Added TODOs and the parser function that populates event.CommentMetadata

Author: yashmaheshwari

event.go

@@ -12,18 +12,19 @@ package slowlog
 // event is expected to define the query and Query_time metric. Other metrics
 // and metadata vary according to MySQL version, distro, and configuration.
 type Event struct {
-	Offset        uint64 // byte offset in file at which event starts
-	Ts            string // raw timestamp of event
-	Admin         bool   // true if Query is admin command
-	Query         string // SQL query or admin command
-	User          string
-	Host          string
-	Db            string
-	TimeMetrics   map[string]float64 // *_time and *_wait metrics
-	NumberMetrics map[string]uint64  // most metrics
-	BoolMetrics   map[string]bool    // yes/no metrics
-	RateType      string             // Percona Server rate limit type
-	RateLimit     uint               // Percona Server rate limit value
+	Offset          uint64 // byte offset in file at which event starts
+	Ts              string // raw timestamp of event
+	Admin           bool   // true if Query is admin command
+	Query           string // SQL query or admin command
+	User            string
+	Host            string
+	Db              string
+	TimeMetrics     map[string]float64 // *_time and *_wait metrics
+	NumberMetrics   map[string]uint64  // most metrics
+	BoolMetrics     map[string]bool    // yes/no metrics
+	RateType        string             // Percona Server rate limit type
+	RateLimit       uint               // Percona Server rate limit value
+	CommentMetadata map[string]string
 }
 
 // NewEvent returns a new Event with initialized metric maps.

metrics.go

@@ -14,6 +14,7 @@ type Metrics struct {
 	TimeMetrics   map[string]*TimeStats   `json:",omitempty"`
 	NumberMetrics map[string]*NumberStats `json:",omitempty"`
 	BoolMetrics   map[string]*BoolStats   `json:",omitempty"`
+	//	TODO(Yash): Add another metrics for metadata.
 }
 
 // TimeStats are microsecond-based metrics like Query_time and Lock_time.
@@ -26,6 +27,8 @@ type TimeStats struct {
 	P95        float64 `json:",omitempty"` // 95th percentile
 	Max        float64 `json:",omitempty"`
 	outlierSum float64
+	//	TODO(Yash) timeTakeToTraceId as map.  Move this to Metrics class or create another struct parallel to TImeMetrics, NumberMetrics, BoolMetrics
+	//	traceIdOfMaxSql
 }
 
 // NumberStats are integer-based metrics like Rows_sent and Merge_passes.
@@ -72,6 +75,7 @@ func (m *Metrics) AddEvent(e Event, outlier bool) {
 			stats.Sum += val
 		}
 		stats.vals = append(stats.vals, float64(val))
+		//	TODO(Yash): Add traceIdProto to the map defined in TimeMetrics
 	}
 
 	for metric, val := range e.NumberMetrics {
@@ -133,6 +137,7 @@ func (m *Metrics) Finalize(rateLimit uint) {
 
 		// Update sum last because avg ^ needs the original value.
 		s.Sum = (s.Sum * float64(rateLimit)) + s.outlierSum
+		//TODO(Yash): Do timeTakenToTraceIdProto[s.Max] to know the traceId and populate traceIdOfMaxSql
 	}
 
 	for _, s := range m.NumberMetrics {

parser.go

@@ -9,6 +9,7 @@ package slowlog
 
 import (
 	"bufio"
+	"bytes"
 	"errors"
 	"fmt"
 	"io"
@@ -51,6 +52,7 @@ var metricsRe = regexp.MustCompile(`(\w+): (\S+|\z)`)
 var adminRe = regexp.MustCompile(`command: (.+)`)
 var setRe = regexp.MustCompile(`^SET (?:last_insert_id|insert_id|timestamp)`)
 var useRe = regexp.MustCompile(`^(?i)use `)
+var blockCommentsRe = regexp.MustCompile("/\\*[\\w\\W]*?(?=\\*/)\\*/")
 
 // FileParser represents a file-based Parser. This is the canonical Parser
 // because the slow log is a file.
@@ -396,6 +398,92 @@ func (p *FileParser) parseAdmin(line string) {
 	}
 }
 
+// Tested here:  https://play.golang.org/p/u7XZnxu2LLL
+func parseComments(queryWithComments string) map[string]string {
+	startIndex := strings.Index(queryWithComments, "/*")
+	if startIndex != -1 {
+		endIndex := strings.LastIndex(queryWithComments, "*/")
+		sqlComments := queryWithComments[startIndex:endIndex]
+		return parseKeyValuePairToMap(sqlComments)
+	}
+	return map[string]string{}
+}
+
+// Copied verbatim from https://gist.github.com/alexisvisco/4b846978c9346e4eaf618bb632c0693a
+func parseKeyValuePairToMap(msg string) map[string]string {
+
+	type kv struct {
+		key, val string
+	}
+
+	var pair *kv = nil
+	pairs := make(map[string]string)
+	buf := bytes.NewBuffer([]byte{})
+
+	var (
+		escape  = false
+		garbage = false
+		quoted  = false
+	)
+
+	completePair := func(buffer *bytes.Buffer, pair *kv) kv {
+		if pair != nil {
+			return kv{pair.key, buffer.String()}
+		} else {
+			return kv{buffer.String(), ""}
+		}
+	}
+
+	for _, c := range msg {
+		if !quoted && c == ' ' {
+			if buf.Len() != 0 {
+				if !garbage {
+					p := completePair(buf, pair)
+					pairs[p.key] = p.val
+					pair = nil
+				}
+				buf.Reset()
+			}
+			garbage = false
+		} else if !quoted && c == '=' {
+			if buf.Len() != 0 {
+				pair = &kv{key: buf.String(), val: ""}
+				buf.Reset()
+			} else {
+				garbage = true
+			}
+		} else if quoted && c == '\\' {
+			escape = true
+		} else if c == '"' {
+			if escape {
+				buf.WriteRune(c)
+				escape = false
+			} else {
+				quoted = !quoted
+			}
+		} else {
+			if escape {
+				buf.WriteRune('\\')
+				escape = false
+			}
+			buf.WriteRune(c)
+		}
+	}
+
+	if !garbage {
+		p := completePair(buf, pair)
+		pairs[p.key] = p.val
+	}
+
+	// Remove any key value pair where either of them is a blank string
+	for key, value := range pairs {
+		if key == "" || value == "" {
+			delete(pairs, key)
+		}
+	}
+	return pairs
+}
+
 func (p *FileParser) sendEvent(inHeader bool, inQuery bool) {
 	if Debug {
 		log.Println("send event")
@@ -421,6 +509,7 @@ func (p *FileParser) sendEvent(inHeader bool, inQuery bool) {
 	// Clean up the event.
 	p.event.Db = strings.TrimSuffix(p.event.Db, ";\n")
 	p.event.Query = strings.TrimSuffix(p.event.Query, ";")
+	p.event.CommentMetadata = parseComments(p.event.Query)
 
 	// Send the event.  This will block.
 	select {