CSV spreadsheets can include SVG images, yet only the defs (gradients, etc) and
drawing data (paths, shapes, etc) are imported, not the header.
This means that included images have to comply with the SVG XML used by
gnuclad.
If you think the current SVG header is missing something, please file
a bug detailing the problem.
CSV spreadsheets can also include PNG images, but keep in mind that raster images are encoded in base64 and may result in a really huge output file.
All nodes with a starting date newer than endOfTime (usually the current date) will be ignored. All nodes with a stopping date newer or the same as endOfTime will have their stopping date removed (unset == "still alive").
For best results with CSV output, the recommended optimisation is 0 and treeMode conf setting is "lower inverse":
optimise = 0 treeMode = 2
This way the trees will be grouped by children and sorted by date, while your single nodes won't be mangled.
The optimisation parameter can be set to a value between 0 and 99. 0 turns off optimisation completely while 99 tries to stuff the resulting image closely together.
The first digit refers to the optimisation factor for isolated nodes, those which are not part of the tree. The higher it is, the more aggressively they will be appended to each other. The second digit sets the tree optimisation. Higher values will try to cramp the tree together more than lower ones.
The current mapping looks as follows (the exact parametrisation may be subject to change). Each higher scoring optimisation includes all the lines above it in the same group.
0 - off x1 - queue pairwise consecutive node lines; fast x3 - pull nodes to parent x6 - pull nodes towards root x8 - pull nodes towards root strongly 1x - inject single nodes among other single nodes in a greedy manner - don't pass over trees 2x - pass over trees of size <= 2 3x - pass over trees of size <= 3 4x - pass over trees of size <= 5 5x - pass over trees of size <= 10 6x - pass over trees of size <= 20 7x - pass over trees of size <= 50 8x - pass over any tree
The current optimisation is greedy and therefore not exactly stellar.
For example, this optimises well:
4 -> 2 ___ -> ___ ___ ___ ___ -> _______ _______ -> ___ ->
But a slightly different map doesn't (a human can easily reduce this to 2):
4 -> 3 ___ -> ___ ___ ___ -> _______ _______ -> _____ _____ ->
Sorting by start date (see config options) might help a bit.