feat: format bulk commit from other sources in release notes (#2665)

Fixes #2599

Author: JoeWang1127

internal/config/state.go

@@ -26,6 +26,8 @@ import (
 const (
 	StatusNew      = "new"
 	StatusExisting = "existing"
+	// BulkChangeThreshold is a threshold to determine whether a commit is a bulk change.
+	BulkChangeThreshold = 10
 )
 
 // LibrarianState defines the contract for the state.yaml file.
@@ -161,7 +163,7 @@ type Commit struct {
 // IsBulkCommit returns true if the commit is associated with 10 or more
 // libraries.
 func (c *Commit) IsBulkCommit() bool {
-	return len(strings.Split(c.LibraryIDs, ",")) >= 10
+	return len(strings.Split(c.LibraryIDs, ",")) >= BulkChangeThreshold
 }
 
 var (

internal/librarian/release_init.go

@@ -277,7 +277,7 @@ func (r *initRunner) updateLibrary(library *config.LibraryState, commits []*gitr
 
 	// Update the previous version, we need this value when creating release note.
 	library.PreviousVersion = library.Version
-	library.Changes = toCommit(commits)
+	library.Changes = toCommit(commits, library.ID)
 	library.Version = nextVersion
 	library.ReleaseTriggered = true
 	return nil
@@ -310,16 +310,24 @@ func (r *initRunner) determineNextVersion(commits []*gitrepo.ConventionalCommit,
 // toCommit converts a slice of gitrepo.ConventionalCommit to a slice of config.Commit.
 // If the ConventionalCommit has NestedCommits, they are also extracted and
 // converted.
-func toCommit(c []*gitrepo.ConventionalCommit) []*config.Commit {
+// Set LibraryIDs to the given libraryID if the conventional commit doesn't have key `Library-IDs` in the Footers;
+// otherwise use the value in the Footers as LibraryIDs.
+func toCommit(c []*gitrepo.ConventionalCommit, libraryID string) []*config.Commit {
 	var commits []*config.Commit
 	for _, cc := range c {
+		var libraryIDs string
+		libraryIDs, ok := cc.Footers["Library-IDs"]
+		if !ok {
+			libraryIDs = libraryID
+		}
+
 		commits = append(commits, &config.Commit{
 			Type:          cc.Type,
 			Subject:       cc.Subject,
 			Body:          cc.Body,
 			CommitHash:    cc.CommitHash,
 			PiperCLNumber: cc.Footers["PiperOrigin-RevId"],
-			LibraryIDs:    cc.Footers["Library-IDs"],
+			LibraryIDs:    libraryIDs,
 		})
 	}
 	return commits

