vc.py

from bs4 import BeautifulSoup
import httplib2
import unicodedata
import utils
import io, json, os

portfolio_dict = {
    "a16z": "http://a16z.com/portfolio/venture-growth/",
    "khosla": "http://www.khoslaventures.com/portfolio",
    #"accel": "http://www.accel.com/companies/",
    "sequoia": "https://www.sequoiacap.com/companies/",
    #"first round": "http://firstround.com/companies",
    "kpcb": "http://www.kpcb.com/companies",
    "ff": "http://foundersfund.com/portfolio/",
    "greylock": "http://www.greylock.com/greylock-companies/",
    "gv": "http://www.gv.com/portfolio/"
}

name_dict = {
    "a16z": "Andreessen Horowitz",
    "sequoia": "Sequoia Capital",
    "khosla": "Khosla Ventures",
    "kpcb": "KPCB Partners",
    "ff": "Founders Fund",
    "greylock": "Greylock Patrners",
    "gv": "Google Ventures",
}

industries = ["advertising", "agriculture-food", "big-data", "chemical-fuels", "consumer", "education", "efficiency", "enterprise", "financial-services", "health", "materials", "power", "robotics", "space", "storage", "transportation"]


# Crawl the portfolio list of a16z
def geta16zPortfilio():
    url = portfolio_dict['a16z']
    http = httplib2.Http()
    status, response = http.request(url)
    soup = BeautifulSoup(response)
    companies = {}
    company = ""
    for li in soup.find(id="main").findAll('li'):
        t = li.get_text()
        if t[0] == "C":
            company = t[9:]
        if t[1] == "W":
            companies[company] = "http://" + t[10:]
    return companies

# Crawl the portfolio list of Khosla Venture
def getKhoslaPortfolio():
    url = portfolio_dict['khosla']
    http = httplib2.Http()
    status, response = http.request(url)
    soup = BeautifulSoup(response)
    companies = {}
    cstring = ""
    for industry in industries:
        url = portfolio_dict['khosla'] + "/" + industry
        status, response = http.request(url)
        soup = BeautifulSoup(response)
        for i in range(1, 10):
            cstring = "company-" + str(i)
            company = soup.find(id=cstring)
            if (company == None):
                break
            else:
                for li in company.findAll('a'):
                    t = li.get_text()
                    u = li.get('href')
                    companies[t] = "http://www.khoslaventures.com" + u
    return companies

# Crawl the portfolio list of Sequoia Cap
def getSequoiaPortfolio():
    url = portfolio_dict['sequoia']
    http = httplib2.Http()
    status, response = http.request(url)
    soup = BeautifulSoup(response)
    companies = {}
    name = ""
    for li in soup.find(id="allColumn").findAll('li'):
        name = li.get_text()
        companies[name] = "https://www.sequoiacap.com"+ li.find('div').get('data-url')
    return companies

# Crawl the portfolio list of KPCB ventures
def getKPCBPortfolio():
    url = portfolio_dict['kpcb']
    http = httplib2.Http()
    status, response = http.request(url)
    soup = BeautifulSoup(response)
    companies = {}
    name = ""
    for li in soup.find("div", {"class":"medium-9"}).findAll("li"):
        if li != None:
            a = li.find('a')
            if a != None:
                url = a.get('href')
                if url[0] == '/':
                    url = "http://www.kpcb.com" + url
                name = utils.extract_name_from_url(a.get('href'))
                companies[name] = url
    return companies

# crawl portfolio list of Founder Fund
def getFoundersFundPortfolio():
    url = portfolio_dict['ff']
    # mock a browser
    user_agent = 'Mozilla/4.0 (compatible; MSIE 5.5; Windows NT)'
    headers = { 'User-Agent' : user_agent }

    http = httplib2.Http()
    status, response = http.request(url, 'GET', None, headers)
    soup = BeautifulSoup(response)
    companies = {}
    name = ""
    for a in soup.find("ul", {"class":"portfolio"}).findAll('a'):
        url = str(a.get('href'))
        url_l = url.split('/')
        name = url_l[len(url_l)-2]

