Implemented an autocorrect module at the postprocessing stage

2026-04-26 19:33:23 -05:00
parent 99d0c425d6
commit 3928e0c4a7
7 changed files with 482 additions and 6 deletions
--- a/internal/autocorrect/autocorrect_test.go
+++ b/internal/autocorrect/autocorrect_test.go
@@ -0,0 +1,190 @@
+package autocorrect
+
+import (
+	"os"
+	"path/filepath"
+	"strings"
+	"testing"
+)
+
+func TestLoadValidRules(t *testing.T) {
+	dir := t.TempDir()
+	path := writeAutocorrect(t, dir, `autocorrect:
+  - target: "Hrank"
+    match:
+      - "Frank"
+      - "frank"
+`)
+
+	rules, err := Load(path)
+	if err != nil {
+		t.Fatalf("load rules: %v", err)
+	}
+
+	got, count := rules.Apply("Frank and frank")
+	if got != "Hrank and Hrank" {
+		t.Fatalf("text = %q, want %q", got, "Hrank and Hrank")
+	}
+	if count != 2 {
+		t.Fatalf("count = %d, want 2", count)
+	}
+}
+
+func TestLoadValidation(t *testing.T) {
+	tests := []struct {
+		name    string
+		content string
+		want    string
+	}{
+		{
+			name:    "missing top-level autocorrect",
+			content: `other: []`,
+			want:    "must contain at least one autocorrect rule",
+		},
+		{
+			name:    "empty rules list",
+			content: `autocorrect: []`,
+			want:    "must contain at least one autocorrect rule",
+		},
+		{
+			name: "empty target",
+			content: `autocorrect:
+  - target: ""
+    match: ["Frank"]
+`,
+			want: "must include target",
+		},
+		{
+			name: "empty match list",
+			content: `autocorrect:
+  - target: "Hrank"
+    match: []
+`,
+			want: "must include at least one match string",
+		},
+		{
+			name: "empty match string",
+			content: `autocorrect:
+  - target: "Hrank"
+    match: [" "]
+`,
+			want: "contains empty match string",
+		},
+		{
+			name: "duplicate match across rules",
+			content: `autocorrect:
+  - target: "Hrank"
+    match: ["Frank"]
+  - target: "Other"
+    match: ["Frank"]
+`,
+			want: `appears in both rule 0 and rule 1`,
+		},
+		{
+			name: "duplicate match within rule",
+			content: `autocorrect:
+  - target: "Hrank"
+    match: ["Frank", "Frank"]
+`,
+			want: `contains duplicate match string "Frank"`,
+		},
+	}
+
+	for _, test := range tests {
+		t.Run(test.name, func(t *testing.T) {
+			dir := t.TempDir()
+			path := writeAutocorrect(t, dir, test.content)
+
+			_, err := Load(path)
+			if err == nil {
+				t.Fatal("expected error")
+			}
+			if !strings.Contains(err.Error(), test.want) {
+				t.Fatalf("expected error to contain %q, got %v", test.want, err)
+			}
+		})
+	}
+}
+
+func TestApplyReplacementBehavior(t *testing.T) {
+	rules := Rules{rules: []Rule{
+		{
+			Target: "Hrank",
+			Match:  []string{"Frank"},
+		},
+		{
+			Target: "Mike Brown",
+			Match:  []string{"Mike Pat"},
+		},
+		{
+			Target: "Godfrey",
+			Match:  []string{"God-free"},
+		},
+	}}
+
+	tests := []struct {
+		name      string
+		input     string
+		want      string
+		wantCount int
+	}{
+		{
+			name:      "case sensitive",
+			input:     "Frank and FRANK",
+			want:      "Hrank and FRANK",
+			wantCount: 1,
+		},
+		{
+			name:      "punctuation boundary",
+			input:     "Frank, are you there?",
+			want:      "Hrank, are you there?",
+			wantCount: 1,
+		},
+		{
+			name:      "no substring in larger token",
+			input:     "Franklin and xFrank Frank_y Frank2",
+			want:      "Franklin and xFrank Frank_y Frank2",
+			wantCount: 0,
+		},
+		{
+			name:      "multi word match",
+			input:     "Hello Mike Pat.",
+			want:      "Hello Mike Brown.",
+			wantCount: 1,
+		},
+		{
+			name:      "hyphenated match",
+			input:     "God-free is here.",
+			want:      "Godfrey is here.",
+			wantCount: 1,
+		},
+		{
+			name:      "hyphen outside match is boundary",
+			input:     "x-Frank-y",
+			want:      "x-Hrank-y",
+			wantCount: 1,
+		},
+	}
+
+	for _, test := range tests {
+		t.Run(test.name, func(t *testing.T) {
+			got, count := rules.Apply(test.input)
+			if got != test.want {
+				t.Fatalf("text = %q, want %q", got, test.want)
+			}
+			if count != test.wantCount {
+				t.Fatalf("count = %d, want %d", count, test.wantCount)
+			}
+		})
+	}
+}
+
+func writeAutocorrect(t *testing.T, dir string, content string) string {
+	t.Helper()
+
+	path := filepath.Join(dir, "autocorrect.yml")
+	if err := os.WriteFile(path, []byte(content), 0o600); err != nil {
+		t.Fatalf("write autocorrect file: %v", err)
+	}
+	return path
+}