From 57c3c48f6660f905fb974cff7ec58f746a1a6970 Mon Sep 17 00:00:00 2001
From: Aria Shrimpton <me@aria.rip>
Date: Mon, 25 Mar 2024 14:01:37 +0000
Subject: cleanup analysis code

---
 analysis/vis.livemd | 139 ++++++++++++++++++++++++++++------------------------
 1 file changed, 76 insertions(+), 63 deletions(-)

(limited to 'analysis')

diff --git a/analysis/vis.livemd b/analysis/vis.livemd
index 360e2fe..4913e76 100644
--- a/analysis/vis.livemd
+++ b/analysis/vis.livemd
@@ -88,42 +88,8 @@ cost_model_points =
   |> DF.concat_rows()
 ```
 
-```elixir
-cost_models
-|> DF.filter(op == "contains")
-```
-
 ## Cost model exploratory plots
 
-```elixir
-startn = 1
-endn = 5000
-resolution = 1
-
-points_for = fn impl, op ->
-  %{"coeffs" => [coeffs]} =
-    DF.filter(cost_models, impl == ^impl and op == ^op)
-    |> DF.to_columns()
-
-  Enum.map(startn..endn//resolution, fn n ->
-    t =
-      (coeffs
-       |> Enum.take(3)
-       |> Enum.with_index()
-       |> Enum.map(fn {coeff, idx} -> coeff * n ** idx end)
-       |> Enum.sum()) + Enum.at(coeffs, 3) * Math.log2(n)
-
-    %{
-      impl: String.split(impl, "::") |> List.last(),
-      op: op,
-      n: n,
-      t: max(t, 0)
-    }
-  end)
-  |> DF.new()
-end
-```
-
 <!-- livebook:{"reevaluate_automatically":true} -->
 
 ```elixir
@@ -131,36 +97,83 @@ set_impls = ["BTreeSet", "HashSet", "VecSet", "SortedVecSet"]
 mapping_impls = ["HashMap", "BTreeMap", "VecMap", "SortedVecMap"]
 other_impls = ["Vec", "LinkedList", "SortedVec"]
 
-inspect_op = "remove"
-impls = ["Vec", "LinkedList"]
+impls = other_impls
 
-Tucan.layers([
-  cost_models
-  |> DF.filter(op == ^inspect_op)
-  |> DF.distinct(["impl"])
-  |> DF.to_rows()
-  |> Enum.map(fn %{"impl" => impl} -> points_for.(impl, inspect_op) end)
-  |> DF.concat_rows()
-  |> DF.filter(impl in ^impls)
-  |> Tucan.lineplot("n", "t", color_by: "impl", clip: true)
-  # Tucan.scatter(
-  #   cost_model_points
-  #   |> DF.filter(op == ^inspect_op and impl in ^impls)
-  #   |> DF.group_by(["impl", "n"]),
-  #   # |> DF.summarise(t: mean(t))
-  #   "n",
-  #   "t",
-  #   color_by: "impl",
-  #   clip: true
-  # )
-])
-|> Tucan.Axes.set_y_title("Estimated cost")
-|> Tucan.Axes.set_x_title("Size of container (n)")
-|> Tucan.Scale.set_x_domain(startn, endn)
-# |> Tucan.Scale.set_y_domain(0, 10_000)
-|> Tucan.set_size(500, 250)
-|> Tucan.Legend.set_title(:color, "Implementation")
-|> Tucan.Legend.set_orientation(:color, "bottom")
+defmodule CostModel do
+  @defaults %{y_domain: nil, ns: 1..60_000//100, draw_points: true}
+
+  def points_for(cost_models, ns, impl, op) do
+    %{"coeffs" => [coeffs]} =
+      DF.filter(cost_models, impl == ^impl and op == ^op)
+      |> DF.to_columns()
+
+    Enum.map(ns, fn n ->
+      t =
+        (coeffs
+         |> Enum.take(3)
+         |> Enum.with_index()
+         |> Enum.map(fn {coeff, idx} -> coeff * n ** idx end)
+         |> Enum.sum()) + Enum.at(coeffs, 3) * Math.log2(n)
+
+      %{
+        impl: String.split(impl, "::") |> List.last(),
+        op: op,
+        n: n,
+        t: max(t, 0)
+      }
+    end)
+    |> DF.new()
+  end
+
+  def plot(cost_models, cost_model_points, impls, op, opts \\ []) do
+    %{y_domain: y_domain, ns: ns, draw_points: draw_points} = Enum.into(opts, @defaults)
+
+    plot =
+      Tucan.layers(
+        [
+          cost_models
+          |> DF.filter(op == ^op)
+          |> DF.distinct(["impl"])
+          |> DF.to_rows()
+          |> Enum.map(fn %{"impl" => impl} -> points_for(cost_models, ns, impl, op) end)
+          |> DF.concat_rows()
+          |> DF.filter(impl in ^impls)
+          |> Tucan.lineplot("n", "t", color_by: "impl", clip: true)
+        ] ++
+          if(draw_points,
+            do: [
+              Tucan.scatter(
+                cost_model_points
+                |> DF.filter(op == ^op and impl in ^impls)
+                |> DF.group_by(["impl", "n"])
+                |> DF.summarise(t: mean(t)),
+                "n",
+                "t",
+                color_by: "impl",
+                clip: true
+              )
+            ],
+            else: []
+          )
+      )
+
+    plot =
+      plot
+      |> Tucan.Axes.set_y_title("Estimated cost")
+      |> Tucan.Axes.set_x_title("Size of container (n)")
+      |> Tucan.set_size(500, 250)
+      |> Tucan.Legend.set_title(:color, "Implementation")
+
+    case y_domain do
+      [lo, hi] -> Tucan.Scale.set_y_domain(plot, lo, hi)
+      _ -> plot
+    end
+  end
+end
+```
+
+```elixir
+CostModel.plot(cost_models, cost_model_points, other_impls, "remove")
 ```
 
 ## Read benchmark data
-- 
cgit v1.2.3