baseline fix for bar plots using log scales

src/args.jl

@@ -531,6 +531,9 @@ const _suppress_warnings = Set{Symbol}([
  :relative_bbox,
  :layout_insets,
  :force_minpad,
+ :x_extrema,
+ :y_extrema,
+ :z_extrema,
 ])
 
 is_subplot_attr(k) = k in _all_subplot_args

src/pipeline.jl

@@ -151,6 +151,14 @@ function RecipesPipeline.plot_setup!(plt::Plot, plotattributes, kw_list)
 end
 
 function RecipesPipeline.process_sliced_series_attributes!(plt::Plots.Plot, kw_list)
+ # determine global extrema
+ xe = ye = ze = NaN, NaN
+ for kw in kw_list
+ xe = ignorenan_min_max(get(kw, :x, nothing), xe)
+ ye = ignorenan_min_max(get(kw, :y, nothing), ye)
+ ze = ignorenan_min_max(get(kw, :z, nothing), ze)
+ end
+
  # swap errors
  err_inds =
  findall(kw -> get(kw, :seriestype, :path) in (:xerror, :yerror, :zerror), kw_list)
@@ -163,6 +171,10 @@ function RecipesPipeline.process_sliced_series_attributes!(plt::Plots.Plot, kw_l
  end
 
  for kw in kw_list
+ kw[:x_extrema] = xe
+ kw[:y_extrema] = ye
+ kw[:z_extrema] = ze
+
  rib = get(kw, :ribbon, default(:ribbon))
  fr = get(kw, :fillrange, default(:fillrange))
  # map ribbon if it's a Function

src/recipes.jl

@@ -292,17 +292,15 @@ end
 # create vertical line segments from fill
 @recipe function f(::Type{Val{:sticks}}, x, y, z) # COV_EXCL_LINE
  n = length(x)
- fr = plotattributes[:fillrange]
- if fr === nothing
+ if (fr = plotattributes[:fillrange]) === nothing
  sp = plotattributes[:subplot]
  fr = if sp[:yaxis][:scale] === :identity
  0.0
  else
  NaNMath.min(axis_limits(sp, :y)[1], ignorenan_minimum(y))
  end
  end
- newx, newy = zeros(3n), zeros(3n)
- newz = z !== nothing ? zeros(3n) : nothing
+ newx, newy, newz = zeros(3n), zeros(3n), z !== nothing ? zeros(3n) : nothing
  for (i, (xi, yi, zi)) in enumerate(zip(x, y, z !== nothing ? z : 1:n))
  rng = (3i - 2):(3i)
  newx[rng] = [xi, xi, NaN]
@@ -368,8 +366,7 @@ end
 @recipe function f(::Type{Val{:curves}}, x, y, z; npoints = 30) # COV_EXCL_LINE
  args = z !== nothing ? (x, y, z) : (x, y)
  newx, newy = zeros(0), zeros(0)
- fr = plotattributes[:fillrange]
- newfr = fr !== nothing ? zeros(0) : nothing
+ newfr = (fr = plotattributes[:fillrange]) !== nothing ? zeros(0) : nothing
  newz = z !== nothing ? zeros(0) : nothing
 
  # for each line segment (point series with no NaNs), convert it into a bezier curve
@@ -406,14 +403,14 @@ end
 
 # create a bar plot as a filled step function
 @recipe function f(::Type{Val{:bar}}, x, y, z) # COV_EXCL_LINE
- procx, procy, xscale, yscale, baseline = _preprocess_barlike(plotattributes, x, y)
+ procx, procy, xscale, yscale, _ = _preprocess_barlike(plotattributes, x, y)
  nx, ny = length(procx), length(procy)
  axis = plotattributes[:subplot][isvertical(plotattributes) ? :xaxis : :yaxis]
- cv = [discrete_value!(plotattributes, :x, xi)[1] for xi in procx]
+ cv = map(xi -> discrete_value!(plotattributes, :x, xi)[1], procx)
  procx = if nx == ny
  cv
  elseif nx == ny + 1
- 0.5diff(cv) + cv[1:(end - 1)]
+ 0.5diff(cv) + @view(cv[1:(end - 1)])
  else
  error(
  "bar recipe: x must be same length as y (centers), or one more than y (edges).\n\t\tlength(x)=$(length(x)), length(y)=$(length(y))",
@@ -437,26 +434,21 @@ end
  fillto = 0
  end
  if yscale in _logScales && !all(_is_positive, fillto)
- fillto = map(x -> _is_positive(x) ? typeof(baseline)(x) : baseline, fillto)
+ # github.com/JuliaPlots/Plots.jl/issues/4502
+ T = float(eltype(y))
+ min_y, _ = plotattributes[:y_extrema]
+ baseline = floor_base(min_y, _logScaleBases[yscale])
+ fillto = map(x -> _is_positive(x) ? T(x) : T(baseline), fillto)
  end
 
- xseg, yseg = Segments(), Segments()
+ xseg, yseg = map(_ -> Segments(), 1:2)
  for i in 1:ny
- yi = procy[i]
- if !isnan(yi)
- center = procx[i]
- hwi = _cycle(hw, i)
- fi = _cycle(fillto, i)
- push!(
- xseg,
- center - hwi,
- center - hwi,
- center + hwi,
- center + hwi,
- center - hwi,
- )
- push!(yseg, yi, fi, fi, yi, yi)
- end
+ (yi = procy[i]) |> isnan && continue
+ center = procx[i]
+ hwi = _cycle(hw, i)
+ fi = _cycle(fillto, i)
+ push!(xseg, center - hwi, center - hwi, center + hwi, center + hwi, center - hwi)
+ push!(yseg, yi, fi, fi, yi, yi)
  end
 
