Add support for serializing as numpy array and as a zip file with num…

…py array and meta info

Project.toml

@@ -5,10 +5,13 @@ version = "1.0.1"
 
 [deps]
 Distributions = "31c24e10-a181-5473-b8eb-7969acd0382f"
+JSON = "682c06a0-de6a-54ab-a142-c8b1cf79cde6"
 Plots = "91a5bcdd-55d7-5caf-9e0b-520d859cae80"
 ProgressLogging = "33c8b6b6-d38a-422a-b730-caa89a2f386c"
+PyCall = "438e738f-606a-5dbb-bf0a-cddfbfd45ab0"
 Serialization = "9e88b42a-f829-5b0c-bbe9-9e923198166b"
 StatsBase = "2913bbd2-ae8a-5f71-8c99-4fb6c76f3a91"
+ZipFile = "a5390f91-8eb1-5f08-bee0-b1d1ffed6cea"
 
 [compat]
 julia = "1"

notebooks/Project.toml

@@ -8,4 +8,6 @@ PlutoLinks = "0ff47ea0-7a50-410d-8455-4348d5de0420"
 PlutoTest = "cb4044da-4d16-4ffa-a6a3-8cad7f73ebdc"
 PlutoUI = "7f904dfe-b85e-4ff6-b463-dae2292396a8"
 Printf = "de0858da-6303-5e67-8744-51eddeeeb8d7"
+PyCall = "438e738f-606a-5dbb-bf0a-cddfbfd45ab0"
 Unzip = "41fe7b60-77ed-43a1-b4f0-825fd5a5650d"
+ZipFile = "a5390f91-8eb1-5f08-bee0-b1d1ffed6cea"

src/GridShielding.jl

@@ -1,17 +1,22 @@
 module GridShielding
 
 using Plots
-using Serialization
 using ProgressLogging
 
+# Serialization stuff
+using Serialization
+using JSON
+using PyCall
+using ZipFile
+
 # TODO: Only export relevant names (remember to update the testing-notebooks).
 export Bounds, get_dim, bounded, magnitude
 include("Bounds.jl")
 
 export Grid, Partition, get_size, box, bounds, set_value!, get_value, clear!, initialize!, indexof, draw, cover
 include("Grids.jl")
 
-export robust_grid_serialization, robust_grid_deserialization, get_c_library_header, get_libshield
+export robust_grid_serialization, robust_grid_deserialization, get_c_library_header, get_libshield, export_numpy_array, get_meta_info, export_meta_info, numpy_zip_file
 include("Serialisation.jl")
 
 export SupportingPoints, get_spacing_sizes

src/Serialisation.jl

@@ -1,5 +1,7 @@
 ⨝ = joinpath
 
+## Robust Grid (De)Serialisation ##
+
 """
     robust_grid_serialization(file, grid::Grid)	
 
@@ -43,6 +45,8 @@ function robust_grid_deserialization(file)
 	grid
 end
 
+## Exporting as C library ##
+
 function file_replace(file_path, replacements...)
 	lines = readlines(file_path)
 	open(file_path, "w") do file_buffer
@@ -193,3 +197,86 @@ function stringdump(grid)
 
 	return String(result)
 end
+
+## Exporting to Numpy array + JSON ##
+
+function export_numpy_array(grid, destination)
+	np = pyimport("numpy")
+	pyshield = np.array(grid.array)
+	open(destination, write=true, create=true) do 🗋
+		np.save(🗋, pyshield)
+	end
+end
+
+"""
+	get_meta_info(grid::Grid; variables::A, binary_variables::A, actions::Type, env_id::S) 
+		where {A<:AbstractArray, S<:AbstractString}
+
+Returns a dictionary containing structured meta-info on the grid. 
+
+Arguments: (NB: Mostly KW)
+ - `grid`: The grid in question. Bounds and granularity are inferred from this.
+ - `variables`: Axis labels for axes that are not binary.
+ - `binary_variables`: Labels for axes that are binary.
+ - `actions`: Enum representing the available axes. This will be used to create an `id_to_actionset` list.
+ - `env_id`: Descriptive name of the grid.
+
+Example: 
+
+	get_meta_info(grid′,
+		variables=["x", "y"], 
+		binary_variables=["a"],
+		actions=Action,
+		env_id="test grid")
+"""
+function get_meta_info(grid::Grid; variables::A, binary_variables::AA, actions::Type, env_id::S) where {A<:AbstractArray, AA<:AbstractArray, S<:AbstractString}
+	meta = (
+		"variables" => variables,
+		"env_id" => env_id,
+
+		"id_to_actionset" => [
+			(a => [a′ ∈ int_to_actions(actions, a) for a′ in instances(actions)])
+			for a in unique(grid.array) ],
+
+		"n_actions" => length(instances(actions)),
+		"actions" => instances(actions),
+		"bounds" => [grid.bounds.lower, grid.bounds.upper],
+		"granularity" => grid.granularity,
+		"bvars" => binary_variables
+	)
+end
+
+"""
+	numpy_zip_file(grid::Grid, destination; variables::A, binary_variables::AA, actions::Type, env_id::S) 
+	where {A<:AbstractArray, AA<:AbstractArray, S<:AbstractString}
+
+Returns a dictionary containing structured meta-info on the grid. 
+
+Arguments: (NB: Mostly KW)
+ - `grid`: The grid in question. Bounds and granularity are inferred from this.
+ - `destination`: File path to save to. Errors if exists.
+ - `variables`: Axis labels for axes that are not binary.
+ - `binary_variables`: Labels for axes that are binary.
+ - `actions`: Enum representing the available axes. This will be used to create an `id_to_actionset` list.
+ - `env_id`: Descriptive name of the grid.
+
+Example: 
+
+	numpy_zip_file(grid′, "path/to/save/grid.zip",
+		variables=["x", "y"], 
+		binary_variables=["a"],
+		actions=Action,
+		env_id="test grid")
+"""
+function numpy_zip_file(grid::Grid, destination; variables::A, binary_variables::AA, actions::Type, env_id::S) where {A<:AbstractArray, AA<:AbstractArray, S<:AbstractString}
+	w = ZipFile.Writer(destination)
+
+	np = pyimport("numpy")
+	pyshield = np.array(grid.array)
+	np.save(ZipFile.addfile(w, "grid.npy"), pyshield)
+
+	meta_info = get_meta_info(grid; variables, binary_variables, actions, env_id)
+
+	write(ZipFile.addfile(w, "meta.json"), JSON.json(meta_info))
+	close(w)
+end

