MAJOR: add aspell as additional check for spelling errors

Author: oktalz

.aspell.yml

@@ -0,0 +1,34 @@
+mode: all
+minlength: 3
+ignore:
+  - '*_test.go'
+allowed:
+  - aspell
+  - repo
+  - yaml
+  - config
+  - Github
+  - Gitlab
+  - env
+  - failsafe
+  - golang
+  - mkdir
+  - WORKDIR
+  - apk
+  - ENTRYPOINT
+  - ubuntu
+  - golangci
+  - splitted
+  - sudo
+  - mri
+  - IID
+  - repo
+  - OPTIM
+  - Submatch
+  - Lshortfile
+  - MAXSUBJECTLEN
+  - MAXSUBJECTPARTS
+  - MINSUBJECTLEN
+  - MINSUBJECTPARTS
+  - malformatted
+  - cmdline

.github/workflows/actions.yaml

@@ -6,8 +6,13 @@ jobs:
     runs-on: ubuntu-latest
     steps:
       - uses: actions/checkout@v4
+      - name: Set up Go
+        uses: actions/setup-go@v4
+        with:
+          go-version-file: 'go.mod'
+          check-latest: true
       - name: golangci-lint
-        uses: golangci/golangci-lint-action@v3
+        uses: golangci/golangci-lint-action@v6
   go_build:
     name: Go build
     runs-on: ubuntu-latest
@@ -32,6 +37,8 @@ jobs:
     runs-on: ubuntu-latest
     needs: ["go_build"]
     steps:
+    - name: Install Aspell
+      run: sudo apt-get update && sudo apt-get install -y aspell aspell-en
     - uses: actions/checkout@v4
     - name: Set up Go
       uses: actions/setup-go@v4
@@ -51,6 +58,8 @@ jobs:
     runs-on: ubuntu-latest
     needs: ["go_build"]
     steps:
+    - name: Install Aspell
+      run: sudo apt-get update && sudo apt-get install -y aspell aspell-en
     - uses: actions/checkout@v4
       with:
         fetch-depth: 0

Dockerfile

@@ -1,10 +1,11 @@
-FROM golang:alpine as builder
+FROM golang:alpine AS builder
 RUN mkdir /build
 ADD . /build/
 WORKDIR /build
 RUN go build -o check
 
 FROM alpine:latest
+RUN apk --no-cache add aspell aspell-en
 COPY --from=builder /build/check /check
 WORKDIR /
 ENTRYPOINT ["/check"]

README.md

@@ -32,7 +32,7 @@ bug: fix set-var parsing bug in config-parser
 ```
 BUG/MEDIUM: fix set-var
 ```
-- Unkown severity
+- Unknown severity
 ```
 BUG/MODERATE: fix set-var parsing bug in config-parser
 ```
@@ -95,3 +95,35 @@ TagOrder:
 ### Optional parameters
 
 The program accepts an optional parameter to specify the location (path) of the base of the git repository. This can be useful in certain cases where the checked-out repo is in a non-standard location within the CI environment, compared to the running path from which the check-commit binary is being invoked.
+
+### aspell
+
+to check also spellcheck errors aspell was added. it can be configured with `.aspell.yml`
+
+example
+```yaml
+mode: subject
+min_length: 3
+ignore:
+  - go.mod
+  - go.sum
+  - '*test.go'
+  - 'gen/*'
+allowed:
+  - aspell
+  - config
+```
+
+`min_length` is minimal word size that is checked (default: 3)
+
+`mode` can be set as
+
+- `subject`
+  - `default` option
+  - only subject of commit message will be checked
+- `commit`
+  - whole commit message will be checked
+- `all`
+  - both commit message and all code committed
+- `disabled`
+  - check is disabled

