this is dead, see rozap/corpcrawl

This is a python scraper for the Securities and Exchange Commission EDGAR database. It looks at the Form 10k filings that publicly held corporations are required to file with the SEC. It then attempts to extract the subsidiary relationships from the 10k exhibit 21.1.

Corpcrawl is a storage agnostic scraper, so you'll need to implement your own storage scheme.

You can get the package from PyPi through Pip.

pip install corpcrawl

From a python console you can try

import corpcrawl

If it works, then you can get started

First import the required pieces

from corpcrawl.crawler import CorpCrawl

from corpcrawl.backend import Backend

def main()

my_backend = MyBackend() crawler = CorpCrawl(cache_path = '/an/absolute/path/to/some/dir', backend = my_backend) c.crawl(years = [2011, 2012], quarters = [1, 2, 3, 4])

class MyBackend(Backend):

def get_company(self, name):

pass

def add_company(self, comp):

print "Adding %s" % str(comp)

This is about as simple of a backend as can be. This code will crawl the filings and print out the name of each company and subsidiary that it finds for all quarters of years 2011 and 2012. Obviously you will want to hook it up to a database or something.

The SEC data is very unstructured. As such, there are a lot of errors. This is version 0.0.1, so instead of implementing your own methods to massage the data that it returns, it would be better if you could contribute so make the core parser better.