WIP: Use Rust NIF for word checking

.gitignore

@@ -26,3 +26,4 @@ checkcheckcheck-*.tar
 /tmp/
 
 .elixir_ls
+priv

README.md

@@ -10,14 +10,14 @@ To get errors and suggestions from an input file or a string:
 ## Examples
 
 ```
-$ mix correct "hello wrld waht is up?
+$ mix correct "hello wrld waht is up?"
 Errors in input:
 Word: wrld, Suggestions: [world, wild, weld]
 Word: waht, Suggestions: [what, wat, wah]
 ```
 
 ```
-$ ex: `mix correct ./data/inputs/ashford-short.txt`
+$ mix correct ./data/inputs/ashford-short.txt
 Errors in input:
 Word: peaple, Suggestions: [people, peale, pepple]
 Word: salteena, Suggestions: []
@@ -32,16 +32,16 @@ Simply run: `mix test`
 ## Performance tests
 
 `./data/inputs/ashford-short.txt`  
-Total time: 1.176746s  
+Total time: 1.17s  
 Words: 490  
 Words per second: 416.41 word/s
 
 `./data/inputs/ashford.txt`  
-Total time: 58.445784s  
+Total time: 58.44s  
 Words: 12,484  
 Words per second: 213.6 word/s
 
 `"hello wrld waht is up?"`  
-Total time: 0.104436s  
+Total time: 0.10s  
 Words: 5  
 Words per second: 47.88 word/s

data/dictionnary-sm.txt

@@ -0,0 +1,2 @@
+hello
+world

lib/mix/tasks/correct.ex

@@ -16,10 +16,11 @@ defmodule Mix.Tasks.Correct do
       end
 
     result = SpellChecker.check(input)
+    # IO.inspect(result)
 
-    Enum.reduce(result, "Errors in input:\n", fn error, acc ->
-      acc <> "Word: #{elem(error, 1)}, Suggestions: [#{Enum.join(elem(error, 2), ", ")}]\n"
-    end)
+    # Enum.reduce(result, "Errors in input:\n", fn error, acc ->
+    #   acc <> "Word: #{elem(error, 1)}, Suggestions: [#{Enum.join(elem(error, 2), ", ")}]\n"
+    # end)
   end
 
   @impl Mix.Task

lib/rust_word_checker.ex

@@ -0,0 +1,7 @@
+defmodule RustWordChecker do
+  use Rustler, otp_app: :checkcheckcheck, crate: "rustwordchecker"
+
+  # When your NIF is loaded, it will override this function.
+  # def check(_word, _dictionnary), do: :erlang.nif_error(:nif_not_loaded)
+  def correct(_text, _dictionnary), do: :erlang.nif_error(:nif_not_loaded)
+end

lib/spell_checker.ex

@@ -6,7 +6,7 @@ defmodule SpellChecker do
   end
 
   defp check_sentence(sentence, dictionnary) do
-    Enum.map(String.split(sentence), &WordChecker.check(cleaned(&1), dictionnary, 0))
+    Enum.map(String.split(sentence), &WordChecker.check(cleaned(&1), dictionnary))
     |> Enum.filter(&match?({:error, _, _}, &1))
   end
 
@@ -27,11 +27,23 @@ defmodule SpellChecker do
       Task.await_many(tasks, 100_000) |> List.flatten()
     end
 
+    RustWordChecker.correct(text, dictionnary)
+
     {u_secs, result} = :timer.tc(run_checker)
     word_count = length(String.split(text))
     Logger.info("Total time: #{u_secs / 1_000_000}s")
-    Logger.info("Words: #{word_count}")
-    Logger.info("Words per second: #{(word_count / (u_secs / 1_000_000)) |> Float.ceil(2)} word/s")
+    # Logger.info("Words: #{word_count}")
+    # Logger.info("Words per second: #{(word_count / (u_secs / 1_000_000)) |> Float.ceil(2)} word/s")
+
+    # IO.inspect(result)
+
     result
+
+    # {u_secs, result} = :timer.tc(run_checker)
+    # word_count = length(String.split(text))
+    # Logger.info("Total time: #{u_secs / 1_000_000}s")
+    # Logger.info("Words: #{word_count}")
+    # Logger.info("Words per second: #{(word_count / (u_secs / 1_000_000)) |> Float.ceil(2)} word/s")
+    # result
   end
 end

mix.exs

@@ -21,6 +21,7 @@ defmodule SpellChecker.MixProject do
   # Run "mix help deps" to learn about dependencies.
   defp deps do
     [
+      {:rustler, "~> 0.25.0"}
       # {:dep_from_hexpm, "~> 0.3.0"},
       # {:dep_from_git, git: "https://github.com/elixir-lang/my_dep.git", tag: "0.1.0"}
     ]

mix.lock

@@ -0,0 +1,5 @@
+%{
+  "jason": {:hex, :jason, "1.3.0", "fa6b82a934feb176263ad2df0dbd91bf633d4a46ebfdffea0c8ae82953714946", [:mix], [{:decimal, "~> 1.0 or ~> 2.0", [hex: :decimal, repo: "hexpm", optional: true]}], "hexpm", "53fc1f51255390e0ec7e50f9cb41e751c260d065dcba2bf0d08dc51a4002c2ac"},
+  "rustler": {:hex, :rustler, "0.25.0", "32526b51af7e58a740f61941bf923486ce6415a91c3934cc16c281aa201a2240", [:mix], [{:jason, "~> 1.0", [hex: :jason, repo: "hexpm", optional: false]}, {:toml, "~> 0.6", [hex: :toml, repo: "hexpm", optional: false]}], "hexpm", "6b43a11a37fe79c6234d88c4102ab5dfede7a6a764dc5c7b539956cfa02f3cf4"},
+  "toml": {:hex, :toml, "0.6.2", "38f445df384a17e5d382befe30e3489112a48d3ba4c459e543f748c2f25dd4d1", [:mix], [], "hexpm", "d013e45126d74c0c26a38d31f5e8e9b83ea19fc752470feb9a86071ca5a672fa"},
+}

native/rustwordchecker/.cargo/config

@@ -0,0 +1,5 @@
+[target.'cfg(target_os = "macos")']
+rustflags = [
+    "-C", "link-arg=-undefined",
+    "-C", "link-arg=dynamic_lookup",
+]

native/rustwordchecker/.gitignore

@@ -0,0 +1 @@
+/target

native/rustwordchecker/Cargo.toml

@@ -0,0 +1,13 @@
+[package]
+name = "rustwordchecker"
+version = "0.1.0"
+authors = []
+edition = "2018"
+
+[lib]
+name = "rustwordchecker"
+path = "src/lib.rs"
+crate-type = ["cdylib"]
+
+[dependencies]
+rustler = "0.25.0"

native/rustwordchecker/README.md

@@ -0,0 +1,20 @@
+# NIF for Elixir.RustWordChecker
+
+## To build the NIF module:
+
+- Your NIF will now build along with your project.
+
+## To load the NIF:
+
+```elixir
+defmodule RustWordChecker do
+    use Rustler, otp_app: :checkcheckcheck, crate: "rustwordchecker"
+
+    # When your NIF is loaded, it will override this function.
+    def add(_a, _b), do: :erlang.nif_error(:nif_not_loaded)
+end
+```
+
+## Examples
+
+[This](https://github.com/hansihe/NifIo) is a complete example of a NIF written in Rust.