Files
forge/services/repository/generate.go
zeekayandHanzo Dev aa87442fdd Record upstream authorship centrally in NOTICE instead of in every file header
Source files keep their SPDX-License-Identifier and Hanzo copyright line; the
Gitea, Gogs and LFS Test Server copyrights they used to repeat now live once in
NOTICE, alongside LICENSE. The header lint accepts any copyright line above the
SPDX line, so it no longer demands an upstream one.

Co-authored-by: Hanzo Dev <[email protected]>
2026-09-10 17:50:35 -07:00

322 lines
11 KiB
Go

// Copyright 2026 Hanzo AI, Inc. All rights reserved.
// SPDX-License-Identifier: MIT
package repository
import (
"bufio"
"bytes"
"context"
"errors"
"fmt"
"io/fs"
"os"
"path/filepath"
"regexp"
"strconv"
"strings"
"sync"
"time"
git_model "github.com/hanzoai/git/models/git"
repo_model "github.com/hanzoai/git/models/repo"
"github.com/hanzoai/git/modules/git"
"github.com/hanzoai/git/modules/gitrepo"
"github.com/hanzoai/git/modules/glob"
"github.com/hanzoai/git/modules/log"
repo_module "github.com/hanzoai/git/modules/repository"
"github.com/hanzoai/git/modules/setting"
"github.com/hanzoai/git/modules/util"
"github.com/huandu/xstrings"
)
type transformer struct {
Name string
Transform func(string) string
}
type expansion struct {
Name string
Value string
Transformers []transformer
}
var globalVars = sync.OnceValue(func() (ret struct {
defaultTransformers []transformer
fileNameSanitizeRegexp *regexp.Regexp
},
) {
ret.defaultTransformers = []transformer{
{Name: "SNAKE", Transform: xstrings.ToSnakeCase},
{Name: "KEBAB", Transform: xstrings.ToKebabCase},
{Name: "CAMEL", Transform: xstrings.ToCamelCase},
{Name: "PASCAL", Transform: xstrings.ToPascalCase},
{Name: "LOWER", Transform: strings.ToLower},
{Name: "UPPER", Transform: strings.ToUpper},
{Name: "TITLE", Transform: util.ToTitleCase},
}
// invalid filename contents, based on https://github.com/sindresorhus/filename-reserved-regex
// "COM10" needs to be opened with UNC "\\.\COM10" on Windows, so itself is valid
ret.fileNameSanitizeRegexp = regexp.MustCompile(`(?i)[<>:"/\\|?*\x{0000}-\x{001F}]|^(con|prn|aux|nul|com\d|lpt\d)$`)
return ret
})
func generateExpansion(ctx context.Context, src string, templateRepo, generateRepo *repo_model.Repository) string {
transformers := globalVars().defaultTransformers
year, month, day := time.Now().Date()
expansions := []expansion{
{Name: "YEAR", Value: strconv.Itoa(year), Transformers: nil},
{Name: "MONTH", Value: fmt.Sprintf("%02d", int(month)), Transformers: nil},
{Name: "MONTH_ENGLISH", Value: month.String(), Transformers: transformers},
{Name: "DAY", Value: fmt.Sprintf("%02d", day), Transformers: nil},
{Name: "REPO_NAME", Value: generateRepo.Name, Transformers: transformers},
{Name: "TEMPLATE_NAME", Value: templateRepo.Name, Transformers: transformers},
{Name: "REPO_DESCRIPTION", Value: generateRepo.Description, Transformers: nil},
{Name: "TEMPLATE_DESCRIPTION", Value: templateRepo.Description, Transformers: nil},
{Name: "REPO_OWNER", Value: generateRepo.OwnerName, Transformers: transformers},
{Name: "TEMPLATE_OWNER", Value: templateRepo.OwnerName, Transformers: transformers},
{Name: "REPO_LINK", Value: generateRepo.Link(), Transformers: nil},
{Name: "TEMPLATE_LINK", Value: templateRepo.Link(), Transformers: nil},
{Name: "REPO_HTTPS_URL", Value: generateRepo.CloneLinkGeneral(ctx).HTTPS, Transformers: nil},
{Name: "TEMPLATE_HTTPS_URL", Value: templateRepo.CloneLinkGeneral(ctx).HTTPS, Transformers: nil},
{Name: "REPO_SSH_URL", Value: generateRepo.CloneLinkGeneral(ctx).SSH, Transformers: nil},
{Name: "TEMPLATE_SSH_URL", Value: templateRepo.CloneLinkGeneral(ctx).SSH, Transformers: nil},
}
expansionMap := make(map[string]string)
for _, e := range expansions {
expansionMap[e.Name] = e.Value
for _, tr := range e.Transformers {
expansionMap[fmt.Sprintf("%s_%s", e.Name, tr.Name)] = tr.Transform(e.Value)
}
}
return os.Expand(src, func(key string) string {
