fix multithreading

Project.toml

@@ -1,7 +1,7 @@
 name = "ParameterSpacePartitions"
 uuid = "64931d76-8770-46b0-8423-0a5d3a7d2d72"
 authors = ["itsdfish"]
-version = "0.3.8"
+version = "0.3.9"
 
 [deps]
 ComponentArrays = "b0b7db55-cfe3-40fc-9ded-d10e2dbeff66"

README.md

@@ -31,20 +31,20 @@ A second function categorizes the predicted data into a qualitative pattern. At
 
 ```julia
 function p_fun(location, hypercube::HyperCube, args...; kwargs...)
-        p_bounds = hypercube.p_bounds
-        nb = length(p_bounds)
-        nd = length(location)
-        vals = fill(-100, nd)
-        for j in 1:nd
-            for i in 1:(nb-1) 
-                if (location[j] ≥ p_bounds[i]) && (location[j] ≤ p_bounds[i+1])
-                    vals[j] = i 
-                    continue
-                end
+    p_bounds = hypercube.p_bounds
+    nb = length(p_bounds)
+    nd = length(location)
+    vals = fill(-100, nd)
+    for j in 1:nd
+        for i in 1:(nb-1) 
+            if (location[j] ≥ p_bounds[i]) && (location[j] ≤ p_bounds[i+1])
+                vals[j] = i 
+                continue
             end
         end
-        return vals
     end
+    return vals
+end
 ```
 
 ## Model Configuration

src/sampler.jl

@@ -73,7 +73,12 @@ Uses threading to generate patterns associated with a vector of proposals
 - `options`: a set of options for configuring the algorithm
 
 """
-function t_eval_patterns(proposals, model, p_fun, options, patterns)
+function t_eval_patterns(proposals, model, p_fun, options, chains)
+    (;bounds) = options
+    return tmap((c,p) -> eval_pattern(p_fun, model, bounds, c, p), chains, proposals)
+end
+
+function t_eval_patterns(proposals, model, p_fun)
     return tmap(p -> p_fun(model(p)), proposals)
 end