internal/devtools/cmd/labelhist: show label history

Add a CLI tool for displaying the the labels added to and removed
from an issue.

This will be useful to evaluate labels added by gaby.

Example:

> go run ./internal/devtools/cmd/labelhist 69000
69000:
  2024-08-21T19:56:28Z cherrymui  +NeedsInvestigation
  2024-08-22T19:03:13Z dmitshur   -NeedsInvestigation
  2024-08-22T19:03:13Z dmitshur   +NeedsFix
  2024-08-22T19:03:30Z dmitshur   +Testing

For #64.

Change-Id: Ide5f429df92bfebd7978a8dfbb46b93708a6ab3d
Reviewed-on: https://go-review.googlesource.com/c/oscar/+/635876
LUCI-TryBot-Result: Go LUCI <[email protected]>
Reviewed-by: Tatiana Bradley <[email protected]>

internal/devtools/cmd/labelhist/main.go

@@ -0,0 +1,101 @@
+// Copyright 2024 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+/*
+Labelhist displays the events of a GitHub issue that affect its labels.
+
+Usage:
+
+	labelhist issues...
+
+Each argument can be a single issue number or a range of numbers "from-to".
+*/
+package main
+
+import (
+	"context"
+	"flag"
+	"fmt"
+	"log"
+	"log/slog"
+	"os"
+	"strconv"
+	"strings"
+
+	"golang.org/x/oscar/internal/gcp/firestore"
+	"golang.org/x/oscar/internal/github"
+)
+
+var project = flag.String("project", "golang/go", "GitHub project")
+
+func usage() {
+	fmt.Fprintf(os.Stderr, "usage: labelhist issues...\n")
+	flag.PrintDefaults()
+	os.Exit(2)
+}
+
+func main() {
+	log.SetFlags(0)
+	log.SetPrefix("labelhist: ")
+	flag.Usage = usage
+	flag.Parse()
+	if err := run(context.Background()); err != nil {
+		log.Fatal(err)
+	}
+}
+
+func run(ctx context.Context) error {
+	var ranges []Range
+	for _, arg := range flag.Args() {
+		r, err := parseIssueArg(arg)
+		if err != nil {
+			return err
+		}
+		ranges = append(ranges, r)
+	}
+	lg := slog.New(slog.NewTextHandler(os.Stderr, nil))
+	db, err := firestore.NewDB(ctx, lg, "oscar-go-1", "prod")
+	if err != nil {
+		return err
+	}
+
+	for _, r := range ranges {
+		for ev := range github.Events(db, *project, r.min, r.max) {
+			switch ev.API {
+			case "/issues":
+				fmt.Printf("%d:\n", ev.Issue)
+			case "/issues/events":
+				ie := ev.Typed.(*github.IssueEvent)
+				if ie.Event == "labeled" || ie.Event == "unlabeled" {
+					c := '+'
+					if ie.Event == "unlabeled" {
+						c = '-'
+					}
+					fmt.Printf("  %s %-10s %c%s\n", ie.CreatedAt, ie.Actor.Login, c, ie.Label.Name)
+				}
+			}
+		}
+	}
+	return nil
+}
+
+type Range struct {
+	min, max int64
+}
+
+func parseIssueArg(s string) (Range, error) {
+	sfrom, sto, found := strings.Cut(s, "-")
+	from, err := strconv.ParseInt(sfrom, 10, 64)
+	if err != nil {
+		return Range{}, err
+	}
+	if !found {
+		return Range{from, from}, nil
+	}
+	to, err := strconv.ParseInt(sto, 10, 64)
+	if err != nil {
+		return Range{}, err
+	}
+	return Range{from, to}, nil
+}

internal/github/data.go

@@ -62,7 +62,7 @@ func LookupIssue(db storage.DB, project string, issue int64) (*Issue, error) {
 // only consulting the database (not actual GitHub).
 func LookupIssues(db storage.DB, project string, issueMin, issueMax int64) iter.Seq[*Issue] {
 	return func(yield func(*Issue) bool) {
-		for e := range events(db, project, issueMin, issueMax) {
+		for e := range Events(db, project, issueMin, issueMax) {
 			if e.API == "/issues" {
 				if !yield(e.Typed.(*Issue)) {
 					break
@@ -107,18 +107,19 @@ func CleanBody(body string) string {
 	return body
 }
 
+// Events calls [Events] with the client's db.
+func (c *Client) Events(project string, issueMin, issueMax int64) iter.Seq[*Event] {
+	return Events(c.db, project, issueMin, issueMax)
+}
+
 // Events returns an iterator over issue events for the given project,
 // limited to issues in the range issueMin ≤ issue ≤ issueMax.
 // If issueMax < 0, there is no upper limit.
 // The events are iterated over in (Project, Issue, API, ID) order,
 // so "/issues" events come first, then "/issues/comments", then "/issues/events".
 // Within a specific API, the events are ordered by increasing ID,
 // which corresponds to increasing event time on GitHub.
-func (c *Client) Events(project string, issueMin, issueMax int64) iter.Seq[*Event] {
-	return events(c.db, project, issueMin, issueMax)
-}
-
-func events(db storage.DB, project string, issueMin, issueMax int64) iter.Seq[*Event] {
+func Events(db storage.DB, project string, issueMin, issueMax int64) iter.Seq[*Event] {
 	return func(yield func(*Event) bool) {
 		start := o(project, issueMin)
 		if issueMax < 0 {
@@ -206,6 +207,7 @@ type IssueEvent struct {
 	URL        string    `json:"url"`
 	Actor      User      `json:"actor"`
 	Event      string    `json:"event"`
+	Label      Label     `json:"label"` // for "labeled" and "unlabeled" events
 	Labels     []Label   `json:"labels"`
 	LockReason string    `json:"lock_reason"`
 	CreatedAt  string    `json:"created_at"`