Differences between revisions 101 and 127 (spanning 26 versions)
⇤ ← Revision 101 as of 2007-02-16 04:14:05 →
Size: 2711
Editor: anonymous
Comment:
← Revision 127 as of 2008-11-14 13:41:59 → ⇥
Size: 3798
Editor: anonymous
Comment: converted to 1.6 markup
Deletions are marked like this. Additions are marked like this.
Line 1: Line 1:
[[TableOfContents]] <<TableOfContents>>

The SAGE Graph Theory Project
Line 4: Line 6:
We are implementing graph objects and algorithms in ["SAGE"]. The main people working on this project are Emily Kirkman and Robert Miller.  . We are implementing graph objects and algorithms in [[SAGE]]. The main people working on this project are Emily Kirkman and Robert Miller.
Line 7: Line 9:
 * NetworkX base classes have been interfaced, and all functions are soon to follow (SAGE Days 3 project).
 * Graph plotting has been implemented:
attachment:9cube.png		  * NetworkX base classes have been interfaced, and all functions are soon to follow.
 * Jason Grout's (Brigham Young) SQL lite graph database is almost finished, and when it is done, it will be interfaced by Emily.
 * Jim Morrow (UW) has expressed interest in using SAGE for his [[http://www.math.washington.edu/~morrow/reu07/reu.html|summer REU]] on graphs. Robert is working on implementing electrical networks, which includes implementing edge weights and graphs with boundary.
 * Chris Godsil (Waterloo) has expressed interest in helping design a more general discrete math package in SAGE.
 * Brendan !McKay's isomorphism algorithm has been replicated for the open source community (it is called "nice", as opposed to "nauty").
 * Graph plotting has been implemented in two dimensions:
Line 11: Line 16:
 * We are currently seeking possible additions to our [http://sage.math.washington.edu:9001/graph_survey survey] of existing graph theory software.
 * The initial [http://sage.math.washington.edu:9001/graph_benchmark benchmarking] has proven that ["NetworkX"] is the solution for SAGE.
 * On Friday, October 20th, Robert Miller gave a [http://sage.math.washington.edu/home/rlmill/talk_2001-10-20/2006-10-20SAGE.pdf talk] about the state of affairs for existing software which shared a few benchmarks and discussed some implementation ideas.
 * Emily Kirkman (and soon Jason Grout (Brigham Young)) is working on a database of well-known graphs.
 * Robert Miller has wrapped the basic functionality of NetworkX into SAGE's Graph class, and implemented plotting of graphs.
 * Jim Morrow has expressed interest in using SAGE for his summer REU on graphs. Robert Miller is currently working on implementing some of the algorithms of this group.
 * Chris Godsil (Waterloo) has expressed interest in helping design a more general discrete math package in SAGE, including perhaps an open source version of nauty.
== Wiki Pages ==
=== Survey of Existing Software ===
 * [http://sage.math.washington.edu:9001/graph_survey Link]
 * We have attempted to make a complete list of existing graph theory software. We posted functionality lists and some algorithm/construction summaries. We are very interested in feedback!
=== Benchmarks ===
 * [http://sage.math.washington.edu:9001/graph_benchmark Link]
 * Our initial tests are designed to compare the constructions and very basic functionality found in our survey of existing software. At this stage in the game, we are testing to find the best way to represent graph objects in SAGE.
 * We will post results on the wiki as we get them. And as always, we love feedback!
=== Plotting ===
 * [http://sage.math.washington.edu:9001/graph_plotting Link]
 * So far: NetworkX graphics primitive
=== Database ===
 * [http://sage.math.washington.edu:9001/graph_database Link]
 * So far: Basic graph structures with intuitive graphics
 * Plan: Extensive educational docstrings and many, many more graph constructors
=== Survey of Existing Database Software ===
 * [http://sage.math.washington.edu:9001/graph_db_survey Link]
 * I've found some resources, but please recommend...		 {{{
C = graphs.CubeGraph(9)
color = {}
for v in C.vertices():
    b = v.count('1')/9
    g = 1 - b
    if color.has_key((0, g, b)): color[(0, g, b)].append(v)
    else: color[(0, g, b)] = [v]
C.show(vertex_labels=False, node_size=60, graph_border=True, color_dict=color, figsize=[9,8])
}}}

{{attachment:9_cube.png}}

== Talks ==
 * On Oct. 20, 2006, Robert gave a [[http://sage.math.washington.edu/home/rlmill/talk_2001-10-20/2006-10-20SAGE.pdf|talk]] about the state of affairs for existing software which shared a few benchmarks and discussed some implementation ideas.
 * On Feb. 15, 2007, Robert and Emily gave an Informal Introduction to the Graph Theory Package. This talk was essentially a review of newly available features.
 * On Feb. 17, 2007, Robert gave a [[http://www.robertlmiller.com/talks/graph_intro.pdf|talk]] at Sage Days 3.
 * On Mar. 2, 2007, Robert gave a talk entitled "Opening Graph Automorphisms".
 * On Mar. 29, 2007, he followed up with a [[http://www.robertlmiller.com/talks/nauty.pdf|part two]] of this talk, after implementing "nice".
 * On Apr. 21, 2007, Emily presented a [[http://sage.math.washington.edu/home/ekirkman/talks/ECCAD.ppt|poster (ppt)]] introducing the SAGE graph theory package at East Coast Computer Algebra Day.
 * On Apr. 27, 2007, Emily gave a [[http://sage.math.washington.edu/home/ekirkman/talks/2007_04_27.html|talk]] introducing the (almost complete) sqlite database.