if val, ok := expansionMap[key]; ok {
return val
}
return key
})
}
// gitTemplateFileMatcher holds information about a .hanzo/template file
type gitTemplateFileMatcher struct {
relPath string
globs []glob.Glob
}
func newGitTemplateFileMatcher(relPath string, content []byte) *gitTemplateFileMatcher {
gt := &gitTemplateFileMatcher{relPath: relPath}
gt.globs = make([]glob.Glob, 0)
scanner := bufio.NewScanner(bytes.NewReader(content))
for scanner.Scan() {
line := strings.TrimSpace(scanner.Text())
if line == "" || strings.HasPrefix(line, "#") {
continue
}
g, err := glob.Compile(line, '/')
if err != nil {
log.Debug("Invalid glob expression '%s' (skipped): %v", line, err)
continue
}
gt.globs = append(gt.globs, g)
}
return gt
}
func (gt *gitTemplateFileMatcher) HasRules() bool {
return len(gt.globs) != 0
}
func (gt *gitTemplateFileMatcher) Match(s string) bool {
for _, g := range gt.globs {
if g.Match(s) {
return true
}
}
return false
}
func readGitTemplateFile(tmpDir string) (*gitTemplateFileMatcher, error) {
templateRelPath := filepath.Join(".hanzo", "template")
content, err := util.ReadRegularPathFile(tmpDir, templateRelPath, 1024*1024)
if err != nil {
return nil, util.Iif(errors.Is(err, util.ErrNotRegularPathFile), os.ErrNotExist, err)
}
return newGitTemplateFileMatcher(templateRelPath, content), nil
}
func substGitTemplateFile(ctx context.Context, tmpDir, tmpDirSubPath string, templateRepo, generateRepo *repo_model.Repository) error {
content, err := util.ReadRegularPathFile(tmpDir, tmpDirSubPath, 1024*1024)
if err != nil {
if errors.Is(err, fs.ErrNotExist) {
return nil
}
return err
}
if err := os.Remove(util.FilePathJoinAbs(tmpDir, tmpDirSubPath)); err != nil {
return err
}
generatedContent := generateExpansion(ctx, string(content), templateRepo, generateRepo)
substSubPath := filePathSanitize(generateExpansion(ctx, tmpDirSubPath, templateRepo, generateRepo))
return util.WriteRegularPathFile(tmpDir, substSubPath, []byte(generatedContent), 0o755, 0o644)
}
// processGitTemplateFile processes and removes the .hanzo/template file, does variable expansion for template files
// and save the processed files to the filesystem. It returns a list of skipped files that are not regular paths.
func processGitTemplateFile(ctx context.Context, tmpDir string, templateRepo, generateRepo *repo_model.Repository, fileMatcher *gitTemplateFileMatcher) (skippedFiles []string, _ error) {
// Why not use "os.Root" here: symlink is unsafe even in the same root but "os.Root" can't help, it's more difficult to use "os.Root" to do the WalkDir.
if err := os.Remove(util.FilePathJoinAbs(tmpDir, fileMatcher.relPath)); err != nil {
return nil, fmt.Errorf("unable to remove .hanzo/template: %w", err)
}
if !fileMatcher.HasRules() {
return skippedFiles, nil // Avoid walking tree if there are no globs
}
err := filepath.WalkDir(tmpDir, func(fullPath string, d os.DirEntry, walkErr error) error {
if walkErr != nil {
return walkErr
}
if d.IsDir() {
return nil
}
tmpDirSubPath, err := filepath.Rel(tmpDir, fullPath)
if err != nil {
return err
}
if fileMatcher.Match(filepath.ToSlash(tmpDirSubPath)) {
err := substGitTemplateFile(ctx, tmpDir, tmpDirSubPath, templateRepo, generateRepo)
if errors.Is(err, util.ErrNotRegularPathFile) {
skippedFiles = append(skippedFiles, tmpDirSubPath)
} else if err != nil {
return err
}
}
return nil
}) // end: WalkDir
if err != nil {
return nil, err
}
if err = util.RemoveAll(util.FilePathJoinAbs(tmpDir, ".git")); err != nil {
return nil, err
}
return skippedFiles, nil
}
func generateRepoCommit(ctx context.Context, repo, templateRepo, generateRepo *repo_model.Repository, tmpDir string) error {
// set default branch based on whether it's specified in the newly generated repo or not
repo.DefaultBranch = util.IfZero(repo.DefaultBranch, util.IfZero(templateRepo.DefaultBranch, setting.Repository.DefaultBranch))
// Clone to temporary path and do the init commit.
