feat: clang-tidy parser (#26)

Co-authored-by: zjc_he <zjc_he@sjtu.edu.cn>
Reviewed-on: FOCS-dev/JOJ3#26
Reviewed-by: 张泊明518370910136 <bomingzh@sjtu.edu.cn>
Co-authored-by: 张佳澈520370910044 <zjc_he@sjtu.edu.cn>
Co-committed-by: 张佳澈520370910044 <zjc_he@sjtu.edu.cn>

.gitmodules

@@ -22,7 +22,11 @@
 	path = examples/keyword/cpplint/sillycode
 	url = ssh://git@focs.ji.sjtu.edu.cn:2222/FOCS-dev/JOJ3-examples.git
 	branch = keyword/cpplint/sillycode
-[submodule "examples/keyword/clang-tidy/sillycode"]
+[submodule "examples/clangtidy/sillycode"]
-	path = examples/keyword/clang-tidy/sillycode
+	path = examples/clangtidy/sillycode
 	url = ssh://git@focs.ji.sjtu.edu.cn:2222/FOCS-dev/JOJ3-examples.git
-	branch = keyword/clang-tidy/sillycode
+	branch = clangtidy/sillycode
+[submodule "examples/keyword/clangtidy/sillycode"]
+	path = examples/keyword/clangtidy/sillycode
+	url = ssh://git@focs.ji.sjtu.edu.cn:2222/FOCS-dev/JOJ3-examples.git
+	branch = keyword/clangtidy/sillycode

examples/clangtidy/sillycode

@@ -0,0 +1 @@
+Subproject commit a69a7e87fddaddf21fc4f9cd6774e310fa7137c1

internal/parsers/all.go

@@ -1,6 +1,7 @@
 package parsers
 
 import (
+	_ "focs.ji.sjtu.edu.cn/git/FOCS-dev/JOJ3/internal/parsers/clangtidy"
 	_ "focs.ji.sjtu.edu.cn/git/FOCS-dev/JOJ3/internal/parsers/cpplint"
 	_ "focs.ji.sjtu.edu.cn/git/FOCS-dev/JOJ3/internal/parsers/diff"
 	_ "focs.ji.sjtu.edu.cn/git/FOCS-dev/JOJ3/internal/parsers/dummy"