test/Serialisation.jl

@@ -1,5 +1,5 @@
 ### A Pluto.jl notebook ###
-# v0.19.27
+# v0.19.36
 
 using Markdown
 using InteractiveUtils
@@ -25,26 +25,32 @@ begin
 	using PlutoLinks
 	using PlutoUI
 	using Distributions
+	using ZipFile
+	using PyCall
 	TableOfContents()
 end
 
+# ╔═╡ 5db17b6d-62b9-4339-b2b4-64472c9e0318
+Pkg.add("PyCall")
+
 # ╔═╡ 515c5c0b-a734-406c-b89d-6c921001a777
 @revise using GridShielding
 
 # ╔═╡ e4f088b7-b48a-4c6f-aa36-fc9fd4746d9b
 md"""
-# Bouncing Ball Example
+# Serializatoin Tests
 
 ## Preface
 """
 
-# ╔═╡ 7612a253-eda4-47a8-bb93-d5fcccc6c628
-begin
-end
-
 # ╔═╡ 1c03e3a8-6de3-49ff-a05d-c22fc41a5de7
 ⨝ = joinpath
 
+# ╔═╡ 670ff2b2-cb12-4840-857f-7d720bc71834
+md"""
+## Example Grid
+"""
+
 # ╔═╡ 55cd8522-c016-455b-bfb5-0257e8ccc334
 r = 0.9
 
@@ -96,7 +102,7 @@ let
 	draw(grid, slice,
 		xlabel="x",
 		ylabel="y",
-		label="z=0",
+		title="z=0",
 		colors=[:white, :wheat],
 		aspectratio=:equal,
 		size=(500, 300))
@@ -105,13 +111,28 @@ let
 end
   ╠═╡ =#
 
+# ╔═╡ 37d1be0a-8920-4df5-8e4e-8682ef8b86fc
+md"""
+##  Robust Grid (De)Serialization
+"""
+
 # ╔═╡ 9af0dcfa-6699-4eb9-bfe1-e99ff010b12f
 begin
+	# Reactivity
 	@use_file "/home/asger/.julia/dev/GridShielding/misc/shield.c"
 	grid
+
 	working_dir = mktempdir()
 end
 
+# ╔═╡ 292ecf52-e588-4ee6-96f7-f313fb825a64
+@bind open_folder_button CounterButton("Open Folder")
+
+# ╔═╡ c62e53c0-c580-4313-8f1a-1c8f63299a5e
+if open_folder_button > 0
+	run(`nautilus $working_dir`, wait=false)
+end; "This cell opens `$working_dir` in nautilus"
+
 # ╔═╡ 8c814127-99fa-4eb0-99a9-d82c68d2fd4e
 grid
 
@@ -207,20 +228,95 @@ let
 	sort(tests, by=t -> t isa PlutoTest.Fail, rev=true)
 end
 
