Added optional keyword parameter to gplot for plotting Chain Graphs #110
Conversation
Codecov Report
@@ Coverage Diff @@
## master #110 +/- ##
==========================================
+ Coverage 32.02% 37.38% +5.35%
==========================================
Files 9 9
Lines 509 527 +18
==========================================
+ Hits 163 197 +34
+ Misses 346 330 -16
Continue to review full report at Codecov.
|
|
Hi, Sorry for the huge delay here. I like this feature, but I am not sure if |
| @@ -35,9 +60,17 @@ function graphline(g::AbstractGraph{T}, locs_x, locs_y, nodesize::Real, arrowlen | |||
| endx = locs_x[j] + nodesize*cos(θ+π) | |||
| endy = locs_y[j] + nodesize*sin(θ+π) | |||
| lines[e_idx] = [(startx, starty), (endx, endy)] | |||
| arr1, arr2 = arrowcoords(θ, endx, endy, arrowlength, angleoffset) | |||
| arrows[e_idx] = [arr1, (endx, endy), arr2] | |||
| if chainGraph | |||
There was a problem hiding this comment.
You seem to have used tabs here. For consistency with the rest of the project we should spaces instead.
| end | ||
|
|
||
| function filterUndef(array) | ||
| result = [] |
There was a problem hiding this comment.
| result = [] | |
| result = eltype(array)[] |
| return has_edge(g,dst(e),src(e)) | ||
| end | ||
|
|
||
| function filterUndef(array) |
There was a problem hiding this comment.
| function filterUndef(array) | |
| function filterAssigned(array) |
or filterDefined.
Reason is, that when one uses the filter function then predicate specifies what should stay in the array and not what should be filtered out.
| if result == [] | ||
| result = typeof(array[i])[] | ||
| end |
There was a problem hiding this comment.
| if result == [] | |
| result = typeof(array[i])[] | |
| end |
We can remove this as we can already initialize the list correctly above.
| if chainGraph | ||
| if !hasReverseEdge(g,e) | ||
| arr1, arr2 = arrowcoords(θ, endx, endy, arrowlength, angleoffset) | ||
| arrows[e_idx] = [arr1, (endx, endy), arr2] | ||
| end | ||
| else | ||
| arr1, arr2 = arrowcoords(θ, endx, endy, arrowlength, angleoffset) | ||
| arrows[e_idx] = [arr1, (endx, endy), arr2] | ||
| end |
There was a problem hiding this comment.
| if chainGraph | |
| if !hasReverseEdge(g,e) | |
| arr1, arr2 = arrowcoords(θ, endx, endy, arrowlength, angleoffset) | |
| arrows[e_idx] = [arr1, (endx, endy), arr2] | |
| end | |
| else | |
| arr1, arr2 = arrowcoords(θ, endx, endy, arrowlength, angleoffset) | |
| arrows[e_idx] = [arr1, (endx, endy), arr2] | |
| end | |
| if !(chainGraph && !hasReverseEdge(g, e)) | |
| arr1, arr2 = arrowcoords(θ, endx, endy, arrowlength, angleoffset) | |
| arrows[e_idx] = [arr1, (endx, endy), arr2] | |
| end |
| @@ -1,7 +1,24 @@ | |||
| function hasReverseEdge(g, e) | |||
| return has_edge(g,dst(e),src(e)) | |||
There was a problem hiding this comment.
| return has_edge(g,dst(e),src(e)) | |
| return has_edge(g, dst(e), src(e)) |
Hi all,
I added support for "mixed model" graphs or so called "chain graphs" with an optional keyword parameter called "chainGraph". It (selectively) removes the double arrows for bi-directional edges for SimpleDiGraphs.
Visual Examples below:
gives
gives
I also added a (visual) test set for that case. I haven't implemented it for the
linestyle="curve"option though. Should an error be thrown in that case (until it's implemented)?I'm also wondering about the placement of the helper functions
hasReverseEdgeandfilterUndefcurrently they reside in lines.jl where they are used.kr,
Dieter