== Other wiki pages ==
 * Surveys
  . [[http://wiki.sagemath.org/graph_survey|Existing Software]]: We are always seeking additions and comments.
  . [[http://wiki.sagemath.org/graph_db_survey|Existing Databases]]: Includes comments by Jason Grout.

 * [[http://wiki.sagemath.org/graph_benchmark|Benchmarks]]
  . Initial Benchmarks indicated that NetworkX was the appropriate package for inclusion in SAGE. They were designed to compare the constructions and very basic functionality found in our survey of existing software.

 * [[http://wiki.sagemath.org/graph_plotting|Plotting]]
  * 2D plotting is implemented using matplotlib.
  * 3D plotting via Tachyon is coming soon.
  * 3D interactive viewing is on its way, via Java applets (pending some serious implementation issues).

 * [[http://wiki.sagemath.org/graph_generators|Graph Generators]]
  * So far: Basic graph structures with intuitive graphics.
  * Extensive educational docstrings and many graph constructors.
  * Native positionings allow standard layouts and quicker plotting

 * [[http://wiki.sagemath.org/graph_database|Database]]
  . Emily has provided a provisional database of graphs on up to seven vertices:

{{{
graphs_query.show_graphs(with_properties=True, nodes=7, diameter=5)
}}}
{{attachment:database.png}}

Contents

  1. Introduction
  2. Current Status
  3. Talks
  4. Other wiki pages

The SAGE Graph Theory Project

Introduction

  • We are implementing graph objects and algorithms in SAGE. The main people working on this project are Emily Kirkman and Robert Miller.

Current Status

  • NetworkX base classes have been interfaced, and all functions are soon to follow.
  • Jason Grout's (Brigham Young) SQL lite graph database is almost finished, and when it is done, it will be interfaced by Emily.

  • Jim Morrow (UW) has expressed interest in using SAGE for his summer REU on graphs. Robert is working on implementing electrical networks, which includes implementing edge weights and graphs with boundary.

  • Chris Godsil (Waterloo) has expressed interest in helping design a more general discrete math package in SAGE.

  • Brendan McKay's isomorphism algorithm has been replicated for the open source community (it is called "nice", as opposed to "nauty").

  • Graph plotting has been implemented in two dimensions: 

C = graphs.CubeGraph(9)
color = {}
for v in C.vertices():
    b = v.count('1')/9
    g = 1 - b
    if color.has_key((0, g, b)): color[(0, g, b)].append(v)
    else: color[(0, g, b)] = [v]
C.show(vertex_labels=False, node_size=60, graph_border=True, color_dict=color, figsize=[9,8])

9_cube.png

Talks

  • On Oct. 20, 2006, Robert gave a talk about the state of affairs for existing software which shared a few benchmarks and discussed some implementation ideas.

  • On Feb. 15, 2007, Robert and Emily gave an Informal Introduction to the Graph Theory Package. This talk was essentially a review of newly available features.

  • On Feb. 17, 2007, Robert gave a talk at Sage Days 3.

  • On Mar. 2, 2007, Robert gave a talk entitled "Opening Graph Automorphisms".

  • On Mar. 29, 2007, he followed up with a part two of this talk, after implementing "nice".

  • On Apr. 21, 2007, Emily presented a poster (ppt) introducing the SAGE graph theory package at East Coast Computer Algebra Day.

  • On Apr. 27, 2007, Emily gave a talk introducing the (almost complete) sqlite database.

Other wiki pages

  • Surveys

  • Benchmarks

    • Initial Benchmarks indicated that NetworkX was the appropriate package for inclusion in SAGE. They were designed to compare the constructions and very basic functionality found in our survey of existing software.

  • Plotting

    • 2D plotting is implemented using matplotlib.
    • 3D plotting via Tachyon is coming soon.
    • 3D interactive viewing is on its way, via Java applets (pending some serious implementation issues).

  • Graph Generators

    • So far: Basic graph structures with intuitive graphics.
    • Extensive educational docstrings and many graph constructors.
    • Native positionings allow standard layouts and quicker plotting

  • Database

    • Emily has provided a provisional database of graphs on up to seven vertices: 

graphs_query.show_graphs(with_properties=True, nodes=7, diameter=5)

database.png

graph (last edited 2009-11-29 06:48:46 by newacct)