  # widen limits out a bit
@@ -509,11 +501,11 @@ end
 @recipe function f(::Type{Val{:plots_heatmap}}, x, y, z) # COV_EXCL_LINE
  xe, ye = heatmap_edges(x), heatmap_edges(y)
  m, n = size(z.surf)
- x_pts, y_pts = fill(NaN, 6 * m * n), fill(NaN, 6 * m * n)
+ x_pts, y_pts = fill(NaN, 6m * n), fill(NaN, 6m * n)
  fz = zeros(m * n)
  for i in 1:m, j in 1:n # i ≡ y, j ≡ x
  k = (j - 1) * m + i
- inds = (6 * (k - 1) + 1):(6 * k - 1)
+ inds = (6(k - 1) + 1):(6k - 1)
  x_pts[inds] .= [xe[j], xe[j + 1], xe[j + 1], xe[j], xe[j]]
  y_pts[inds] .= [ye[i], ye[i], ye[i + 1], ye[i + 1], ye[i]]
  fz[k] = z.surf[i, j]
@@ -539,7 +531,7 @@ RecipesPipeline.is_surface(::Type{Val{:hexbin}}) = true
 # ---------------------------------------------------------------------------
 # Histograms
 
-_bin_centers(v::AVec) = (v[1:(end - 1)] + v[2:end]) / 2
+_bin_centers(v::AVec) = (@view(v[1:(end - 1)]) + @view(v[2:end])) / 2
 
 _is_positive(x) = (x > 0) && !(x ≈ 0)
 
@@ -549,16 +541,11 @@ _scale_adjusted_values(
  ::Type{T},
  V::AbstractVector,
  scale::Symbol,
-) where {T<:AbstractFloat} =
- if scale in _logScales
- [_positive_else_nan(T, x) for x in V]
- else
- [T(x) for x in V]
- end
+) where {T<:AbstractFloat} = scale in _logScales ? _positive_else_nan.(T, V) : T.(V)
 
 _binbarlike_baseline(min_value::T, scale::Symbol) where {T<:Real} =
  if scale in _logScales
- isnan(min_value) ? T(1e-3) : min_value / T(_logScaleBases[scale]^log10(2))
+ isnan(min_value) ? T(1e-3) : floor_base(min_value, _logScaleBases[scale])
  else
  zero(T)
  end
@@ -734,9 +721,11 @@ function _auto_binning_nbins(
 
  # The nd estimator is the key to most automatic binning methods, and is modified for twodimensional histograms to include correlation
  nd = n_samples^(1 / (2 + N))
- nd =
- N == 2 ?
- min(n_samples^(1 / (2 + N)), nd / (1 - cor(first(vs), last(vs))^2)^(3 // 8)) : nd # the >2-dimensional case does not have a nice solution to correlations
+ nd = if N == 2
+ min(n_samples^(1 / (2 + N)), nd / (1 - cor(first(vs), last(vs))^2)^(3 // 8))
+ else # the >2-dimensional case does not have a nice solution to correlations
+ nd
+ end
 
  v = vs[dim]
  mode === :auto && (mode = :fd)
@@ -1062,20 +1051,17 @@ function intersection_point(xA, yA, xB, yB, h, w)
  hh, hw = h / 2, w / 2
  # left or right?
  if -hh <= s * hw <= hh
- if xA > xB
- # right
- return xB + hw, yB + s * hw
- else # left
- return xB - hw, yB - s * hw
+ if xA > xB # right
+ xB + hw, yB + s * hw
+ else # left
+ xB - hw, yB - s * hw
  end
  # top or bot?
  elseif -hw <= hh / s <= hw
- if yA > yB
- # top
- return xB + hh / s, yB + hh
- else
- # bottom
- return xB - hh / s, yB - hh
+ if yA > yB # top
+ xB + hh / s, yB + hh
+ else # bottom
+ xB - hh / s, yB - hh
  end
  end
 end
@@ -1126,7 +1112,7 @@ error_tuple(x::Tuple) = x
 
 function error_coords(errorbar, errordata, otherdata...)
  ed = Vector{float_extended_type(errordata)}(undef, 0)
- od = [Vector{float_extended_type(odi)}(undef, 0) for odi in otherdata]
+ od = map(odi -> Vector{float_extended_type(odi)}(undef, 0), otherdata)
  for (i, edi) in enumerate(errordata)
  for (j, odj) in enumerate(otherdata)
  odi = _cycle(odj, i)
@@ -1287,7 +1273,7 @@ function quiver_using_hack(plotattributes::AKW)
  arrow_h = 0.1dist # height of arrowhead
  arrow_w = 0.5arrow_h # halfwidth of arrowhead
  U1 = v ./ dist # vector of arrowhead height
- U2 = P2((-U1[2], U1[1]))  # vector of arrowhead halfwidth
+ U2 = P2((-U1[2], U1[1])) # vector of arrowhead halfwidth
  U1 = U1 .* arrow_h
  U2 = U2 .* arrow_w