsOpWWdn3vsWIQLGR7W08plxXlZ6nPk4NLLo2C0Zw3L1OUJjwyN/dZxOv4yahn0/FwNBMBhVlMDoY8xO4iUZOPJ/EsxAIxbhVuJ2k4uRDMIPCTYSkjGiJgBhSO5LlWP0X8Q5rsy3O+Wlh1LEpBuqH26MRPoc6pGKJ0r2ol+tMVw654HuSdWirn7JB2bfXbMRkQumlTUjp8qCgtqzcy70ptNfyiBL4k5UslePBSUgHvNG+QHA1413L45NrBjGxIhSOxrfXPkulQsIA2GxNotxdPxEnrDz8bf5+iTPHd35CqiAApkkTf4PLYSG7IIRPf2swmEnWPhYxaotsDbtVCcU+kjlNCgxEeiUlFZgwToz9oWQFHXQbVZYLP/vUOOSVFZit2Loodepw+Sm6OXl7+iVcVYzkfOr10xkuJciTdp6ZoXrzu8029coIIkizElZKH+NBoEE5BISbJJSUmSBRoiJWCfiJKW6TYXNt2/L8ZK/n+o2EJQr65BDLuwC/eWLy488WlWzY/t3zT/mxybnGMHQT/k31gx1dwM87GhCQ5rOO//vj0H7/aT/Ss98059ZLokAxShOSgTUo0BHKjzgdMF+cgI6c1S5SAnW6nzgpuoAjMnIq2qfnWmca/H7uE8XFpcKt7r6y1CYpw1RmplUVMbDgDfvWbOjpqRQW2a1/ge8PSrRr8b9x4OC2nn1o/lhVqgatSsTc3zr0hv1qpEGKwrE6ZdwJPFYDrhI+7fqt6xV/pvmzUS9N9jsJiovawUB7zualRPv0jn2FyqIoewu0KuD3kwRWn/s3p6x7icIsEdIZJVngm+/rSqh0w+ceSYQk/G8CXopNZJiz0VrEU8oC57ChMCYnCNgCYghO9n2L5oSv+4iZ2Rb2+8WCZ4/Te9r37f8h9qA0Ro6ETwgVew3rl7y2y3Ppe5+trG3Zs/WhNVsZ+6BmMu3aLGrOFiAFDCKTkUraGOvLTI1PbOn/wH9UAkbA1Fddhn1jxcjYWawH5WbMykhfe8c8fy+nCXzLrKATaaMYT3YQGU7CeaHUYJFjgP2DAv/l93dEbTKn4/ug1/O6Lq/gDzIeIzUVa3hvP1Xa++kLdJmVSGnbhZ//AwS+/xTOwGGZT7v12shdeO72vrXPnLmlk7jEc6zdGfELamE2pZo7uQkZyLjWi3pxOHfROTjJocijq/9l4fQqS5PbYuHHEpnncTp2qiHn1iQgphu6APYQopguO080nwKSpce2xk5gmhPfTXxVADudtIWFEpfA8vg6fHMeLflfrl/obm0prsDD/N7+6FZdyM3dZYNiCOY3xYe2F2ehdS24qnTxfQ9QmXluzr3rVdi1t1db2+i3KjqbWmOVPB9TePj//R736+/6JwzLwFgGKtvQKm5UfWMMKGWu9cNPbh51xAf6z5oT4+V/vOkzZO/x2Jb9J2LzuuHcippefulk0LB0ugMirebcOFO65M2dQ3hZTJX9ybJLdxklEVW6DRKMliPiZ6IIrb3/9ogEH8et5r3zNck8+pY6/MmasMiagPwpaQvQJRoEYi+fN2gpWd/fPMx3aMrqkOnweB9SjokKaIE2khKErdZauHsATjVwX7DGQSBJgvVNxipbWIPLBOrx9AuJOAtVIhlriI1FT94Yey0jNfUfFbRdUfBgV1xN7lx1TWFFwTE2Av0Asbx0d2LKoyc/fDyPSWQbrbVnUXl0HxWR94xX2meM6lKTHYOfZY9galQ7NJ0hTkr1HxNXmfHdrCs5rWpDRluEt7idTI/ehPKidZmZcyKIbSA8kLKT+OC6y843bzlB+6B7cCwuhLp5w3z3XUxa6QBxJx1mH8lJBhgpci+dc2RAMh7SIxnGrHXjiXAwEd8Vi39kEkILjMvLS4jOpVn7EV3vVt0WjlEBeOU26n9D6kJQvFfqyZpRH0RmBMEpqq880CLuUPfWIGeVR6UgkQcAkyiBZsRk94JH1uaUvenCLPJI7qXYwMu9IIk2mg1nVWJv2h/uwi2AXjk/AvVhNzA3n2ikHs2gy2e1wDmMvXmeReA9OZ2vXepvwLP14Lcn28ymUSyPTsdF7nUVTrtNOiQLchN1hMVAqtyOBimlIB4aSVgaIjJSXjnR2XjdlibHPHCDoPpJUvM9urbZ0jNrXoIA40MB5t/dQqixbmBMKm7XdSxYtbKhRUu6+r2iwr2btvNScAp6wrxOjXrPV7luOq/clKyNcP6zJr4qxxKnjKctKxOZWS4JZd+Im3ZDLNjD4eMXKfd8DrX8dBw8+cezGmhUb3sfz5/Dfu9dsxjNHrq1/C1/lupB9tLH20ETZ+HHHkR+VIpzw4vKvtCe0vjUrYfSd+e0V2SaM+qwYdPpyJpkj3IHEWmoQlZkFcY8pQZZt+Dme/u16uP9vn77f+tQ0dkL7bMKEiBdvwnj49lu8Mv1IyZ8gevNYbtyOVfAYLlLh375rL1cPiVek0BEvQeJgtHCK4nbyVvpIlhKnXfV0syveBrbyR2q9XbLMHw0iv9YIh8BBijyYclC5ixdtoa2sSIF0bb1BiLpIb5eilCn3COG5M4OmwpnF/fq83G4IAOK6KE+iOYizvLivHnkGQh7MZkw2K5PjKi0HIl/J+Cc+4H1qRFdPqR+IZaSpvnkBkQRRqEcly+dRHfY7/WqjkGumztSJMYm8qpPcqcNGXpFSBvcoFEVs/9n07kSLkeys2wJRYJEU9tYAxYd7FnC+iFBRgTf+/DrsyHVh8pp3y/YFGOsPZseUZw3ITKtjtXJ1KPEhx88jkmFDSECId9Zuf3N3cZ3R7I+OlJZTlA0xmHF5KsyBmGo8hh08hrlXr6ELe3fdc28r1QstwznpMJtZtVmu8nkZM6pcMY1ZRaw10cYdWrEQcRTsGW4ha1W5Fy+KnKLAug4IZWVU3+zgsDjAR3ogbBFe1Ge28vBl3EWcF0g6Fy/dJ+VQyfRsQrHKpMsN2UA9epA51AdBHpIYAozwUGUUg83D9Tm8Tbcqqkf3AyNexeM22G1BKqkWSdXFBTjIfpFnx153NizD+pKSBleKi5dBhnvCBGwrl+u1GyHKJJcC77fKsjZPH7lAe+sMMoqnSvHqtXosi4+D+opMLfS6LEd+wqh6MtTD4Tpqzbr+Y94ZtfSCeUuLUtmevuUvc97UtNi0gbRmaMDBi1dolwa4jJ/gua05x7saz23Hm53ZaW1nmsDcg8bDVM2/j5f+d3a9TURLgZKJUGJItKg2g8Op0hqT3BIxkUEw/n6/ovQvNDxAMWXyh1C/As3wY+mAiPYCbMQ2wvbDEPTal8x92wvswnu4UdLz7nCCn58YkMaSf4qjtJup13gEXkl2VX86QdJlpI/5TymPUtQgu856Zt8U+c6wSn/M41H0MTFh1+1UPJx2I8oGbylGucKhhXoVJTs3q52DLE70r+ryD96zYOnQcfzIw2EMb6Dq+jHgWel+Wh9vi41eO83Nf8v57bxejCNu+9gbTicu+R3/mu3X2lnMCSUYQzhPl+VgimiKVwZ1NQvXCm7KMwoDlrJA/ZHFXz5qwYFMf5nMvkj95fTUR5+MnkcXmRKDQwW/eJNehXh3btsmRuqLPepflWukrfG6RlAbkSK4rXq+YD5zc2c864FEs4cHJ1qp1UZ6ciJEGOw8MMZPm0NbfYlPY8nEHoMBDt5LzsW/XcuPTWf5u7RgmXu7fPgxNhgD/XLmBQeGkkXcgoO85jhIEWMWC/HA7GHp+Nz7tXnYRBCLYcWv4z+Gt+CDG6ord4eFH154hPhRfwYzoJ4XLdQTFAN04TcaEgMptlr/V90NfNp07zb+iNKt8Ke99c3cXJKgPXWkQ0uMXgJPsm86YCwY8Af8V4f3/PCCdEp1f9A2YqtbusP/snpblBCXxUqJPjTYfA8JTDY3B33+OuL5fSNq31Sb9ttJH013QLdj1keWMFPZ1U4J5qg8dmhLuwDxdkqktgSv7tLmmDPYhg3eMzzfLLTRcI0vAJc+l9P8/AJ5s7cFYo8MLSO+Yr2wU5RolSEU6ES+E94Y9M/LLoA2bG1OLVwKrqaS+ibCizX8ecYaFjZho/CfuWuKooxwsTxNpOhPeexWPc9bDbofdSpmTlspu70fEMeHa9tgEK9B+nz0UEhmv+eToUleuRMDFZDf5TwElz0R+5PnDVT/JU7T9e6W7NIkqohvNKsn0EB9yjaeGuAno+XrgSCPnsbBd+RftM7Jnc8nXR08iWfiZ1zjudO1l/j6jne631sDA8SEfecG5fL2ZVXb5kHxutK8HSmiKbeiNpmqh2YtFbsyssCfNbD6Cmrbl9if47VyrPVvIvTKFS5sw1QXNDG/WGzHjDhouqOnu/ll5Udduz3mEYW2JSVGGFzWSNKlRJLyQMWn5LrW/voPRzrfw1+tfAaewm/faGn/+PA17Nn6EX4+eOa5z59ZXweOLggervhLwaeb8fhc2Xh07WfD0qw7n/MvOUWY9TkK+EKGmpQGiQmy7hQCmEmnAf3RksXpSgNPGvAZGxqKD/0y58NVvSCt33/wWX3fQWgHwUWBnCkPbN25e7VcsvylBb/a/1D++QNvDT5bX1WW80bdS9Q12XFT6997d1nthz7865rZqs8K7QR4MLgtSoJvNOJU9Zhh1edTervqD6V0qqRdcdgVSf3AlY9heO5eoJ0GBd2y/ORk3N8Op6lxh7Xvc91sGoR6v+QlHQSQEiHWYA+ug7SLrgJ4is309lP4yE17Fi+TVK1TlCYYDa7TKGN0dhx8zka1QWnHsQm/15e07U4fxalvihOU0Kb4+oj9Z2KmzGm0+OZsdyb9Otf6tGpkjG1Oojc2eXp2b19r45auU9gbnQ1pJ0/1LjvV3Vjk3cWiM3MaFyZXMIei1LB9GwbKTvobuyrXYVx7u3wTW3EZ9vSWXb+IO64m5w6QXMeBuRD7roAEJa2NyzI7q5JZeCWDXbfXoZcxMas4mLUr5UOBdOv/A2hsV9YAAAB4nGNgZGBgAOK7/x/WxfPbfGXg5mAAgVv8B+7A6P8P/23hEGd7B+RyMDCBRAGW2Q6pAAAAeJxjYGRg4Cj/+4LhMwfD/4f/X3KIMwBFUIA+ALnKB3h4nONgAIIUBgaWjQwMHAyoGFXs/1v88gjMfhFVDsZmXQg25yGYzYoQR6j9/xKrO9IQ6pDlAYv1EB0AAAAAAAAAAAwALABiAJwBGAF2AZ4B6gKuAzwDpgQ4BNIFQgWIBgIGeAa8B0oHmgfYCBQIYAjcCRgJqgnWCmwLCgtCC3ALpAvADFAM1A0ADYQN2g4KDmoOkg7qDx4PWg+uEAx4nGNgZGBg0GeoZOBkAAEmIOYCQgaG/2A+AwAZ3wHIAHicfZLNSsNAFIVPbFVsRUHBlcqsRFBTf3buRNFuitBFod2l6UyNpJkwGQs+h+/g0/gM4pOIJ+lVqUIz5PLdc8+duQMDYAvvCDD79vjPOMAmsxkvYRXHwjVs4EK4Tr4SXkYT98Ir1AfCDRzhQbiJbbxwh6C+xuwSr8IB9vEhvMTeT+EadoN14Tr5UHgZO8GN8Ar1gXADvWAq3MRB8NZQ6trpyOuRGj4rYzN/EkfOJdqx0kliZwtrvOpHbZ109fgpjdyPWonzWU+7IrGZOgtP5wt3OtPu+5hiOj733ijj7ETd8kydplblzj7q2IcP3ueXrZYRPYzthHMrrms4aETwjCPmQzwzGlhk1E4Qs+a4Etad9HSYxcwsCv6GPoU+fW16EnQZx3hCWnX+9/46F9V61XkFqZxE4QwhThd23DFmVdff2xSYcqJzqp495e3KHSakW7mn5rQpWSGvao9UYuohX1HZlfPdtLjMH39IF3f6Al+4hLgAAHicbdDJUsQgFIVh/nZo53lutZ1nbZIQSJYhad7FjTurfHzLnCxl8xXce7gUZmS0Jub/NWPEAossscyYFVZZY50NNtlimx122WOfAw454pgTTjljwjkXXDLlimtuuOWOex545IlnXnjljXc+mGENP+Pvr8+5bVrZ+d6sqmXdyKi6c5kMOndxLpPOyyrIaGWrus9KWeS9waq/Cqm3abVvkuZGr3ys9Z443NfaojeVyqUq/plsEWTwsun7U1Y00nWyU3/unez6uanI1FdUyjs76NXv9A/JDfky6X5vNdfnqvu5ciHkUv+RQqyN+QVIKGtIAA==
總結
- 字體反爬的關鍵在于字體文件轉換成xml文件，雖然編碼變了但是中間的規律就是我們破解的依據
- 如果字體和編碼都會發生變化的字體反爬，那么這種方式就不適合，只能使用OCR來識別文字與編碼