internal/parsers/clangtidy/convert.go

@@ -0,0 +1,134 @@
+// Referenced from https://github.com/yuriisk/clang-tidy-converter/blob/master/clang_tidy_converter/parser/clang_tidy_parser.py
+package clangtidy
+
+import (
+	"path/filepath"
+	"regexp"
+	"strconv"
+)
+
+type Level int
+
+const (
+	UNKNOWN Level = iota
+	NOTE
+	REMARK
+	WARNING
+	ERROR
+	FATAL
+)
+
+type ClangMessage struct {
+	filepath       string
+	line           int
+	column         int
+	level          Level
+	message        string
+	diagnosticName string
+	detailsLines   []string
+	children       []ClangMessage
+}
+
+func newClangMessage(filepath string, line int, column int, level Level, message string, diagnosticName string, detailsLines []string, children []ClangMessage) *ClangMessage {
+	if detailsLines == nil {
+		detailsLines = make([]string, 0)
+	}
+	if children == nil {
+		children = make([]ClangMessage, 0)
+	}
+
+	return &ClangMessage{
+		filepath:       filepath,
+		line:           line,
+		column:         column,
+		level:          level,
+		message:        message,
+		diagnosticName: diagnosticName,
+		detailsLines:   detailsLines,
+		children:       children,
+	}
+}
+
+func levelFromString(levelString string) Level {
+	switch levelString {
+	case "note":
+		return NOTE
+	case "remark":
+		return REMARK
+	case "warning":
+		return WARNING
+	case "error":
+		return ERROR
+	case "fatal":
+		return FATAL
+	default:
+		return UNKNOWN
+	}
+}
+
+func isIgnored(line string) bool {
+	ignoreRegex := regexp.MustCompile("^error:.*$")
+	return ignoreRegex.MatchString(line)
+}
+
+func parseMessage(line string) ClangMessage {
+	messageRegex := regexp.MustCompile(`^(?P<filepath>.+):(?P<line>\d+):(?P<column>\d+): (?P<level>\S+): (?P<message>.*?)(?: \[(?P<diagnostic_name>.*)\])?\n$`)
+	regexRes := messageRegex.FindStringSubmatch(line)
+	if len(regexRes) == 0 {
+		return *newClangMessage("", 0, 0, UNKNOWN, "", "", nil, nil)
+	} else {
+		filepath := regexRes[1]
+		line, _ := strconv.Atoi(regexRes[2])
+		column, _ := strconv.Atoi(regexRes[3])
+		level := levelFromString(regexRes[4])
+		message := regexRes[5]
+		diagnosticName := regexRes[6]
+
+		return ClangMessage{
+			filepath:       filepath,
+			line:           line,
+			column:         column,
+			level:          level,
+			message:        message,
+			diagnosticName: diagnosticName,
+			detailsLines:   make([]string, 0),
+			children:       make([]ClangMessage, 0),
+		}
+	}
+}
+
+func groupMessages(messages []ClangMessage) []ClangMessage {
+	groupedMessages := make([]ClangMessage, 0)
+	for _, message := range messages {
+		if message.level == NOTE {
+			groupedMessages[len(groupedMessages)-1].children = append(groupedMessages[len(groupedMessages)-1].children, message)
+		} else {
+			groupedMessages = append(groupedMessages, message)
+		}
+	}
+	return groupedMessages
+}
+
+func convertPathsToRelative(messages *[]ClangMessage, conf Conf) {
+	currentDir := conf.RootDir
+	for i := range *messages {
+		(*messages)[i].filepath, _ = filepath.Rel(currentDir, (*messages)[i].filepath)
+	}
+}
+
+func ParseLines(lines []string, conf Conf) []ClangMessage {
+	messages := make([]ClangMessage, 0)
+	for _, line := range lines {
+		if isIgnored(string(line)) {
+			continue
+		}
+		message := parseMessage(string(line))
+		if message.level == UNKNOWN && len(messages) > 0 {
+			messages[len(messages)-1].detailsLines = append(messages[len(messages)-1].detailsLines, string(line))
+		} else {
+			messages = append(messages, message)
+		}
+	}
+	convertPathsToRelative(&messages, conf)
+	return groupMessages(messages)
+}

