There are numerous graph theoretic tools/packages. Each with its pros and cons. What should be the semantics/syntax of a programming language meant to solve graph theoretic problems?
You might want to look at The Graph Programming Language GP. From the linked page:
GP (for Graph Programs) is a rule-based, nondeterministic programming language for solving graph problems at a high level of abstraction, freeing programmers from handling low-level data structures. The core of GP consists of four constructs: single-step application of a set of conditional graph-transformation rules, sequential composition, branching and iteration.
Sandra Steinert devoted her PhD thesis to the topic.
There's also a Hoare logic for reasoning about the correctness of such programs.
Take a look at the Sage Project, which makes python more mathy, and includes networkx and nice visualization stuff. Their list of graph function might be wiki.sagemath.org/graph_survey or wiki.sagemath.org/graph.
Sazzad - I'm a bit frustrated with your responses. You have been given useful answers and suggestions, yet you remain unsatisfied. Either use the best-of-breed tools that are available on your system for several aspects of graph research, or write your own.
Gremlin is one. It has extensive documentation and is well developed. I believe it is in use by a group at AT&T. Also, it is cross platform (if you care about such things). There's a presentation by the author on his website, but I can't add an additional hyperlink to this post because I'm a new user.