Show a legend for barplot layers using length scaling #354
Assignees
Labels
Comments
fedarko
added a commit
to fedarko/empress
that referenced
this issue
Aug 26, 2020
(still gotta test and document the new code, ofc)
kwcantrell
pushed a commit
that referenced
this issue
Aug 31, 2020
…gth-scaling (#357) * ENH: draw circular layout fm barplots! (hacky tho) * BUG/ENH: correctly set "maxX" in c.bps+use 2 rects * ENH: make s.m. barplots work in circular layout how is this actually not that bad what * DOC: upd8 comment * DOC: update README re #297 * DOC: document polar -> Cartesian stuff * DOC/PERF: mark some places to speed up cbarplots since this kinda uh crashes the EMP tree LMAO * PERF: reduce unnecessary work in c.barplot drawing * PERF: chop out more work in c.sm.b. drawing * MNT: abstract some more c. barplot drawing code * DOC: document new util funcs for circ barplots * DOC: clean up inner/outer radius terminology * Typo fixes * MNT: max X -> max displacement in bp drawing code a more elegant way of representing this for circ barplots * DOC: document precomputed y/angle stuff * STY: pret * MNT: shorten nodeAngleInfo var to just angleInfo * MNT: Catch bad-layout case in _getNodeAngleInfo() * ENH: Update "barplot unavail" msg re circ barplots I would still like to add some tests to the coordinate functions, but I think this is sufficient to close #297. * PERF: don't draw barplots by dflt (close #343) * DOC: punctuation * TST: Test _getNodeAngleInfo() * TST: add early addCircularBarCoords tst * TST: finish x-coord circ bar coord test just gotta test y-coords, with a similar switch stmt * STY: prettify test * TST: finish _addCircularBarCoords() test * TST/MNT: abstract redundant test code also add approxEqual func which I FEEL LIKE I ALREADY ADDED but w/e * TST/DOC: ref toFixedIfy et al in js layout tsts * MNT: rm check for unsupported layouts in bpdrawing * DOC: update comments in drawBarplots() start * TST: Add more context/detail to circ bar comp tst * DOC: clarify comment * PERF: don't compute halfAngleRange w rect barplots Makes things a bit more efficient. * STY: prettify test error msg * MNT: abstract rect coord-adding to sep func * TST: test _addRectangularBarCoords() * ENH: add length barplot legends; close #354 (still gotta test and document the new code, ofc) * MNT: Simplify+doc layer drawing code re lenlegends * TST: Clarify test util funcs * DOC/TST: refurbish testing utilities docs * TST: unbreak assignBarplotLengths() tests * DOC/TST: polish up and test assignBarplotLengths esp the new functionality. but also i apparently wasn't testing non-numeric and numeric combos! that's wack. so now that's tested. * TST: unbreak circ layout js tests (since nodes are now just referred to by postorder positions) * DOC: document more length legend stuff * TST: test Legend.addLengthKey() * STY: fix improperly named variable - jshint thing * DOC: fix outdated comment thanks @ElDeveloper! * MNT: Simplify maxD computation in drawBarplots() Thanks @ElDeveloper! * DOC: fix old node desc in _getNodeAngleInfo() thanks @kwcantrell! * DOC: max displacement (not just x) in fm/sm coords Forgot to update the docs. Thanks @kwcantrell for pointing this out. Also changed a small block of code to just compute thisLayerMaxD once and not like 3 times (?? why was I doing that lol) * MNT: Compute max displacement in getLayoutInfo() So that this is only done once per layout. Addresses comment from @kwcantrell. * STY: add missing semicolon * TST: restructure approxDeepEqual funcs to use math ... and not stringification :P Default epsilon is 1e-5, but for a few tests I only bothered writing things out to 4 decimal places, so for these I use 1e-4. * TST/STY: clean up utilities for testing * DOC: tidy up max displacement wording * BUG: Fix error when searching for invalid name * DOC: update getNodesWithName() docs * TST: fix approxDeepEqualMulti docs thanks @ElDeveloper * DOC: Update maxDisplacement description thanks @kwcantrell * MNT: add .gitattributes file marking vendored code
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
This would just show the min / max value of the field used for length scaling. It'd be a nice way to convey some extra information to the user, and could help with determining what an adequate min / max length for the visualization should be: for example, in the MP dataset
Confidencefield, the min value is ~0.688 (with the max being ~0.999), so it makes sense to set the minimum length to 69 if the max length is 100 (rather than using the default min length of 1).I already have a demo of this ready on a branch continuing from #340 (see screenshot below). However, it required some refactoring, so I think it'd make sense to merge #340 in first and then add this in when it's ready in a few days.
The text was updated successfully, but these errors were encountered: