Python通过商业库编号爬smiles

2020-02-21

好记性不如烂笔头

直接上代码

# -*- coding: utf-8 -*-
"""
Created on Tue Feb 11 15:14:58 2020

@author: dakki

E-mail: chendaoxing@126.com

Description: 
"""

from urllib.request import urlopen
import re
import pandas as pd

# 获得化合物系统命名
url = 'https://product.pharmablock.com/Home/Search?cas=ids'
identifiers=[]
no=[]
id_file=open('C:\\Users\\cdx07\\Desktop\\wuchao_job\\id.txt','r',encoding='utf-8')
id_line=id_file.readlines()

def Getname(id_url):
    try:
        ans1 = urlopen(id_url).read().decode('utf8')
        #name1 = re.search(r'data-name='+'\"(.+?)\"',ans1)
        name1 = re.search(r'productName='+'(.+?)\&',ans1)
        name2 = name1.group(1)
        return name2
    except:
        return 'Did not work'
    
for i in range(len(id_line)):
    no.append(id_line[i])
    id_url=re.sub(r'ids',id_line[i],url)
    identifiers.append(Getname(id_url))

# 获得化合物smiles格式
smiles=[]
def CIRconvert(ids):
    try:
        url = 'http://cactus.nci.nih.gov/chemical/structure/' + ids + '/smiles'
        ans2 = urlopen(url).read().decode('utf8')
        return ans2
    except:
        return 'Did not work'
    
for ids in identifiers :
    smiles.append(CIRconvert(ids))

# 合并成DataFrame并导出csv格式
df = {'smiles':smiles,
      'no':no,
      'names':identifiers}
result = pd.DataFrame(df)
result.to_csv('result.csv')

得到一个csv格式的文件，里边包括smiles,商业库的编号,化合物的名称。