search

Macaulay2 / Workshop-2020-Warwick

6 stars 9 forks source link

Future: Properties of MixedGraphs #158

Open olgakuznetsova opened 4 years ago

olgakuznetsova commented 4 years ago
  1. This is already in Graphs. Should it be fixed?

If you run descendants of a vertex, then the output includes the vertex itself. But if you run children then the vertex is not included https://faculty.math.illinois.edu/Macaulay2/doc/Macaulay2-1.15/share/doc/Macaulay2/Graphs/html/_descendants.html https://faculty.math.illinois.edu/Macaulay2/doc/Macaulay2-1.15/share/doc/Macaulay2/Graphs/html/_children.html

  1. We do not implement most of the properties of Graphs for MixedGraphs.

    • edges
    • findPaths
    • distance
    • distanceMatrix
    • diameter
    • Display Methods ("displayGraph", "showTikZ", "writeDotFile") and most others but these appear most relevant

  2. How should we treat Bigraphs? Right now they just inherit neighbors and non neighbors functions for Graphs.

  3. Sadeghi and Lauritzen Section 3.1 have such concepts:

    • an arrowhead at j or pointing to j
    • tripath
    • collider/non-collider
    • end-point identical paths

  4. In GraphicalModels we introduce neighbours of a vertex in the underlying graph of a mixed graph. Should this be implemented in StatGraphs instead? https://github.com/Macaulay2/Workshop-2020-Warwick/blob/ce0c5d8326687b6e989465b68f1ad91d03433128/AlgebraicStatistics/MLE/GraphicalModels.m2#L537-L551

  5. FYI: I changed the code for descendants, etc, so now it first collates the vertices and then takes the function. It doesn't affect descendants, parents, children, foreFathers and neighbours but IMO makes more sense for non-descendants and non-neighbours.

olgakuznetsova commented 4 years ago

@lukeamendola @harshitmotwani2015 what are your thoughts on this? I suspect some things in (4) have already been implemented in the TrekSeparation package. Should we move it here?