internal/librarian/release_init_test.go

@@ -1246,6 +1246,7 @@ func TestRunInitCommand(t *testing.T) {
 								Type:       "feat",
 								Subject:    "bump version",
 								CommitHash: "1234560000000000000000000000000000000000",
+								LibraryIDs: "another-example-id",
 							},
 						},
 						ReleaseTriggered: true,
@@ -1266,6 +1267,7 @@ func TestRunInitCommand(t *testing.T) {
 								Type:       "feat",
 								Subject:    "bump version",
 								CommitHash: "1234560000000000000000000000000000000000",
+								LibraryIDs: "example-id",
 							},
 						},
 						ReleaseTriggered: true,
@@ -1556,14 +1558,16 @@ func TestUpdateLibrary(t *testing.T) {
 				PreviousVersion: "1.2.3",
 				Changes: []*config.Commit{
 					{
-						Type:    "fix",
-						Subject: "change a typo",
+						Type:       "fix",
+						Subject:    "change a typo",
+						LibraryIDs: "one-id",
 					},
 					{
 						Type:          "feat",
 						Subject:       "add a config file",
 						Body:          "This is the body.",
 						PiperCLNumber: "12345",
+						LibraryIDs:    "one-id",
 					},
 				},
 				ReleaseTriggered: true,
@@ -1594,14 +1598,16 @@ func TestUpdateLibrary(t *testing.T) {
 				PreviousVersion: "1.2.3",
 				Changes: []*config.Commit{
 					{
-						Type:    "fix",
-						Subject: "change a typo",
+						Type:       "fix",
+						Subject:    "change a typo",
+						LibraryIDs: "one-id",
 					},
 					{
 						Type:          "feat",
 						Subject:       "add a config file",
 						Body:          "This is the body.",
 						PiperCLNumber: "12345",
+						LibraryIDs:    "one-id",
 					},
 				},
 				ReleaseTriggered: true,
@@ -1686,13 +1692,15 @@ func TestUpdateLibrary(t *testing.T) {
 				PreviousVersion: "1.2.3",
 				Changes: []*config.Commit{
 					{
-						Type:    "feat",
-						Subject: "add another config file",
-						Body:    "This is the body",
+						Type:       "feat",
+						Subject:    "add another config file",
+						Body:       "This is the body",
+						LibraryIDs: "one-id",
 					},
 					{
-						Type:    "feat",
-						Subject: "change a typo",
+						Type:       "feat",
+						Subject:    "change a typo",
+						LibraryIDs: "one-id",
 					},
 				},
 				ReleaseTriggered: true,
@@ -1746,8 +1754,9 @@ func TestUpdateLibrary(t *testing.T) {
 				ReleaseTriggered: true,
 				Changes: []*config.Commit{
 					{
-						Type:    "chore",
-						Subject: "a chore",
+						Type:       "chore",
+						Subject:    "a chore",
+						LibraryIDs: "one-id",
 					},
 				},
 			},

internal/librarian/release_notes.go

@@ -176,24 +176,22 @@ type commitSection struct {
 // formatReleaseNotes generates the body for a release pull request.
 func formatReleaseNotes(state *config.LibrarianState, ghRepo *github.Repository) (string, error) {
 	librarianVersion := cli.Version()
+	// Separate commits to bulk changes (affects multiple libraries) or library-specific changes because they
+	// appear in different section in the release notes.
+	bulkChangesMap, libraryChanges := separateCommits(state)
+	// Process library specific changes.
 	var releaseSections []*releaseNoteSection
-	// create a map to deduplicate bulk changes based on their commit hash
-	// and subject
-	bulkChangesMap := make(map[string]*config.Commit)
 	for _, library := range state.Libraries {
 		if !library.ReleaseTriggered {
 			continue
 		}
-
-		for _, commit := range library.Changes {
-			if commit.IsBulkCommit() {
-				bulkChangesMap[commit.CommitHash+commit.Subject] = commit
-			}
-		}
-
-		section := formatLibraryReleaseNotes(library)
+		// No need to check the existence of the key, library.ID, because a library without library-specific changes
+		// may appear in the release notes, i.e., in the bulk changes section.
+		commits := libraryChanges[library.ID]
+		section := formatLibraryReleaseNotes(library, commits)
 		releaseSections = append(releaseSections, section)
 	}
+	// Process bulk changes
 	var bulkChanges []*config.Commit
 	for _, commit := range bulkChangesMap {
 		bulkChanges = append(bulkChanges, commit)
@@ -222,18 +220,19 @@ func formatReleaseNotes(state *config.LibrarianState, ghRepo *github.Repository)
 
 // formatLibraryReleaseNotes generates release notes in Markdown format for a single library.
 // It returns the generated release notes and the new version string.
-func formatLibraryReleaseNotes(library *config.LibraryState) *releaseNoteSection {
+func formatLibraryReleaseNotes(library *config.LibraryState, commits []*config.Commit) *releaseNoteSection {
 	// The version should already be updated to the next version.
 	newVersion := library.Version
 	tagFormat := config.DetermineTagFormat(library.ID, library, nil)
 	newTag := config.FormatTag(tagFormat, library.ID, newVersion)
 	previousTag := config.FormatTag(tagFormat, library.ID, library.PreviousVersion)
 
