From 8ca211f96d96ce9fd663415d9502848f4a1e50fc Mon Sep 17 00:00:00 2001
From: Tobie Morgan Hitchcock <tobie@abcum.com>
Date: Mon, 4 Jul 2022 10:41:18 +0100
Subject: [PATCH] Simplify SQL string::slug function implementation

---
 Cargo.lock                     | 15 +++------------
 lib/Cargo.toml                 |  2 +-
 lib/src/fnc/string.rs          |  4 ++--
 lib/src/fnc/util/mod.rs        |  1 +
 lib/src/fnc/util/string/mod.rs | 23 +++++++++++++++++++++++
 5 files changed, 30 insertions(+), 15 deletions(-)
 create mode 100644 lib/src/fnc/util/string/mod.rs

diff --git a/Cargo.lock b/Cargo.lock
index c30fcb45..74574b3e 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -614,9 +614,9 @@ dependencies = [
 
 [[package]]
 name = "deunicode"
-version = "0.4.3"
+version = "1.3.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "850878694b7933ca4c9569d30a34b55031b9b139ee1fc7b94a527c4ef960d690"
+checksum = "f2c9736e15e7df1638a7f6eee92a6511615c738246a052af5ba86f039b65aede"
 
 [[package]]
 name = "digest"
@@ -2575,15 +2575,6 @@ version = "0.4.6"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "eb703cfe953bccee95685111adeedb76fabe4e97549a58d16f03ea7b9367bb32"
 
-[[package]]
-name = "slug"
-version = "0.1.4"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "b3bc762e6a4b6c6fcaade73e77f9ebc6991b676f88bb2358bddb56560f073373"
-dependencies = [
- "deunicode",
-]
-
 [[package]]
 name = "smallvec"
 version = "1.8.0"
@@ -2694,6 +2685,7 @@ dependencies = [
  "bigdecimal",
  "boa_engine",
  "chrono",
+ "deunicode",
  "dmp",
  "echodb",
  "futures 0.3.21",
@@ -2715,7 +2707,6 @@ dependencies = [
  "serde",
  "sha-1 0.10.0",
  "sha2",
- "slug",
  "storekey",
  "surrealdb-derive",
  "thiserror",
diff --git a/lib/Cargo.toml b/lib/Cargo.toml
index ff343c0a..4b6bb95e 100644
--- a/lib/Cargo.toml
+++ b/lib/Cargo.toml
@@ -22,6 +22,7 @@ boa = { version = "0.15.0", package = "boa_engine", optional = true }
 channel = { version = "1.6.1", package = "async-channel" }
 chrono = { version = "0.4.19", features = ["serde"] }
 derive = { version = "0.3.0", package = "surrealdb-derive" }
+deunicode = "1.3.1"
 dmp = "0.1.1"
 echodb = { version = "0.3.0", optional = true }
 executor = { version = "1.4.1", package = "async-executor", optional = true }
@@ -44,7 +45,6 @@ scrypt = "0.10.0"
 serde = { version = "1.0.138", features = ["derive"] }
 sha-1 = "0.10.0"
 sha2 = "0.10.2"
-slug = "0.1.4"
 storekey = "0.3.0"
 thiserror = "1.0.31"
 tikv = { version = "0.1.0", package = "tikv-client", optional = true }
diff --git a/lib/src/fnc/string.rs b/lib/src/fnc/string.rs
index 7b86345c..67b5c831 100644
--- a/lib/src/fnc/string.rs
+++ b/lib/src/fnc/string.rs
@@ -1,7 +1,7 @@
 use crate::ctx::Context;
 use crate::err::Error;
+use crate::fnc::util::string;
 use crate::sql::value::Value;
-use slug::slugify;
 
 pub fn concat(_: &Context, args: Vec<Value>) -> Result<Value, Error> {
 	Ok(args.into_iter().map(|x| x.as_string()).collect::<Vec<_>>().concat().into())
@@ -56,7 +56,7 @@ pub fn slice(_: &Context, mut args: Vec<Value>) -> Result<Value, Error> {
 }
 
 pub fn slug(_: &Context, mut args: Vec<Value>) -> Result<Value, Error> {
-	Ok(slugify(&args.remove(0).as_string()).into())
+	Ok(string::slug(&args.remove(0).as_string()).into())
 }
 
 pub fn split(_: &Context, mut args: Vec<Value>) -> Result<Value, Error> {
diff --git a/lib/src/fnc/util/mod.rs b/lib/src/fnc/util/mod.rs
index a529dc13..2cd099a5 100644
--- a/lib/src/fnc/util/mod.rs
+++ b/lib/src/fnc/util/mod.rs
@@ -1,3 +1,4 @@
 pub mod geo;
 pub mod http;
 pub mod math;
+pub mod string;
diff --git a/lib/src/fnc/util/string/mod.rs b/lib/src/fnc/util/string/mod.rs
new file mode 100644
index 00000000..e9a08259
--- /dev/null
+++ b/lib/src/fnc/util/string/mod.rs
@@ -0,0 +1,23 @@
+use deunicode::deunicode;
+use once_cell::sync::Lazy;
+use regex::Regex;
+
+static SIMPLES: Lazy<Regex> = Lazy::new(|| Regex::new("[^a-z0-9-_]").unwrap());
+static HYPHENS: Lazy<Regex> = Lazy::new(|| Regex::new("-+").unwrap());
+
+pub fn slug<S: AsRef<str>>(s: S) -> String {
+	// Get a reference
+	let s = s.as_ref();
+	// Convert unicode to ascii
+	let s = deunicode(s);
+	// Convert strgin to lowercase
+	let s = s.to_ascii_lowercase();
+	// Replace any non-simple characters
+	let s = SIMPLES.replace_all(s.as_ref(), "-");
+	// Replace any duplicated hyphens
+	let s = HYPHENS.replace_all(s.as_ref(), "-");
+	// Remove any surrounding hyphens
+	let s = s.trim_matches('-');
+	// Return the string
+	s.to_owned()
+}