        formatted_name = ""
        formatted_name += name[0].upper()

        for c in name[1:]:
            formatted_name += c
        companies[formatted_name] = url
    return companies

def getGreylockPortfolio():
    url = portfolio_dict['greylock']
    # mock browser header
    user_agent = 'Mozilla/4.0 (compatible; MSIE 5.5; Windows NT)'
    headers = { 'User-Agent' : user_agent }

    http = httplib2.Http()
    status, response = http.request(url, 'GET', None, headers)

    soup = BeautifulSoup(response)
    companies = {}
    name = ""
    for i in soup.find("div", {"id":"filterset"}).findAll('a'):
        link = url + str(i.get('href'))
        name = i.find('img').get('title')
        companies[name] = link
    return companies

def getGoogleVenturePostfolio():
    url = portfolio_dict['gv']
    # mock browser header
    user_agent = 'Mozilla/4.0 (compatible; MSIE 5.5; Windows NT)'
    headers = { 'User-Agent' : user_agent }

    http = httplib2.Http()
    status, response = http.request(url, 'GET', None, headers)

    soup = BeautifulSoup(response)
    companies = {}
    name = ""
    for div in soup.find_all("div", {"class":"grid-card-4"}):
        a = div.find('a')
        name = a.get('data-name')
        url = a.get('href')
        companies[name] = url
    return companies

def getInvestors(company):
    investors = {}
    with io.open("companies.txt") as f:
        for line in f:
            c = json.loads(line)
            if c["name"] == company:
                investors = c["investors"]
    return investors

# Crawl company profiles and write into companies.txt
def writeCompany(company):
    url = "http://angel.co/" + company
    #url = "https://www.crunchbase.com/organization/facebook/investors?page=2"
    # mock browser header
    user_agent = 'Mozilla/4.0 (compatible; MSIE 5.5; Windows NT)'
    headers = { 'User-Agent' : user_agent }

    http = httplib2.Http()
    status, response = http.request(url, 'GET', None, headers)

    soup = BeautifulSoup(response)

    investors = {}
    for a in soup.find_all("a", {"class":"startup-link"}):
        name = a.get_text()
        if name != "":
            investors[name] = a.get('href')
    c = {"name":company, "investors":investors}
    with io.open('companies.txt', 'a', encoding='utf-8') as f:
        f.write(unicode(json.dumps(c, ensure_ascii=False)) + u'\n')
        #json.dump(c, f)
        #f.write(os.linesep)

def writeCompanies(companies):
    for company in companies:
        writeCompany(company)

def getPortfolio(vc):
    #vc = vc.lower()
    if vc == "a16z":
        return geta16zPortfilio()
    elif vc == "khosla":
        return getKhoslaPortfolio()
    elif vc == "sequoia":
        return getSequoiaPortfolio()
    elif vc == "ff":
        return getFoundersFundPortfolio()
    elif vc == "kpcb":
        return getKPCBPortfolio()
    elif vc == "greylock":
        return getGreylockPortfolio()
    elif vc == "gv":
        return getGoogleVenturePostfolio()

def getVCName(acronym):
    return name_dict[acronym]

if __name__ == "__main__":
	#print geta16zPortfilio()
    #print getKhoslaPortfolio()
    #print getSequoiaPortfolio()
    #print getKPCBPortfolio()
    #print getFFPortfolio()
    #print getGreylockPortfolio()
    #print getGoogleVenturePostfolio()
    #writeInvestors("slack")

    # crawl and write all related companies into companies.txt
    vcs = ["a16z", "khosla", "sequoia", "ff", "kpcb", "greylock", "gv"]
    companies = {}
    companies_2 = []
    for vc in vcs:
        p = getPortfolio(vc)
        for key in p.keys():
            if key not in companies:
                companies_2.append(key)
                companies[key] = True
    for company in companies_2:
        name = company.replace(u'\xa0', u"-")
        name = name.replace(u'\u2019', u"-")
        name = str(name).lower()
        writeCompany(name)
        print name + " written"

    #print getInvestors("slack")