+# ╔═╡ 65457eee-e990-4797-9929-cefb437c855d
+md"""
+## Python Export
+"""
+
+# ╔═╡ f9a4f12e-d3a9-4875-ace2-e214f7bf1c00
+@enum Action foo bar
+
+# ╔═╡ ffaf9dbd-8608-4bd6-814e-cc6fa6aad57d
+begin
+	grid′ = Grid(grid.granularity, grid.bounds)
+
+	for partition in grid
+		partition′ = Partition(grid′, partition.indices)
+		set_value!(partition′, get_value(partition) >= 1 ? rand((1, 2, 3)) : 0)
+	end
+end
+
+# ╔═╡ be9d9689-c594-4536-9ef5-d6b7f5c1d218
+@doc get_meta_info
+
+# ╔═╡ 5aec68ca-10ea-4705-81c5-6ae7d4d3c299
+typeof([])<:AbstractArray
+
+# ╔═╡ 5db23cbc-8251-4a9a-b4e9-6decf1a6efa6
+let
+	slice::Vector{Any} = partition.indices
+	slice[1] = slice[2] = Colon()
+
+	draw(grid′, slice,
+		xlabel="x",
+		ylabel="y",
+		title="z=0",
+		colors=[:white, :wheat, :beige, :moccasin],
+		aspectratio=:equal,
+		size=(500, 300))
+end
+
+# ╔═╡ d5e30af2-ee54-4132-847e-74e8b1af3eb2
+get_meta_info(grid′,
+	variables=["x", "y"], 
+	binary_variables=String[],
+	actions=Action,
+	env_id="test grid")
+
+# ╔═╡ 70fc0404-ac2e-491c-8776-c49d4bb16bbb
+grid_zip = working_dir ⨝ "grid.zip"
+
+# ╔═╡ e1af4cc0-c631-49e4-8df3-73000e96243a
+numpy_zip_file(grid′, grid_zip,
+	variables=["x", "y"], 
+	binary_variables=String[],
+	actions=Action,
+	env_id="test grid")
+
+# ╔═╡ b10e5910-99bb-4ba3-b366-ebf78ab0b9fe
+np = pyimport("numpy")
+
+# ╔═╡ 24b3e394-3f00-4720-86d0-4acd9cc51b00
+let
+	array = nothing
+	for f in ZipFile.Reader(grid_zip).files
+		if f.name == "grid.npy"
+			grid_npy = working_dir ⨝ "grid.npy"
+			write(grid_npy, f)
+			array = np.load(grid_npy)
+		end
+	end
+	@test grid′.array == array
+end
+
 # ╔═╡ Cell order:
 # ╟─e4f088b7-b48a-4c6f-aa36-fc9fd4746d9b
 # ╠═bb902940-a858-11ed-2f11-1d6f5af61e4a
-# ╠═7612a253-eda4-47a8-bb93-d5fcccc6c628
+# ╠═5db17b6d-62b9-4339-b2b4-64472c9e0318
 # ╠═1c03e3a8-6de3-49ff-a05d-c22fc41a5de7
 # ╠═515c5c0b-a734-406c-b89d-6c921001a777
+# ╟─670ff2b2-cb12-4840-857f-7d720bc71834
 # ╠═55cd8522-c016-455b-bfb5-0257e8ccc334
 # ╠═2a90dc85-62ab-42a9-85ba-5ad4f1d58aae
 # ╠═76231c64-67f4-4763-b26c-5f8fac3e6e10
 # ╠═5827c1bd-9932-4bf4-a7ab-aee75894443a
 # ╠═d1a62487-04c8-4162-815e-8a9257f6d049
 # ╟─6e395fff-e3d9-4b69-9fa2-0b012883a1c4
 # ╟─189387ac-d7fa-4328-ad11-df2740ace9a2
+# ╟─37d1be0a-8920-4df5-8e4e-8682ef8b86fc
 # ╠═9af0dcfa-6699-4eb9-bfe1-e99ff010b12f
+# ╠═292ecf52-e588-4ee6-96f7-f313fb825a64
+# ╠═c62e53c0-c580-4313-8f1a-1c8f63299a5e
 # ╠═8c814127-99fa-4eb0-99a9-d82c68d2fd4e
 # ╠═9d84159b-6bd7-4e86-89cd-fc0673c2b6ef
 # ╠═91ad5f38-27c6-4de7-8514-3af5b0714fcf
@@ -244,3 +340,14 @@ end
 # ╠═4c994427-caac-4106-8c73-0ced38302059
 # ╠═3e50f49b-ae5a-4c9b-9930-d2abf3539dbc
 # ╠═880437c4-f74d-4e0b-9795-acf52fe40332
+# ╟─65457eee-e990-4797-9929-cefb437c855d
+# ╠═f9a4f12e-d3a9-4875-ace2-e214f7bf1c00
+# ╠═ffaf9dbd-8608-4bd6-814e-cc6fa6aad57d
+# ╠═be9d9689-c594-4536-9ef5-d6b7f5c1d218
+# ╠═5aec68ca-10ea-4705-81c5-6ae7d4d3c299
+# ╟─5db23cbc-8251-4a9a-b4e9-6decf1a6efa6
+# ╠═d5e30af2-ee54-4132-847e-74e8b1af3eb2
+# ╠═70fc0404-ac2e-491c-8776-c49d4bb16bbb
+# ╠═e1af4cc0-c631-49e4-8df3-73000e96243a
+# ╠═b10e5910-99bb-4ba3-b366-ebf78ab0b9fe
+# ╠═24b3e394-3f00-4720-86d0-4acd9cc51b00