created interface to parse different databases in a different way

cmd/example/main.go

@@ -6,7 +6,6 @@ import (
 	"time"
 
 	"github.com/devops-works/slowql"
-	"github.com/sirupsen/logrus"
 )
 
 func main() {
@@ -19,37 +18,36 @@ func main() {
 	if err != nil {
 		panic(err)
 	}
-	p := slowql.NewParser(fd)
 
+	p := slowql.NewParser(slowql.MariaDB, fd)
 	var count int
 	start := time.Now()
 	for {
-		q, err := p.GetNext()
-		if err != nil {
-			logrus.Error(err)
-		}
-		if q.Query == "" {
+		q := p.GetNext()
+		if q == (slowql.Query{}) {
 			break
 		}
-		// fmt.Printf("Time: %s\nUser: %s\nHost: %s\nID: %d\nSchema: %s\nLast_errno: %d\nKilled: %d\nQuery_time: %s\nLock_time: %s\nRows_sent: %d\nRows_examined: %d\nRows_affected: %d\nBytes_sent: %d\nQuery: %s\n",
-		// 	q.Time,
-		// 	q.User,
-		// 	q.Host,
-		// 	q.ID,
-		// 	q.Schema,
-		// 	q.LastErrNo,
-		// 	q.Killed,
-		// 	q.QueryTime,
-		// 	q.LockTime,
-		// 	q.RowsSent,
-		// 	q.RowsExamined,
-		// 	q.RowsAffected,
-		// 	q.BytesSent,
-		// 	q.Query,
-		// )
+
+		fmt.Printf("Time: %s\nUser: %s\nHost: %s\nID: %d\nSchema: %s\nLast_errno: %d\nKilled: %d\nQuery_time: %f\nLock_time: %f\nRows_sent: %d\nRows_examined: %d\nRows_affected: %d\nBytes_sent: %d\nQuery: %s\n",
+			q.Time,
+			q.User,
+			q.Host,
+			q.ID,
+			q.Schema,
+			q.LastErrNo,
+			q.Killed,
+			q.QueryTime,
+			q.LockTime,
+			q.RowsSent,
+			q.RowsExamined,
+			q.RowsAffected,
+			q.BytesSent,
+			q.Query,
+		)
 		count++
 	}
 
 	elapsed := time.Since(start)
-	fmt.Printf("parsed %d queries in %s\n", count, elapsed)
+	fmt.Printf("\nparsed %d queries in %s\n", count, elapsed)
+
 }

go.sum

@@ -1,12 +1,9 @@
-github.com/davecgh/go-spew v1.1.1 h1:vj9j/u1bqnvCEfJOwUhtlOARqs3+rkHYY13jYWTU97c=
 github.com/davecgh/go-spew v1.1.1/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
 github.com/magefile/mage v1.10.0 h1:3HiXzCUY12kh9bIuyXShaVe529fJfyqoVM42o/uom2g=
 github.com/magefile/mage v1.10.0/go.mod h1:z5UZb/iS3GoOSn0JgWuiw7dxlurVYTu+/jHXqQg881A=
-github.com/pmezard/go-difflib v1.0.0 h1:4DBwDE0NGyQoBHbLQYPwSUPoCMWR5BEzIk/f1lZbAQM=
 github.com/pmezard/go-difflib v1.0.0/go.mod h1:iKH77koFhYxTK1pcRnkKkqfTogsbg7gZNVY4sRDYZ/4=
 github.com/sirupsen/logrus v1.8.0 h1:nfhvjKcUMhBMVqbKHJlk5RPrrfYr/NMo3692g0dwfWU=
 github.com/sirupsen/logrus v1.8.0/go.mod h1:4GuYW9TZmE769R5STWrRakJc4UqQ3+QQ95fyz7ENv1A=
-github.com/stretchr/testify v1.2.2 h1:bSDNvY7ZPG5RlJ8otE/7V6gMiyenm9RtJ7IUVIAoJ1w=
 github.com/stretchr/testify v1.2.2/go.mod h1:a8OnRcib4nhh0OaRAV+Yts87kKdq0PP7pXfy6kDkUVs=
 golang.org/x/sys v0.0.0-20191026070338-33540a1f6037 h1:YyJpGZS1sBuBCzLAR1VEpK193GlqGZbnPFnPV/5Rsb4=
 golang.org/x/sys v0.0.0-20191026070338-33540a1f6037/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=

query.go → mariadb.go

@@ -1,100 +1,97 @@
 package slowql
 
 import (
-	"crypto/md5"
 	"fmt"
-	"io"
 	"strconv"
 	"strings"
+	"time"
 
 	"github.com/sirupsen/logrus"
 )
 
-// Query contains query informations. Those fields can be accessed from the
-// outside of the package to do calculation. If a field does not exists in the
-// log file, it will be empty in the returned query.
-type Query struct {
-	// Date and time of the query
-	Time string
+// MariaDB is the MariaDB kind
+const MariaDB Kind = 1
 
