Question

I wanted to access a TFS server, and download the dynamically produced “Scrum Burndown chart” on a regular basis, in order to re-use it in a seperate report (ie a pdf)
Yes, there is a lot of reporting going around …

How do I find the correct URL in the server to download, and how do I get around the permissions problems?

Answer 1

This is a question I have answered myself.

Summary of process

• use a HTTP proxy sniffer to determine the format of the URL – see below
• using py-ntlm, connect to TFS server, using a known format of URL to get the “full” report page,
• parse the report page to get the image url, down load that again through py-ntlm, and store.

There is no direct URL ie I cannot ask for reportserver/myteam/burndownchart/20110815 (nice and restful) but there is a param style request for the
wrapper “report” to the image I want. Sadly there is no easy way to find it just browsing in VS2010 extension – I used Charles debugger proxy to see what VS2010 is sending.

The url you will need to directly visit the report for your team is similar to
http://tfs.example.com/reportserver?/TfsReports/ExampleCompany/ExampleProject/Sprint+Burndown+Chart&rc:toolbar=false&pReleaseWorkStreamPath=\Release01\&pSprintTeam=\Release02\Sprint01\Team01

There is some Session based obfuscation going on with the actual image URL, meaning I must call then parse the report page, to find the URL of the image, and then download it. On top of that I needed to get Python to work with Windows Authentication, and generally fiddled around a great deal

The below process is sound but the code is brittle and I just recommend it as a starting point

Identify the URL of the report

Install py-ntlm (http://code.google.com/p/python-ntlm/)

(a) You need to create a hash of user pass and requested url each time (this is invisible by browser usually)
(b) THen pass that as a urllib2 handler to urllib2
(c) process the url as usual through urllib2

import urllib2
from ntlm import HTTPNtlmAuthHandler
#
passman = urllib2.HTTPPasswordMgrWithDefaultRealm()
passman.add_password(None, url, user, password) ## (a) 
auth_NTLM = HTTPNtlmAuthHandler.HTTPNtlmAuthHandler(passman)  #(b)
opener = urllib2.build_opener(auth_NTLM)

urllib2.install_opener(opener)  # (c)
response = urllib2.urlopen(url_moi)
html = response.read()

SO that gets me the standard report web page, with image in it. I now need to extract that image url, for which I used a regex to get
all IMG tags and took the second one in the list. Professional and not at all brittle 🙂

We then rebuild the nltm hash (I have downloaded the HTML report page, now I want the JPG image in it),
call the URL above and write the output to local disk.

I hope that helps.

YMMV