aspell/aspell.go

@@ -0,0 +1,163 @@
+package aspell
+
+import (
+	"bytes"
+	"fmt"
+	"log"
+	"os/exec"
+	"slices"
+	"sort"
+	"strings"
+
+	"check-commit/match"
+
+	"github.com/fatih/camelcase"
+)
+
+type Aspell struct {
+	Mode         mode     `yaml:"mode"`
+	MinLength    int      `yaml:"min_length"`
+	Ignore       []string `yaml:"ignore"`
+	AllowedWords []string `yaml:"allowed"`
+	HelpText     string   `yaml:"-"`
+}
+
+var (
+	camelCaseOK = map[string]struct{}{
+		"HAProxy": {},
+		"golang":  {},
+		"ascii":   {},
+		"api":     {},
+	}
+	camelCaseNotOK = map[string]struct{}{}
+)
+
+func (a Aspell) checkSingle(data string, allowedWords []string) error {
+	var words []string
+	var badWords []string
+
+	checkRes, err := checkWithAspellExec(data)
+	if checkRes != "" {
+		words = strings.Split(checkRes, "\n")
+	}
+	if err != nil {
+		return err
+	}
+
+	for _, word := range words {
+		wordLower := strings.ToLower(word)
+		if len(word) < a.MinLength {
+			continue
+		}
+		if _, ok := camelCaseNotOK[wordLower]; ok {
+			badWords = append(badWords, wordLower)
+			continue
+		}
+		if _, ok := camelCaseOK[wordLower]; ok {
+			continue
+		}
+		if slices.Contains(a.AllowedWords, wordLower) || slices.Contains(allowedWords, wordLower) {
+			continue
+		}
+		splitted := camelcase.Split(word)
+		if len(splitted) < 2 {
+			splitted = strings.FieldsFunc(word, func(r rune) bool {
+				return r == '_' || r == '-'
+			})
+		}
+		if len(splitted) > 1 {
+			for _, s := range splitted {
+				er := a.checkSingle(s, allowedWords)
+				if er != nil {
+					camelCaseNotOK[wordLower] = struct{}{}
+					badWords = append(badWords, word+":"+s)
+					break
+				}
+			}
+		} else {
+			camelCaseNotOK[wordLower] = struct{}{}
+			badWords = append(badWords, word)
+		}
+	}
+
+	if len(badWords) > 0 {
+		m := map[string]struct{}{}
+		for _, w := range badWords {
+			m[w] = struct{}{}
+		}
+		badWords = []string{}
+		for k := range m {
+			badWords = append(badWords, k)
+		}
+		sort.Strings(badWords)
+		return fmt.Errorf("aspell: %s", badWords)
+	}
+	return nil
+}
+
+func (a Aspell) Check(subjects []string, commitsFull []string, content []map[string]string) error {
+	var response string
+	var checks []string
+	switch a.Mode {
+	case modeDisabled:
+		return nil
+	case modeSubject:
+		checks = subjects
+	case modeCommit:
+		checks = commitsFull
+	case modeAll:
+		for _, file := range content {
+			for name, v := range file {
+				nextFile := false
+				for _, filter := range a.Ignore {
+					if match.MatchFilter(name, filter) {
+						// log.Println("File", name, "in ignore list")
+						nextFile = true
+						continue
+					}
+				}
+				if nextFile {
+					continue
+				}
+				var imports []string
+				if strings.HasSuffix(name, ".go") {
+					imports = match.GetImportWordsFromGoFile(name)
+				}
+				if err := a.checkSingle(v, imports); err != nil {
+					log.Println(name, err.Error())
+					response += fmt.Sprintf("%s\n", err)
+				}
+			}
+		}
+		checks = []string{}
+	default:
+		checks = subjects
+	}
+
+	for _, subject := range checks {
+		if err := a.checkSingle(subject, []string{}); err != nil {
+			response += fmt.Sprintf("%s\n", err)
+		}
+	}
+
+	if len(response) > 0 {
+		return fmt.Errorf("%s", response)
+	}
+	return nil
+}
+
+func checkWithAspellExec(subject string) (string, error) {
+	cmd := exec.Command("aspell", "--list")
+	cmd.Stdin = strings.NewReader(subject)
+
+	var stdout, stderr bytes.Buffer
+	cmd.Stdout = &stdout
+	cmd.Stderr = &stderr
+	err := cmd.Run()
+	if err != nil {
+		log.Printf("aspell error: %s, stderr: %s", err, stderr.String())
+		return "", err
+	}
+
+	return stdout.String() + stderr.String(), nil
+}

aspell/aspell_test.go

@@ -0,0 +1,33 @@
+package aspell
+
+import "testing"
+
+func Test_checkWithAspell(t *testing.T) {
+	aspell := Aspell{
+		Mode:         modeSubject,
+		MinLength:    3,
+		AllowedWords: []string{"config"},
+	}
+	tests := []struct {
+		name    string
+		subject string
+		wantErr bool
+	}{
+		{"OK 1", "BUG/MEDIUM: config: add default location of path to the configuration file", false},
+		{"OK 2", "BUG/MEDIUM: config: add default location of path to the configuration file xtra", false},
+		{"error - flie", "BUG/MEDIUM: config: add default location of path to the configuration flie", true},
+		{"error - locatoin", "CLEANUP/MEDIUM: config: add default locatoin of path to the configuration file", true},
+		{"error - locatoin+flie", "CLEANUP/MEDIUM: config: add default locatoin of path to the configuration flie", true},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			err := aspell.checkSingle(tt.subject, []string{"xtra"})
+			if tt.wantErr && err == nil {
+				t.Errorf("checkWithAspell() error = %v, wantErr %v", err, tt.wantErr)
+			}
+			if !tt.wantErr && err != nil {
+				t.Errorf("checkWithAspell() error = %v, wantErr %v", err, tt.wantErr)
+			}
+		})
+	}
+}

aspell/mode.go

@@ -0,0 +1,10 @@
+package aspell
+
+type mode string
+
+const (
+	modeDisabled mode = "disabled"
+	modeSubject  mode = "subject"
+	modeCommit   mode = "commit"
+	modeAll      mode = "all"
+)