python script to parse sphinx objects.inv file

inventory.py

#!/usr/bin/env python
""" Process URL for intersphinx targets and emit html or text """

def validuri(string):
   return string

from sphinx.ext.intersphinx import read_inventory_v2
from posixpath import join
import pprint
import argparse
import locale
import os,sys,tempfile
import urllib2


parser = argparse.ArgumentParser(description='Process intersphinx link library')

parser.add_argument('--url' , type=validuri, help="URL to retrieve objects.inv from")
parser.add_argument('--file' , help="objects.inv format file")

group = parser.add_mutually_exclusive_group(required=False)

group.add_argument('--html', action='store_true', help="Output HTML")
group.add_argument('--terse', action='store_true', help="Output terse text list")
group.add_argument('--rst', action='store_true', help="Output ReStructuredText")
group.add_argument('--rewrite', action='store_true', help="Output short form and correct form of each link.")


args = parser.parse_args()

def start_role(role):
	if (args.terse):
		return
	elif (args.rewrite):
		return
	elif (args.rst):
		print role
	else:
		print "<dt>Role: {}</dt>\n<dd>\n<dl>\n".format(role)

def start_item(role,item):
	if (args.terse):
		return
	elif (args.rewrite):
		return
	elif (args.rst):
		print "\t:{}:{}:".format(role,item)
	elif (args.html):
		print "<dt>{}:{}</dt>\n".format(role,item)
		print "<dd>"
		print "<table>\n<tbody>"
		
def end_item(role,item):
	if (args.html):
		print "</tbody></table>"
		print "</dd>\n"

def print_link(role,item,domain,title):
	"""Return the correct link form, if no title then extended form."""

	domain = domain.lower()
	if (title == '')|(title=='-'):
		linkStr = ":{}:`{} <{}:{}>`".format(role,item,domain,item)
	else:
		linkStr = ":{}:`{}:{}`".format(role,domain,item)
	
	if (args.terse):
		print linkStr
	if (args.rewrite):
		print ":{}:`{}:{}`".format(role,domain,item), "\t{}".format(linkStr)
	elif (args.rst):
		print "\t\t:Link:\t{}".format(linkStr)
	elif (args.html):
		print "<tr><th>Link:</th><td>{}</td></tr>".format(linkStr)		

def end_role():
	if (args.html):
		print "</dl>\n"
		print "</dd>\n"
		
def print_meta(role,item,domain,version,url,title):
	if (args.terse):
		return
	elif (args.rewrite):
		return
	elif (args.rst):
		print "\t\t:Domain:\t{}".format(domain)
		print "\t\t:Version:\t{}".format(version)
		print "\t\t:URL:\t{}".format(url)
		print "\t\t:Title:\t{}".format(title)
	elif (args.html):
		print "<tr><th>Domain:</th><td>{}</td></tr>".format(domain)
		print "<tr><th>Version:</th><td>{}</td></tr>".format(version)
		print "<tr><th>URL:</th><td>{}</td></tr>".format(url)
		print "<tr><th>Title:</th><td>{}</td></tr>".format(title)
	return

def fetch_data(url,inv):

	f = open(inv, 'rb')
	line = f.readline() # burn a line
	invdata = read_inventory_v2(f, url or '', join)
	if (args.html):
		print "<dl>"
	for role in invdata:
		start_role(role)
		for item in invdata[role]:
			(domain, version, url, title) = invdata[role][item]
			#sys.stderr.write("{}:{} :: {}\n".format(role,item,invdata[role][item]))
			start_item(role,item)
			print_link(role,item,domain,title)
			print_meta(role,item,domain,version,url,title)
			end_item(role,item)
		if (args.html):
			print "</dl>\n"
	
if __name__ == "__main__":

	if (args.file):
		inv = args.file
#		if os.path.exists(inv) == False:
#			raise Exception('File "{}" does not exist'.format(inv))
	else:
		inv = False
	if (args.url):
		url = args.url			
	else:
		url = False
		
	
	# sys.stderr.write('URL({}) FN({})\n'.format(url,fn))
		
	if (inv == False) & (url == False):
		raise Exception("need to specify a file or URL")
   
	if (inv != False ) & (url != ''):
		#sys.stderr.write("Reading from: {}\n".format(inv))
		#sys.stderr.write("Using: {} as base HREF\n".format(url))
		fetch_data(url,inv)
	elif (url != False):
		# fetch URL into inv
		#sys.stderr.write("Retrieving objects.inv from {}\n".format(url))
		if (url.rfind('objects.inv')>5):
			invdata = urllib2.urlopen(url)
		else:
			invdata = urllib2.urlopen(url + '/objects.inv')
		sys.stderr.write('URL resolved to: {}\n '.format(invdata.geturl()))
		#print invdata.read()
		f = tempfile.NamedTemporaryFile()
		f.write(invdata.read())
		sys.stderr.write("objects.inv written to: {}\n".format(f.name))
		sys.stderr.write("Using: {} as base HREF\n".format(url))
		fetch_data(url,f.name)
	else:
		raise Exception("You need to specify a --URL")

Also on Windows it fails, you should change from posixos import join to from os.path import join

You can now call sphinx.ext.intersphinx directly to output cleartext from objects.inv

[Your_Sphinx_venv]/bin/python -m sphinx.ext.intersphinx.py [path or URL] > output.txt

The output uses justified text instead of tabs for better display in a terminal. To output fully tabbed csv you can tweak the print output of the inventory_main function at the end.

here is the link to my gist with the pure tab-delimited output https://gist.github.com/acsr/fb7b5cf97627f70876ceeae031cd37f1#file-sphinx-ext-intersphinx-py

In case anyone is 1) still using this and 2) has comments, please send them to me or call attention to me, for whatever reason I never get notifications on this gist. I'm @epc on Twitter and epc at epcostello dot com via email.