internal/parsers/clangtidy/formatter.go

@@ -0,0 +1,184 @@
+// Referenced from https://github.com/yuriisk/clang-tidy-converter/blob/master/clang_tidy_converter/parser/clang_tidy_parser.py
+package clangtidy
+
+import (
+	"fmt"
+	"strings"
+)
+
+type JsonMessage struct {
+	Type        string                 `json:"type"`
+	CheckName   string                 `json:"checkname"`
+	Description string                 `json:"description"`
+	Content     map[string]interface{} `json:"content"`
+	Categories  []string               `json:"categories"`
+	Location    map[string]interface{} `json:"location"`
+	Trace       map[string]interface{} `json:"trace"`
+	Severity    string                 `json:"severity"`
+}
+
+func Format(messages []ClangMessage) []JsonMessage {
+	formattedMessages := make([]JsonMessage, len(messages))
+	for i, message := range messages {
+		formattedMessages[i] = formatMessage(message)
+	}
+	return formattedMessages
+}
+
+func formatMessage(message ClangMessage) JsonMessage {
+	result := JsonMessage{
+		Type:        "issue",
+		CheckName:   message.diagnosticName,
+		Description: message.message,
+		Content:     extractContent(message),
+		Categories:  extractCategories(message),
+		Location:    extractLocation(message),
+		Trace:       extractTrace(message),
+		Severity:    extractSeverity(message),
+	}
+	return result
+}
+
+func messagesToText(messages []ClangMessage) []string {
+	textLines := []string{}
+	for _, message := range messages {
+		textLines = append(textLines, fmt.Sprintf("%s:%d:%d: %s", message.filepath, message.line, message.column, message.message))
+		textLines = append(textLines, message.detailsLines...)
+		textLines = append(textLines, messagesToText(message.children)...)
+	}
+	return textLines
+}
+
+func extractContent(message ClangMessage) map[string]interface{} {
+	detailLines := ""
+	for _, line := range message.detailsLines {
+		if line == "" {
+			continue
+		}
+		detailLines += (line + "\n")
+	}
+	for _, line := range messagesToText(message.children) {
+		if line == "" {
+			continue
+		}
+		detailLines += (line + "\n")
+	}
+	result := map[string]interface{}{
+		"body": "```\n" + detailLines + "```",
+	}
+	return result
+}
+
+func removeDuplicates(list []string) []string {
+	uniqueMap := make(map[string]bool)
+	for _, v := range list {
+		uniqueMap[v] = true
+	}
+	result := []string{}
+	for k := range uniqueMap {
+		result = append(result, k)
+	}
+	return result
+}
+
+func extractCategories(message ClangMessage) []string {
+	bugriskCategory := "Bug Risk"
+	clarityCategory := "Clarity"
+	compatibilityCategory := "Compatibility"
+	complexityCategory := "Complexity"
+	duplicationCategory := "Duplication"
+	performanceCategory := "Performance"
+	securityCategory := "Security"
+	styleCategory := "Style"
+
+	categories := []string{}
+	if strings.Contains(message.diagnosticName, "bugprone") {
+		categories = append(categories, bugriskCategory)
+	}
+	if strings.Contains(message.diagnosticName, "modernize") {
+		categories = append(categories, compatibilityCategory)
+	}
+	if strings.Contains(message.diagnosticName, "portability") {
+		categories = append(categories, compatibilityCategory)
+	}
+	if strings.Contains(message.diagnosticName, "performance") {
+		categories = append(categories, performanceCategory)
+	}
+	if strings.Contains(message.diagnosticName, "readability") {
+		categories = append(categories, clarityCategory)
+	}
+	if strings.Contains(message.diagnosticName, "cloexec") {
+		categories = append(categories, securityCategory)
+	}
+	if strings.Contains(message.diagnosticName, "security") {
+		categories = append(categories, securityCategory)
+	}
+	if strings.Contains(message.diagnosticName, "naming") {
+		categories = append(categories, styleCategory)
+	}
+	if strings.Contains(message.diagnosticName, "misc") {
+		categories = append(categories, styleCategory)
+	}
+	if strings.Contains(message.diagnosticName, "cppcoreguidelines") {
+		categories = append(categories, styleCategory)
+	}
+	if strings.Contains(message.diagnosticName, "hicpp") {
+		categories = append(categories, styleCategory)
+	}
+	if strings.Contains(message.diagnosticName, "simplify") {
+		categories = append(categories, complexityCategory)
+	}
+	if strings.Contains(message.diagnosticName, "redundant") {
+		categories = append(categories, duplicationCategory)
+	}
+	if strings.HasPrefix(message.diagnosticName, "boost-use-to-string") {
+		categories = append(categories, compatibilityCategory)
+	}
+	if len(categories) == 0 {
+		categories = append(categories, bugriskCategory)
+	}
+	return removeDuplicates(categories)
+}
+
+func extractLocation(message ClangMessage) map[string]interface{} {
+	location := map[string]interface{}{
+		"path": message.filepath,
+		"lines": map[string]interface{}{
+			"begin": message.line,
+		},
+	}
+	return location
+}
+
+func extractOtherLocations(message ClangMessage) []map[string]interface{} {
+	locationList := []map[string]interface{}{}
+	for _, child := range message.children {
+		locationList = append(locationList, extractLocation(child))
+		locationList = append(locationList, extractOtherLocations(child)...)
+	}
+	return locationList
+}
+
+func extractTrace(message ClangMessage) map[string]interface{} {
+	result := map[string]interface{}{
+		"locations": extractOtherLocations(message),
+	}
+	return result
+}
+
+func extractSeverity(message ClangMessage) string {
+	switch message.level {
+	case NOTE:
+		return "info"
+	case REMARK:
+		return "minor"
+	case WARNING:
+		return "major"
+	case ERROR:
+		return "critical"
+	case FATAL:
+		return "blocker"
+	default:
+		return "unknown"
+	}
+}

internal/parsers/clangtidy/meta.go

@@ -0,0 +1,9 @@
+package clangtidy
+
+import "focs.ji.sjtu.edu.cn/git/FOCS-dev/JOJ3/internal/stage"
+
+var name = "clangtidy"
+
+func init() {
+	stage.RegisterParser(name, &ClangTidy{})
+}