+	sort.Slice(commits, func(i, j int) bool {
+		return commits[i].CommitHash < commits[j].CommitHash
+	})
 	commitsByType := make(map[string][]*config.Commit)
-	for _, commit := range library.Changes {
-		if !commit.IsBulkCommit() {
-			commitsByType[commit.Type] = append(commitsByType[commit.Type], commit)
-		}
+	for _, commit := range commits {
+		commitsByType[commit.Type] = append(commitsByType[commit.Type], commit)
 	}
 
 	var sections []*commitSection
@@ -259,3 +258,78 @@ func formatLibraryReleaseNotes(library *config.LibraryState) *releaseNoteSection
 
 	return section
 }
+
+// separateCommits analyzes all commits associated with triggered releases in the
+// given state and categorizes them into two groups:
+//
+// 1. Bulk Changes: Commits that affect multiple libraries. This includes:
+//   - Commits identified by IsBulkCommit() (e.g., librarian generation PRs).
+//   - Commits that appear in multiple libraries' change sets but are not
+//     marked as bulk commits (e.g., dependency updates, README changes).
+//     The Library-IDs for these are concatenated.
+//
+// 2. Library Changes: Commits that are unique to a single library.
+//
+// It returns two maps:
+//   - The first map contains bulk changes, keyed by a composite of commit hash and subject.
+//   - The second map contains library-specific changes, keyed by LibraryID.
+func separateCommits(state *config.LibrarianState) (map[string]*config.Commit, map[string][]*config.Commit) {
+	maybeBulkChanges := make(map[string][]*config.Commit)
+	for _, library := range state.Libraries {
+		if !library.ReleaseTriggered {
+			continue
+		}
+
+		for _, commit := range library.Changes {
+			key := commit.CommitHash + commit.Subject
+			maybeBulkChanges[key] = append(maybeBulkChanges[key], commit)
+		}
+	}
+
+	bulkChanges := make(map[string]*config.Commit)
+	libraryChanges := make(map[string][]*config.Commit)
+	for key, commits := range maybeBulkChanges {
+		// A commit has multiple library IDs in the footer, this should come from librarian generation PR.
+		// All commits should be identical.
+		if commits[0].IsBulkCommit() {
+			bulkChanges[key] = commits[0]
+			continue
+		}
+		// More than ten commits have the same commit subject and sha, this should come from other sources,
+		// e.g., dependency updates, README updates, etc.
+		// All commits should be identical except for the library id.
+		// We assume this type of commits has only one library id in Footers and each id is unique among all
+		// commits.
+		if len(commits) >= config.BulkChangeThreshold {
+			bulkChanges[key] = concatenateLibraryIDs(commits)
+			continue
+		}
+		// We assume the rest of commits are library-specific.
+		for _, commit := range commits {
+			// Non-bulk commits may have 1 - 9 library IDs.
+			libraryIDs := strings.Split(commit.LibraryIDs, ",")
+			for _, libraryID := range libraryIDs {
+				if libraryID == "" {
+					continue
+				}
+				libraryChanges[libraryID] = append(libraryChanges[libraryID], commit)
+			}
+		}
+	}
+
+	return bulkChanges, libraryChanges
+}
+
+// concatenateLibraryIDs merges the LibraryIDs from a slice of commits into the first commit.
+func concatenateLibraryIDs(commits []*config.Commit) *config.Commit {
+	var libraryIDs []string
+	for _, commit := range commits {
+		libraryIDs = append(libraryIDs, commit.LibraryIDs)
+	}
+
+	sort.Slice(libraryIDs, func(i, j int) bool {
+		return libraryIDs[i] < libraryIDs[j]
+	})
+	commits[0].LibraryIDs = strings.Join(libraryIDs, ",")
+	return commits[0]
+}