Ensure Pluto-defined structs have a stable hash (#44)

A bug in detecting anonymous types meant that Pluto workspace defined structs generated a different hash id each time the cell was updated. This ensures a stable ID for pluto-defined structs, and improves type-name validation.

Project.toml

@@ -1,7 +1,7 @@
 name = "StableHashTraits"
 uuid = "c5dd0088-6c3f-4803-b00e-f31a60c170fa"
 authors = ["Beacon Biosignals, Inc."]
-version = "1.1.3"
+version = "1.1.4"
 
 [deps]
 Compat = "34da2185-b29b-5c13-b0c7-acf172513d20"

src/StableHashTraits.jl

@@ -424,12 +424,25 @@ function cleanup_name(str)
     # later versions of julia, they do not
     str = replace(str, "AbstractVector{T} where T" => "AbstractVector")
     str = replace(str, "AbstractMatrix{T} where T" => "AbstractMatrix")
+    # cleanup pluto workspace names
+
+    # handle pluto symbols
+
+    # TODO: eventually, when we create hash version 3 (which will generate strings from
+    # scratch rather than leveraging `string(T)`), we should handle pluto symbols by
+    # checking `is_inside_pluto` as defined here
+    # https://github.com/JuliaPluto/PlutoHooks.jl/blob/f6bc0a3962a700257641c3449db344cf0ddeae1d/src/notebook.jl#L89-L98
+
+    # NOTE: in more recent julia versions (>= 1.8) the values are surrounded by `var`
+    # qualifiers
+    str = replace(str, r"var\"workspace#[0-9]+\"" => "PlutoWorkspace")
+    str = replace(str, r"workspace#[0-9]+" => "PlutoWorkspace")
     return str
 end
 
 function validate_name(str)
-    if occursin(r"\.#[^.]*$", str)
-        throw(ArgumentError("Anonymous types (those containing `#`) cannot be hashed to a reliable value"))
+    if occursin("#", str)
+        throw(ArgumentError("Anonymous types (those containing `#`) cannot be hashed to a reliable value: found type $str"))
     end
     return str
 end

test/Project.toml

@@ -5,6 +5,7 @@ BenchmarkTools = "6e4b80f9-dd63-53aa-95a3-0cdb28fa8baf"
 CRC32c = "8bf52ea8-c179-5cab-976a-9e18b702a9bc"
 DataFrames = "a93c6f00-e57d-5684-b7b6-d8193f3e46c0"
 Dates = "ade2ca70-3891-5945-98fb-dc099432e06a"
+Pluto = "c3e4b0f8-55cb-11ea-2926-15256bba5781"
 ReferenceTests = "324d217c-45ce-50fc-942e-d289b448e8cf"
 SHA = "ea8e919c-243c-51af-8825-aaa63cd721ce"
 Tables = "bd369af6-aec1-5ad0-b16a-f7cc5008161c"