internal/parsers/clangtidy/parser.go

@@ -0,0 +1,62 @@
+package clangtidy
+
+import (
+	"fmt"
+	"strings"
+
+	"focs.ji.sjtu.edu.cn/git/FOCS-dev/JOJ3/internal/stage"
+	"github.com/criyle/go-judge/envexec"
+)
+
+type Match struct {
+	Keyword []string
+	Score   int
+}
+
+type Conf struct {
+	Score   int    `default:"100"`
+	RootDir string `default:"/w"`
+	Matches []Match
+}
+
+type ClangTidy struct{}
+
+func Parse(executorResult stage.ExecutorResult, conf Conf) stage.ParserResult {
+	stdout := executorResult.Files["stdout"]
+	stderr := executorResult.Files["stderr"]
+
+	lines := strings.SplitAfter(stdout, "\n")
+	messages := ParseLines(lines, conf)
+	formattedMessages := Format(messages)
+
+	if executorResult.Status != stage.Status(envexec.StatusAccepted) {
+		if !((executorResult.Status == stage.Status(envexec.StatusNonzeroExitStatus)) &&
+			(executorResult.ExitStatus == 1)) {
+			return stage.ParserResult{
+				Score: 0,
+				Comment: fmt.Sprintf(
+					"Unexpected executor status: %s.\nStderr: %s",
+					executorResult.Status, stderr,
+				),
+			}
+		}
+	}
+	return stage.ParserResult{
+		Score:   GetScore(formattedMessages, conf),
+		Comment: GetComment(formattedMessages),
+	}
+}
+
+func (*ClangTidy) Run(results []stage.ExecutorResult, confAny any) (
+	[]stage.ParserResult, bool, error,
+) {
+	conf, err := stage.DecodeConf[Conf](confAny)
+	if err != nil {
+		return nil, true, err
+	}
+	var res []stage.ParserResult
+	for _, result := range results {
+		res = append(res, Parse(result, *conf))
+	}
+	return res, false, nil
+}

internal/parsers/clangtidy/score.go

@@ -0,0 +1,77 @@
+package clangtidy
+
+import (
+	"fmt"
+	"strings"
+)
+
+func contains(arr []string, element string) bool {
+	for i := range arr {
+		// TODO: The keyword in json report might also be an array, need to split it
+		if strings.Contains(arr[i], element) {
+			return true
+		}
+	}
+	return false
+}
+
+func GetScore(jsonMessages []JsonMessage, conf Conf) int {
+	fullmark := conf.Score
+	for _, jsonMessage := range jsonMessages {
+		keyword := jsonMessage.CheckName
+		for _, match := range conf.Matches {
+			if contains(match.Keyword, keyword) {
+				fullmark -= match.Score
+				break
+			}
+		}
+	}
+	return fullmark
+}
+
+func GetComment(jsonMessages []JsonMessage) string {
+	res := "### Test results summary\n\n"
+	keys := [...]string{
+		"codequality-unchecked-malloc-result",
+		"codequality-no-global-variables",
+		"codequality-no-header-guard",
+		"codequality-no-fflush-stdin",
+		"readability-function-size",
+		"readability-duplicate-include",
+		"readability-identifier-naming",
+		"readability-redundant",
+		"readability-misleading-indentation",
+		"readability-misplaced-array-index",
+		"cppcoreguidelines-init-variables",
+		"bugprone-suspicious-string-compare",
+		"google-global-names-in-headers",
+		"clang-diagnostic",
+		"clang-analyzer",
+		"misc",
+		"performance",
+		"others",
+	}
+	mapping := map[string]int{}
+	for _, key := range keys {
+		mapping[key] = 0
+	}
+	for _, jsonMessage := range jsonMessages {
+		keyword := jsonMessage.CheckName
+		flag := true
+		for key := range mapping {
+			if strings.Contains(keyword, key) {
+				mapping[key] += 1
+				flag = false
+				break
+			}
+		}
+		if flag {
+			mapping["others"] += 1
+		}
+	}
+
+	for i, key := range keys {
+		res = fmt.Sprintf("%s%d. %s: %d\n", res, i+1, key, mapping[key])
+	}
+	return res
+}