From 41734affb3cef63ef7c98643e9202cf0274b3950 Mon Sep 17 00:00:00 2001
From: joe-conigliaro <joe-conigliaro@users.noreply.github.com>
Date: Fri, 13 Sep 2019 21:10:24 +1000
Subject: [PATCH] compiler: detect typos in function/variable/module names

---
 compiler/fn.v             | 20 +++++++++++++
 compiler/parser.v         |  5 ++++
 compiler/table.v          | 63 +++++++++++++++++++++++++++++++++++++++
 vlib/strings/similarity.v | 59 ++++++++++++++++++++++++++++++++++++
 4 files changed, 147 insertions(+)
 create mode 100644 vlib/strings/similarity.v

diff --git a/compiler/fn.v b/compiler/fn.v
index 41e0f85745..fb5dac609e 100644
--- a/compiler/fn.v
+++ b/compiler/fn.v
@@ -1023,3 +1023,23 @@ fn (f &Fn) str_args(table &Table) string {
 	}
 	return s
 }
+
+// find local function variable with closest name to `name`
+fn (f &Fn) find_misspelled_local_var(name string, min_match f64) string {
+	mut closest := f64(0)
+	mut closest_var := ''
+	for var in f.local_vars {
+		n := '${f.mod}.$var.name'
+		if var.name == '' || !name.starts_with(f.mod) || (n.len - name.len > 3 || name.len - n.len > 3) { continue }
+		p := strings.dice_coefficient(name, n)
+		println(' ## $name - $n: $p')
+		if p > closest {
+			closest = p
+			closest_var = n
+		}
+	}
+	if closest >= min_match {
+		return closest_var
+	}
+	return ''
+}
diff --git a/compiler/parser.v b/compiler/parser.v
index 61e34fba3d..7b36357f13 100644
--- a/compiler/parser.v
+++ b/compiler/parser.v
@@ -1648,6 +1648,11 @@ fn (p mut Parser) name_expr() string {
 				f = p.table.find_fn(name)
 			}
 			if f.name == '' {
+				// check for misspelled function / variable / module
+				suggested := p.table.identify_typo(name, p.cur_fn, p.import_table)
+				if suggested != '' {
+					p.error('undefined: `$name`. did you mean:$suggested')
+				}
 				// If orig_name is a mod, then printing undefined: `mod` tells us nothing
 				// if p.table.known_mod(orig_name) {
 				if p.table.known_mod(orig_name) || p.import_table.known_alias(orig_name) {
diff --git a/compiler/table.v b/compiler/table.v
index da8e79594e..4366136c44 100644
--- a/compiler/table.v
+++ b/compiler/table.v
@@ -926,3 +926,66 @@ fn (t &Type) contains_field_type(typ string) bool {
 	}
 	return false
 }
+
+// check for a function / variable / module typo in `name`
+fn (table &Table) identify_typo(name string, current_fn &Fn, fit &FileImportTable) string {
+	// dont check if so short
+	if name.len < 2 { return '' }
+	min_match := 0.8 // for dice coefficient between 0.0 - 1.0
+	name_orig := name.replace('__', '.').replace('_dot_', '.')
+	mut output := ''
+	// check functions
+	mut n := table.find_misspelled_fn(name_orig, min_match)
+	if n != '' {
+		output += '\n  * function: `$n`'
+	}
+	// check function local variables
+	n = current_fn.find_misspelled_local_var(name_orig, min_match)
+	if n != '' {
+		output += '\n  * variable: `$n`'
+	}
+	// check imported modules
+	n = table.find_misspelled_imported_mod(name_orig, fit, min_match)
+	if n != '' {
+		output += '\n  * module: `$n`'
+	}
+	return output
+}
+
+// find function with closest name to `name`
+fn (table &Table) find_misspelled_fn(name string, min_match f64) string {
+	mut closest := f64(0)
+	mut closest_fn := ''
+	for _, f in table.fns {
+		n := '${f.mod}.$f.name'
+		if !name.starts_with(f.mod) || (n.len - name.len > 3 || name.len - n.len > 3) { continue }
+		p := strings.dice_coefficient(name, n)
+		if p > closest {
+			closest = p
+			closest_fn = n
+		}
+	}
+	if closest >= min_match {
+		return closest_fn
+	}
+	return ''
+}
+
+// find imported module with closest name to `name`
+fn (table &Table) find_misspelled_imported_mod(name string, fit &FileImportTable, min_match f64) string {
+	mut closest := f64(0)
+	mut closest_mod := ''
+	for alias, mod in fit.imports {
+		n := '${fit.module_name}.$alias'
+		if !name.starts_with(fit.module_name) || (n.len - name.len > 3 || name.len - n.len > 3) { continue }
+		p := strings.dice_coefficient(name, n)
+		if p > closest {
+			closest = p
+			closest_mod = '$alias ($mod)'
+		}
+	}
+	if closest >= min_match {
+		return closest_mod
+	}
+	return ''
+}
diff --git a/vlib/strings/similarity.v b/vlib/strings/similarity.v
new file mode 100644
index 0000000000..4b18c86bf0
--- /dev/null
+++ b/vlib/strings/similarity.v
@@ -0,0 +1,59 @@
+module strings
+
+// use levenshtein distance algorithm to calculate
+// the distance between between two strings (lower is closer)
+pub fn levenshtein_distance(a, b string) int {
+	mut f := [int(0); b.len+1]
+	for ca in a {
+		mut j := 1
+		mut fj1 := f[0]
+		f[0]++
+		for cb in b {
+			mut mn := if f[j]+1 <= f[j-1]+1 { f[j]+1 } else { f[j-1]+1 }
+			if cb != ca {
+				mn = if mn <= fj1+1 { mn } else { fj1+1 }
+			} else {
+				mn = if mn <= fj1 { mn } else { fj1 }
+			}
+			fj1 = f[j]
+			f[j] = mn
+			j++
+		}
+	}
+	return f[f.len-1]
+}
+
+// use levenshtein distance algorithm to calculate
+// how similar two strings are as a percentage (higher is closer)
+pub fn levenshtein_distance_percentage(a, b string) f64 {
+	d := levenshtein_distance(a, b)
+	l := if a.len >= b.len { a.len } else { b.len }
+	return (1.00 - f64(d)/f64(l)) * 100.00
+}
+
+// implementation of Sørensen–Dice coefficient.
+// find the similarity between two strings.
+// returns f64 between 0.0 (not similar) and 1.0 (exact match).
+pub fn dice_coefficient(s1, s2 string) f64 {
+	if s1.len == 0 || s2.len == 0 { return 0.0 }
+	if s1 == s2 { return 1.0 }
+	if s1.len < 2 || s2.len < 2 { return 0.0 }
+	mut first_bigrams := map[string]int
+	for i := 0; i < s1.len-1; i++ {
+		a := s1[i]
+		b := s1[i+1]
+		bigram := (a+b).str()
+		first_bigrams[bigram] = if bigram in first_bigrams { first_bigrams[bigram]+1 } else { 1 }
+	}
+	mut intersection_size := 0
+	for i := 0; i < s2.len-1; i++ {
+		a := s2[i]
+		b := s2[i+1]
+		bigram := (a+b).str()
+		count := if bigram in first_bigrams { first_bigrams[bigram] } else { 0 }
+		if count > 0 {
+			intersection_size++
+		}
+	}
+	return (2.0 * intersection_size) / (f64(s1.len) + f64(s2.len) - 2)
+}