test/references/pluto01_1_crc.txt

@@ -0,0 +1 @@
+e2568b25890efb3e569dc44cc24409249a2d2b1126bbf80e9b82fac526eb99db

test/references/pluto01_1_sha1.txt

@@ -0,0 +1 @@
+e2568b25890efb3e569dc44cc24409249a2d2b1126bbf80e9b82fac526eb99db

test/references/pluto01_1_sha256.txt

@@ -0,0 +1 @@
+e2568b25890efb3e569dc44cc24409249a2d2b1126bbf80e9b82fac526eb99db

test/references/pluto01_2_crc32c.txt

@@ -0,0 +1 @@
+e2568b25890efb3e569dc44cc24409249a2d2b1126bbf80e9b82fac526eb99db

test/references/pluto01_2_sha1.txt

@@ -0,0 +1 @@
+e2568b25890efb3e569dc44cc24409249a2d2b1126bbf80e9b82fac526eb99db

test/references/pluto01_2_sha256.txt

@@ -0,0 +1 @@
+e2568b25890efb3e569dc44cc24409249a2d2b1126bbf80e9b82fac526eb99db

test/references/pluto02_1_crc.txt

@@ -0,0 +1 @@
+c50004f5a3abd3c6c98926f354b08eb2557188086db60d661f3a0b7b7503fa7f

test/references/pluto02_1_sha1.txt

@@ -0,0 +1 @@
+c50004f5a3abd3c6c98926f354b08eb2557188086db60d661f3a0b7b7503fa7f

test/references/pluto02_1_sha256.txt

@@ -0,0 +1 @@
+c50004f5a3abd3c6c98926f354b08eb2557188086db60d661f3a0b7b7503fa7f

test/references/pluto02_2_crc32c.txt

@@ -0,0 +1 @@
+c50004f5a3abd3c6c98926f354b08eb2557188086db60d661f3a0b7b7503fa7f

test/references/pluto02_2_sha1.txt

@@ -0,0 +1 @@
+c50004f5a3abd3c6c98926f354b08eb2557188086db60d661f3a0b7b7503fa7f

test/references/pluto02_2_sha256.txt

@@ -0,0 +1 @@
+c50004f5a3abd3c6c98926f354b08eb2557188086db60d661f3a0b7b7503fa7f

test/runtests.jl

@@ -77,8 +77,8 @@ include("setup_tests.jl")
                 @test_reference("references/ref20_$(V)_$(nameof(hashfn)).txt",
                                 bytes2hex_(test_hash(==("test"))))
             end
-
             # verifies that transform can be called recursively
+
             @testset "FnHash" begin
                 @test test_hash(GoodTransform(2)) == test_hash(GoodTransform("-0.2"))
                 @test test_hash(GoodTransform(3)) != test_hash(GoodTransform("-0.2"))
@@ -202,6 +202,80 @@ include("setup_tests.jl")
                 @test_throws ArgumentError test_hash(BadHashMethod())
             end
 
+            @testset "Pluto-defined strucst are stable" begin
+                notebook_project_dir = joinpath(@__DIR__, "..")
+                @info "Notebook project: $notebook_project_dir"
+
+                notebook_str = """
+                ### A Pluto.jl notebook ###
+                # v0.19.36
+
+                using Markdown
+                using InteractiveUtils
+
+                # ╔═╡ 3592b099-9c96-4939-94b8-7ef2614b0955
+                import Pkg
+
+                # ╔═╡ 72871656-ae6e-11ee-2b23-251ac2aa38a3
+                begin
+                    Pkg.activate("$notebook_project_dir")
+                    using StableHashTraits
+                end
+
+                # ╔═╡ 1c505fa8-75fa-4ed2-8c3f-43e28135b55d
+                begin
+                    bytes2hex_(x::Number) = x
+                    bytes2hex_(x) = bytes2hex(x)
+                end
+
+                # ╔═╡ b449d8e9-7ede-4171-a5ab-044c338ebae2
+                struct MyStruct end
+
+                # ╔═╡ 1e683f1d-f5f6-4064-970c-1facabcf61cc
+                stable_hash(MyStruct()) |> bytes2hex_
+
+                # ╔═╡ f8f3a7a4-544f-456f-ac63-5b5ce91a071a
+                stable_hash((a=MyStruct, b=(c=MyStruct(), d=2))) |> bytes2hex_
+
+                # ╔═╡ Cell order:
+                # ╠═3592b099-9c96-4939-94b8-7ef2614b0955
+                # ╠═72871656-ae6e-11ee-2b23-251ac2aa38a3
+                # ╠═1c505fa8-75fa-4ed2-8c3f-43e28135b55d
+                # ╠═b449d8e9-7ede-4171-a5ab-044c338ebae2
+                # ╠═1e683f1d-f5f6-4064-970c-1facabcf61cc
+                # ╠═f8f3a7a4-544f-456f-ac63-5b5ce91a071a
+                """
+
+                server = Pluto.ServerSession()
+                server.options.evaluation.workspace_use_distributed = false
+                olddir = pwd()
+                nb = mktempdir() do dir
+                    path = joinpath(dir, "notebook.pluto.jl")
+                    write(path, notebook_str)
+                    nb = Pluto.load_notebook(path)
+                    Pluto.update_run!(server, nb, nb.cells)
+                    return nb
+                end
+                # pluto changes pwd
+                cd(olddir)
+
+                # NOTE: V refers to the hash version currently in loose
+                # its the `for` loop at the top of this file
+                if nb.cells[5].output.body isa Dict
+                    throw(Error("Failed notebook eval: $(nb.cells[5].output.body[:msg])"))
+                else
+                    @test_reference("references/pluto01_$(V)_$(nameof(hashfn)).txt",
+                                    strip(nb.cells[5].output.body, '"'))
+                end
+
+                if nb.cells[6].output.body isa Dict
+                    throw(Error("Failed notebook eval: $(nb.cells[6].output.body[:msg])"))
+                else
+                    @test_reference("references/pluto02_$(V)_$(nameof(hashfn)).txt",
+                                    strip(nb.cells[6].output.body, '"'))
+                end
+            end
+
             if V > 2 && hashfn == sha256
                 @testset "Hash-invariance to buffer size" begin
                     data = (rand(Int8, 2), rand(Int8, 2))

test/setup_tests.jl

@@ -9,6 +9,7 @@ using CRC32c
 using DataFrames
 using Tables
 using AWSS3
+using Pluto
 
 struct TestType
     a::Any