I think this impression comes from the culture of experimentation in Ruby and more widespread willingness to adopt new tools and libs very quickly. Python definitely has a diverse web development community and more options in some areas (like templates), but it's more conservative, not necessarily evolving as rapidly and, like Python overall, resists influence from other languages.
For instance, Haml and Saas are pretty widely accepted by Ruby web developers, whereas Django templates are very popular with Python web developers precisely because because they are close to straight HTML. Ruby developers (like Ruby itself) readily adopt good ideas from other languages, including Python (rack from wsgi, rip from pip, merb slices and rails engines from django apps), while anything that resembles anything from ruby or rails, like some Pylons components, gets criticized by many python developers. It's noticeable with version control, too, as svn is still very widely used by Python developers whereas Ruby and Ruby/Py developers moved completely to git and hg over a year ago.
FWIW, there are lots of Python folks moving to git and hg as well. But I don't understand what VCS choice has to do with language culture. What's next, a critique of editor choice? :)
Python is such a large tent that it's quite inaccurate to say that "the Python culture" lacks a sense of experimentation and adoption of new tools. What you might be getting at is the fact that there are already so many existing libraries in Python for doing most things that there's not a lot of public reinvention of wheels going on.
"FWIW, there are lots of Python folks moving to git and hg as well"
It's not remotely as universal as it was for Ruby developers. When working with Python you still can't go a day without hitting an svn repo, but with Ruby you'd easily not touch one for months.
"But I don't understand what VCS choice has to do with language culture."
A lot, and I already described it. Ruby developers and the Ruby language overall are less conservative and more open to dramatic change than Python developers or the language. Each model has its benefits.
"'Python is such a large tent that it's quite inaccurate to say that "the Python culture'"
It's a large tent, but the common thread is Python and Python's principles, and they define the culture.
"there are already so many existing libraries in Python for doing most things"
But individual developers don't do most things, each works on certain things. Python has lots of libraries, but that doesn't mean it's the best choice for any specific job. As a web developer, there are a huge number of Python packages that don't matter at all in that domain. If sheer number of libraries mattered, we'd all be using perl.
Can't compare to ruby as I haven't used it much but the python world is seriously overflowing with web-related frameworks and libraries.
Django is the most popular single framework, but at least as many people prefer to roll their own based on various modules. Some of the most popular backbone kits would be werkzeug, cherrypy, webpy, pylons. Then there's literally dozens of template languages (jinja, genshi, cheetah, to name a few) and at least three mature ORMs (sqlalchemy, storm, sqlobject) to choose from.
So, while I have no idea about ruby libraries I can assure you that in python-land you get to choose between at least two mature options for pretty much any web-related task.
Ruby On Rails is a framework not a collection of third party libraries (like pypi), that said if you really want to compare RoR to something then that's http://www.djangoproject.com/
I get the impression that there are more 3rd-party libraries in Ruby, so it is easier to get web code up and running with ruby.