if err := gitrepo.CloneRepoToLocal(ctx, templateRepo, tmpDir, git.CloneRepoOptions{
Depth: 1,
Branch: templateRepo.DefaultBranch,
}); err != nil {
return fmt.Errorf("git clone: %w", err)
}
// Get active submodules from the template
submodules, err := git.GetTemplateSubmoduleCommits(ctx, tmpDir)
if err != nil {
return fmt.Errorf("GetTemplateSubmoduleCommits: %w", err)
}
if err = util.RemoveAll(filepath.Join(tmpDir, ".git")); err != nil {
return fmt.Errorf("remove git dir: %w", err)
}
// Variable expansion
fileMatcher, err := readGitTemplateFile(tmpDir)
if err == nil {
_, err = processGitTemplateFile(ctx, tmpDir, templateRepo, generateRepo, fileMatcher)
if err != nil {
return fmt.Errorf("processGitTemplateFile: %w", err)
}
} else if errors.Is(err, fs.ErrNotExist) {
log.Debug("skip processing repo template files: no available .hanzo/template")
} else {
return fmt.Errorf("readGitTemplateFile: %w", err)
}
if err = git.InitRepository(ctx, tmpDir, false, templateRepo.ObjectFormatName); err != nil {
return err
}
if err = git.AddTemplateSubmoduleIndexes(ctx, tmpDir, submodules); err != nil {
return fmt.Errorf("failed to add submodules: %v", err)
}
return initRepoCommit(ctx, tmpDir, repo, repo.Owner)
}
// GenerateGitContent generates git content from a template repository
func GenerateGitContent(ctx context.Context, templateRepo, generateRepo *repo_model.Repository) (err error) {
tmpDir, cleanup, err := setting.AppDataTempDir("git-repo-content").MkdirTempRandom("repos-" + generateRepo.Name)
if err != nil {
return fmt.Errorf("failed to create temp dir for repository %s: %w", generateRepo.FullName(), err)
}
defer cleanup()
if err = generateRepoCommit(ctx, generateRepo, templateRepo, generateRepo, tmpDir); err != nil {
return fmt.Errorf("generateRepoCommit: %w", err)
}
if err = gitrepo.SetDefaultBranch(ctx, generateRepo, generateRepo.DefaultBranch); err != nil {
return fmt.Errorf("setDefaultBranch: %w", err)
}
if err = repo_model.UpdateRepositoryColsNoAutoTime(ctx, generateRepo, "default_branch"); err != nil {
return fmt.Errorf("updateRepository: %w", err)
}
if err := repo_module.UpdateRepoSize(ctx, generateRepo); err != nil {
return fmt.Errorf("failed to update size for repository: %w", err)
}
if err := git_model.CopyLFS(ctx, generateRepo, templateRepo); err != nil {
return fmt.Errorf("failed to copy LFS: %w", err)
}
if _, err := repo_module.SyncRepoBranches(ctx, generateRepo.ID, 0); err != nil {
return fmt.Errorf("SyncRepoBranches: %w", err)
}
return nil
}
// GenerateRepoOptions contains the template units to generate
type GenerateRepoOptions struct {
Name string
DefaultBranch string
Description string
Private bool
GitContent bool
Topics bool
GitHooks bool
Webhooks bool
Avatar bool
IssueLabels bool
ProtectedBranch bool
}
// IsValid checks whether at least one option is chosen for generation
func (gro GenerateRepoOptions) IsValid() bool {
return gro.GitContent || gro.Topics || gro.GitHooks || gro.Webhooks || gro.Avatar ||
gro.IssueLabels || gro.ProtectedBranch // or other items as they are added
}
func filePathSanitize(s string) string {
fields := strings.Split(filepath.ToSlash(s), "/")
for i, field := range fields {
field = strings.TrimSpace(strings.TrimSpace(globalVars().fileNameSanitizeRegexp.ReplaceAllString(field, "_")))
if strings.HasPrefix(field, "..") {
field = "__" + field[2:]
}
if strings.EqualFold(field, ".git") {
field = "_" + field[1:]
}
fields[i] = field
}
return filepath.Clean(filepath.FromSlash(strings.Trim(strings.Join(fields, "/"), "/")))
}