Python accurately determines file type

Determining the file type is a very common requirement in development. How can we accurately determine the file type? The first thing everyone thinks of is the suffix of the file, but unfortunately this method is very unreliable, because the suffix of the file can be changed at will, and everyone knows that suffix does not have this concept under the Linux system, so It is impossible to accurately determine the type of a file just by judging the suffix. There is a second method to determine the file header. Each file will identify the file type in the file header. Let's take a look at how to use python to determine the file type.

How python determines the file type through the file header:

#! /usr/bin/python
# pythontab提醒您注意中文编码问题，指定编码为utf-8
# -*- coding: utf-8 -*- 
import struct
# 支持文件类型 
# 用16进制字符串的目的是可以知道文件头是多少字节 
# 各种文件头的长度不一样，少则2字符，长则8字符 
def typeList(): 
  return { 
    "FFD8FF": "JPEG", 
    "89504E47": "PNG"} 
   
# 字节码转16进制字符串 
def bytes2hex(bytes): 
  num = len(bytes) 
  hexstr = u"" 
  for i in range(num): 
    t = u"%x" % bytes[i] 
    if len(t) % 2: 
      hexstr += u"0"
    hexstr += t 
  return hexstr.upper() 
   
# 获取文件类型 
def filetype(filename): 
  binfile = open(filename, 'rb') # 必需二制字读取 
  tl = typeList() 
  ftype = 'unknown'
  for hcode in tl.keys(): 
    numOfBytes = len(hcode) / 2 # 需要读多少字节 
    binfile.seek(0) # 每次读取都要回到文件头，不然会一直往后读取 
    hbytes = struct.unpack_from("B"*numOfBytes, binfile.read(numOfBytes)) # 一个 "B"表示一个字节 
    f_hcode = bytes2hex(hbytes) 
    if f_hcode == hcode: 
      ftype = tl[hcode] 
      break
  binfile.close() 
  return ftype 
   
if __name__ == '__main__': 
  print filetype('./test.jpg')

File headers of common file formats

文件格式 文件头(十六进制)
JPEG (jpg) FFD8FF
PNG (png) 89504E47
GIF (gif) 47494638
TIFF (tif) 49492A00
Windows Bitmap (bmp) 424D
CAD (dwg) 41433130
Adobe Photoshop (psd) 38425053
Rich Text Format (rtf) 7B5C727466
XML (xml) 3C3F786D6C
HTML (html) 68746D6C3E
Email [thorough only] (eml) 44656C69766572792D646174653A
Outlook Express (dbx) CFAD12FEC5FD746F
Outlook (pst) 2142444E
MS Word/Excel (xls.or.doc) D0CF11E0
MS Access (mdb) 5374616E64617264204A