Color particles by an attribute - #279
Open
WeiqunZhang wants to merge 18 commits into
Open
WeiqunZhang wants to merge 18 commits into
WeiqunZhang wants to merge 18 commits into
Conversation
WeiqunZhang
added this pull request to stack #280
September 29, 2026 22:41
WeiqunZhang
force-pushed
the
particle-color-by-attribute
branch
from
September 29, 2026 23:50
578f76d to
fb1ed4b
Compare
WeiqunZhang
force-pushed
the
particle-color-by-attribute
branch
from
September 29, 2026 23:59
fb1ed4b to
7dcfc45
Compare
Addresses the color item of #275. The Particles dialog gains Color by (any component name a species has), a colormap, a log scale and a fixed range; species without the component keep their color, and a particle color bar appears in the Color Scale dock. The reader keeps component names and reads one attribute per point; sessions and frame specs carry it, and protocol 1.9 sends names and values (a 1.8 peer's attribute is refused by version). A pure-SoA plotfile written by AMReX (with amrex#5698's int-init fix) tests the reader against the real writer.
WeiqunZhang
force-pushed
the
particle-color-by-attribute
branch
from
September 30, 2026 00:01
7dcfc45 to
b345103
Compare
Particles moved under Overlays when View was split, where P is free, so Overlays > Particles... is now Alt+O, P instead of T.
While particles are colored, image and animation exports of the primary's panels carry the particle scale right of the field's. An export layout can hold a second color bar; a frozen movie layout without room for it leaves it out.
The particle scale normalized with its own arithmetic, which overflowed for a range like [-1e308, 1e308]: 0 took the lowest color and the maximum turned NaN before its cast to a slot. Particles now take their slot through ValueMapping, as slices do.
Each palette-slot batch copied the whole species overlay, points and all, then replaced the points: a million particles over every slot meant 253 full copies per view. Batches now take only the style.
A frame's samples are drawn view by view, and only the all-views redraw refreshed the bar, so a frame whose masses ran 10 to 80 kept the previous frame's 1 to 8 legend. Every view redraw now refreshes it. The smoke test's second frame has scaled masses, and its accessor reads the widget's own range.
The range spin boxes were set to 17 decimals, overriding the scientific spin box's own precision, so any bound below about 1e-17, such as a mass in kilograms, rounded to 0 and Apply refused the range.
With Color by back on species colors, a bad fixed range left behind still blocked Apply, though its controls were disabled and could not be corrected. The range is now checked only when an attribute is chosen.
If this frame's species lacked the chosen attribute, Color by fell back to species colors, and an Apply made only to change alpha cleared the coloring for every frame. The attribute now stays listed, marked as not in this frame.
Opening another dataset or a sequence cleared the particle samples but left their color bar up until the next redraw, or for good if the open failed. Both teardowns now refresh it.
The on-screen particle bar now says " (log)" on a log scale, as the exported one and the field's bar do. The batching comment claimed values past the range keep the species color; they take its end colors.
A movie's layout freezes on its first frame, which reserved the particle color bar only when that frame had usable values, so a movie starting on NaN attributes never showed the bar. A movie now keeps the room whenever an attribute is chosen and paints the bar on frames with a range. The materializer gains --mass-scale, NaN allowed.
A movie kept room for the particle color bar whenever an attribute was chosen, so with no species selected, or none that has the component, every frame carried an empty column. It now also needs a selected species with that component.
When a movie's first frame had no usable particle values, its frozen layout kept the label format worked out from the 1-10 stand-in range, and every later frame's bar was labelled with it. The layout now keeps only the room; each frame labels the bar from its own values.
Each view redraw filled a per-point value vector for every sample even when no attribute was chosen, only to leave it unread. It is now filled only for a colored sample. Also correct the fixture's mass comment, which predates --mass-scale.
A sample counted as colored if it carried any attribute, so switching Color by while a sequence frame loaded drew the old attribute's values under the new one's name, for good if the reload failed. A sample is now colored only when its attribute is what the chosen name resolves to for its species.
The settings restore passed the saved format to the field's color bar but not the particle bar, created earlier with the default, so after a restart the two labelled their ticks differently.
The fixed bounds were filled for the attribute the dialog opened on and kept when another was chosen, so switching from mass over [0, 1] to an attribute near 1e6 colored every particle with the top color. Choosing another attribute now unticks Fixed range.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Addresses the color item of #275. The Particles dialog gains Color by (any component name a species has), a colormap, a log scale and a fixed range; species without the component keep their color, and a particle color bar appears in the Color Scale dock.
The reader keeps component names and reads one attribute per point; sessions and frame specs carry it, and protocol 1.9 sends names and values (a 1.8 peer's attribute is refused by version). A pure-SoA plotfile written by AMReX (with amrex#5698's int-init fix) tests the reader against the real writer.