Skip to content

Commit

Permalink
added Glob as package function
Browse files Browse the repository at this point in the history
  • Loading branch information
@ungerik
ungerik committed Sep 19, 2024
1 parent 437d7c2 commit 1b6f933
Show file tree
Hide file tree
Showing 2 changed files with 202 additions and 17 deletions.
100 changes: 84 additions & 16 deletions file.go
View file
Original file line number Diff line number Diff line change
Expand Up @@ -493,7 +493,34 @@ func (file File) ListDirIterContext(ctx context.Context, patterns ...string) ite
}

// MustGlob yields files and wildcard substituting path segments
// matching a pattern.
// matching a path pattern.
//
// The pattern is appended to the current working directory
// if it is not an absolute path.
//
// The yielded path segments are the strings necessary
// to substitute all wildcard containing path segments
// in the pattern to form a valid path for a yielded file.
// This includes the name of the yielded file itself
// if the pattern contains wildcards for the last segment.
// Non wildcard segments are not included.
//
// The syntax of patterns is the same as in [path.Match].
// It always uses slash '/' as path segment separator
// independently of the file's file system.
//
// MustGlob ignores file system errors such as I/O errors reading directories.
// The only possible panic is in case of a malformed pattern.
func MustGlob(pattern string) iter.Seq2[File, []string] {
globIter, err := Glob(pattern)
if err != nil {
panic(err)
}
return globIter
}

// MustGlob yields files and wildcard substituting path segments
// matching a path pattern.
//
// The yielded path segments are the strings necessary
// to substitute all wildcard containing path segments
Expand All @@ -506,18 +533,60 @@ func (file File) ListDirIterContext(ctx context.Context, patterns ...string) ite
// It always uses slash '/' as path segment separator
// independently of the file's file system.
//
// A pattern ending with a slash '/' will match only directories.
//
// MustGlob ignores file system errors such as I/O errors reading directories.
// The only possible panic is in case of a malformed pattern.
func (file File) MustGlob(pattern string) iter.Seq2[File, []string] {
iterator, err := file.Glob(pattern)
globIter, err := file.Glob(pattern)
if err != nil {
panic(err)
}
return iterator
return globIter
}

// Glob yields files and wildcard substituting path segments
// matching a pattern.
// matching a path pattern.
//
// The pattern is appended to the current working directory
// if it is not an absolute path.
//
// The yielded path segments are the strings necessary
// to substitute all wildcard containing path segments
// in the pattern to form a valid path for a yielded file.
// This includes the name of the yielded file itself
// if the pattern contains wildcards for the last segment.
// Non wildcard segments are not included.
//
// The syntax of patterns is the same as in [path.Match].
// It always uses slash '/' as path segment separator
// independently of the file's file system.
//
// A pattern ending with a slash '/' will match only directories.
//
// Glob ignores file system errors such as I/O errors reading directories.
// The only possible returned error is [path.ErrBadPattern],
// reporting that the pattern is malformed.
func Glob(pattern string) (iter.Seq2[File, []string], error) {
// Find the first wildcard
i := strings.IndexAny(pattern, `*?[\`)
if i == -1 {
// No wildcard in pattern, yield the pattern as File
return File(path.Clean(pattern)).Glob("")
}
// Find the last path separator before the first wildcard
i = strings.LastIndexByte(pattern[:i], '/')
if i == -1 {
// No path separator before the first wildcard
// means that the pattern is relative to the current directory
return CurrentWorkingDir().Glob(pattern)
}
// Split pattern into base directory and glob pattern
return File(pattern[:i+1]).Glob(pattern[i+1:])
}

// Glob yields files and wildcard substituting path segments
// matching a path pattern.
//
// The yielded path segments are the strings necessary
// to substitute all wildcard containing path segments
Expand All @@ -536,7 +605,7 @@ func (file File) MustGlob(pattern string) iter.Seq2[File, []string] {
// The only possible returned error is [path.ErrBadPattern],
// reporting that the pattern is malformed.
func (file File) Glob(pattern string) (iter.Seq2[File, []string], error) {
dirOnly := strings.HasSuffix(pattern, "/")
onlyDirs := strings.HasSuffix(pattern, "/")
pattern = strings.Trim(pattern, "/")
// Check if the pattern is valid
if _, err := path.Match(pattern, ""); err != nil {
Expand All @@ -546,10 +615,8 @@ func (file File) Glob(pattern string) (iter.Seq2[File, []string], error) {
pSegments = slices.DeleteFunc(pSegments, func(s string) bool {
return s == "" || s == "."
})
for _, seg := range pSegments {
if seg == ".." {
return nil, fmt.Errorf("%w, must not contain '..': %s", path.ErrBadPattern, pattern)
}
if slices.Contains(pSegments, "..") {
return nil, fmt.Errorf("%w, must not contain '..': %s", path.ErrBadPattern, pattern)
}
switch i := slices.IndexFunc(pSegments, containsWildcard); {
case i < 0:
Expand All @@ -561,19 +628,20 @@ func (file File) Glob(pattern string) (iter.Seq2[File, []string], error) {
file = file.Join(pSegments[:i]...)
pSegments = pSegments[i:]
}
return file.glob(dirOnly, pSegments, nil), nil
file.CheckIsDir()
return file.glob(onlyDirs, pSegments, nil), nil
}

func containsWildcard(pattern string) bool {
return strings.ContainsAny(pattern, `*?[\`)
}

func (file File) glob(dirOnly bool, segments, values []string) iter.Seq2[File, []string] {
func (file File) glob(onlyDirs bool, segments, values []string) iter.Seq2[File, []string] {
return func(yield func(File, []string) bool) {
switch len(segments) {
case 0:
// No more segments, yield the file itself
if dirOnly && file.IsDir() || !dirOnly && file.Exists() {
if onlyDirs && file.IsDir() || !onlyDirs && file.Exists() {
yield(file, values)
}