-	// The user who issued the query
-	User string
-
-	// The host from where the query were issued
-	Host string
-
-	// The ID of the query
-	ID int
-
-	// The table schema
-	Schema string
-
-	// Value of the error message from previous MySQL operation
-	LastErrNo int
-
-	Killed int
-
-	// Duration of the query
-	QueryTime string
-
-	// Duration of the lock time due to the query
-	LockTime string
-
-	// Rows sent during query
-	RowsSent int
-
-	// Rows examined during query
-	RowsExamined int
-
-	// Rows affected by the query
-	RowsAffected int
-
-	// Bytes sent by the query
-	BytesSent int
+type mariadbParser struct {
+	wl chan Query
+}
 
-	// The query itself. If therre were multiple queries, they are appended
-	Query string
+func (p *mariadbParser) parseBlocs(rawBlocs chan []string) {
+	for {
+		select {
+		case bloc := <-rawBlocs:
+			var q Query
+
+			for _, line := range bloc {
+				if strings.HasPrefix(line, "#") {
+					q.parseMariaDBHeader(line)
+				} else {
+					q.Query = q.Query + line
+				}
+			}
+			p.wl <- q
+		}
+	}
 }
 
-// Fingerprint returns Query.query's MD5 fingerprint.
-func (q Query) Fingerprint() (string, error) {
-	h := md5.New()
-	_, err := io.WriteString(h, q.Query)
-	if err != nil {
-		return "", err
+func (p *mariadbParser) GetNext() Query {
+	var q Query
+	select {
+	case q = <-p.wl:
+		return q
+	case <-time.After(2 * time.Second):
+		close(p.wl)
 	}
-	return fmt.Sprintf("%x", h.Sum(nil)), nil
+	return q
 }
 
-// parseHeader parses everything that begin with #
-func (q *Query) parseHeader(line string) {
+func (q *Query) parseMariaDBHeader(line string) {
 	var err error
 	parts := strings.Split(line, " ")
 
 	for idx, part := range parts {
 		part = strings.ToLower(part)
 
 		if strings.Contains(part, "query_time:") {
-			q.QueryTime = parts[idx+1]
+			time := parts[idx+1]
+			q.QueryTime, err = strconv.ParseFloat(time, 64)
+			if err != nil {
+				logrus.Errorf("query_time: error converting %s to time: %s", parts[idx+1], err)
+			}
+
 		} else if strings.Contains(part, "lock_time:") {
-			q.LockTime = parts[idx+1]
+			time := parts[idx+1]
+			q.LockTime, err = strconv.ParseFloat(time, 64)
+			if err != nil {
+				logrus.Errorf("lock_time: error converting %s to time: %s", parts[idx+1], err)
+			}
+
 		} else if strings.Contains(part, "time:") {
-			q.Time = parts[idx+1]
+			date := parts[idx+1] + " " + parts[idx+2]
+			fmt.Printf("\n\n\n%s\n\n\n\n", date)
+			q.Time, err = time.Parse("060102 15:04:05", date)
+			if err != nil {
+				logrus.Errorf("time: error converting %s to time: %s", parts[idx+1], err)
+			}
+
 		} else if strings.Contains(part, "rows_sent:") {
 			q.RowsSent, err = strconv.Atoi(parts[idx+1])
 			if err != nil {
 				logrus.Errorf("row_sent: error converting %s to int: %s", parts[idx+1], err)
 			}
+
 		} else if strings.Contains(part, "rows_examined:") {
 			q.RowsExamined, err = strconv.Atoi(parts[idx+1])
 			if err != nil {
 				logrus.Errorf("rows_examined: error converting %s to int: %s", parts[idx+1], err)
 			}
+
 		} else if strings.Contains(part, "rows_affected:") {
 			q.RowsAffected, err = strconv.Atoi(parts[idx+1])
 			if err != nil {
 				logrus.Errorf("rows_affected: error converting %s to int: %s", parts[idx+1], err)
 			}
+
 		} else if strings.Contains(part, "id:") {
 			// Some IDs can have multiple spaces, so we try to bruteforce the
 			// number of spaces. I tried implementing a version that keeps in
@@ -109,24 +106,29 @@ func (q *Query) parseHeader(line string) {
 			if err != nil {
 				logrus.Errorf("id: error converting %s to int: %s", parts[idx+1], err)
 			}
+
 		} else if strings.Contains(part, "user@host:") {
-			items := re.FindAllString(line, -1)
+			items := stringInBrackets.FindAllString(line, -1)
 			// We remove first and last bytes of the strings because they are
 			// square brackets
 			q.User = items[0][1 : len(items[0])-1]
 			q.Host = items[1][1 : len(items[1])-1]
+
 		} else if strings.Contains(part, "schema:") {
 			q.Schema = parts[idx+1]
+
 		} else if strings.Contains(part, "last_errno:") {
 			q.LastErrNo, err = strconv.Atoi(parts[idx+1])
 			if err != nil {
 				logrus.Errorf("last_errno: error converting %s to int: %s", parts[idx+1], err)
 			}
+
 		} else if strings.Contains(part, "killed:") {
 			q.Killed, err = strconv.Atoi(parts[idx+1])
 			if err != nil {
 				logrus.Errorf("killed: error converting %s to int: %s", parts[idx+1], err)
 			}
+
 		} else if strings.Contains(part, "bytes_sent:") {
 			q.BytesSent, err = strconv.Atoi(parts[idx+1])
 			if err != nil {