Expand All @@ -586,7 +654,7 @@ func (file File) glob(dirOnly bool, segments, values []string) iter.Seq2[File, [
if err != nil {
return
}
if dirOnly && !f.IsDir() || !dirOnly && !f.Exists() {
if onlyDirs && !f.IsDir() || !onlyDirs && !f.Exists() {
continue
}
if !yield(f, append(slices.Clone(values), f.Name())) {
Expand All @@ -596,7 +664,7 @@ func (file File) glob(dirOnly bool, segments, values []string) iter.Seq2[File, [
} else {
// No wildcard in last segment, join path and yield file if it exists
f := file.Join(pattern)
if dirOnly && f.IsDir() || !dirOnly && f.Exists() {
if onlyDirs && f.IsDir() || !onlyDirs && f.Exists() {
yield(f, values)
}
}
Expand All @@ -609,15 +677,15 @@ func (file File) glob(dirOnly bool, segments, values []string) iter.Seq2[File, [
if err != nil {
return
}
for f, v := range matchedFile.glob(dirOnly, segments[1:], append(slices.Clone(values), matchedFile.Name())) {
for f, v := range matchedFile.glob(onlyDirs, segments[1:], append(slices.Clone(values), matchedFile.Name())) {
if !yield(f, v) {
return
}
}
}
} else {
// No wildcard in segment, join path and recurse
for f, v := range file.Join(segments[0]).glob(dirOnly, segments[1:], values) {
for f, v := range file.Join(segments[0]).glob(onlyDirs, segments[1:], values) {
if !yield(f, v) {
return
}
Expand Down
119 changes: 118 additions & 1 deletion file_test.go
View file
Original file line number Diff line number Diff line change
Expand Up @@ -10,7 +10,6 @@ import (

"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"

"github.com/ungerik/go-fs/fsimpl"
)

Expand Down Expand Up @@ -426,3 +425,121 @@ func TestFile_Glob(t *testing.T) {
})
}
}

func TestGlob(t *testing.T) {
dir := MustMakeTempDir()
t.Cleanup(func() { dir.RemoveRecursive() })
xDir := dir.Join("a", "b", "c", "Hello", "World", "x")
yDir := dir.Join("a", "b", "c", "Hello", "World", "y")
require.NoError(t, xDir.MakeAllDirs())
require.NoError(t, yDir.MakeAllDirs())
cFile := dir.Join("a", "b", "c", "cFile")
require.NoError(t, cFile.Touch())
xFile1 := xDir.Join("file1.txt")
require.NoError(t, xFile1.Touch())
xFile2 := xDir.Join("file2.txt")
require.NoError(t, xFile2.Touch())
xFile3 := xDir.Join("file3.txt")
require.NoError(t, xFile3.Touch())

type result struct {
file File
values []string
}

tests := []struct {
name string
pattern string
want []result
wantErr bool
}{
{
name: "empty pattern for current dir",
pattern: "",
want: []result{{".", nil}},
},
{
name: "no wildcard pattern",
pattern: dir.PathWithSlashes() + "/a/b/c/Hello/World/x/file1.txt",
want: []result{
{xFile1, nil},
},
},
{
name: "no wildcard non-canonical pattern",
pattern: dir.PathWithSlashes() + "/./a/b//c/Hello/./././/World/x/file1.txt",
want: []result{
{xFile1, nil},
},
},
{
name: "root files",
pattern: dir.PathWithSlashes() + "/*",
want: []result{
{dir.Join("a"), []string{"a"}},
},
},
{
name: "root dirs",
pattern: dir.PathWithSlashes() + "/*/",
want: []result{
{dir.Join("a"), []string{"a"}},
},
},
{
name: "file and dir",
pattern: dir.PathWithSlashes() + "/a/b/c/*",
want: []result{
{dir.Join("a", "b", "c", "Hello"), []string{"Hello"}},
{cFile, []string{"cFile"}},
},
},
{
name: "directories only",
pattern: dir.PathWithSlashes() + "/a/b/c/*/",
want: []result{
{dir.Join("a", "b", "c", "Hello"), []string{"Hello"}},
},
},
{
name: "complexer pattern",
pattern: dir.PathWithSlashes() + "/*/b/c/*/W???d/x/file[1-2].txt",
want: []result{
{xFile1, []string{"a", "Hello", "World", "file1.txt"}},
{xFile2, []string{"a", "Hello", "World", "file2.txt"}},
},
},
{
name: "*.txt",
pattern: dir.PathWithSlashes() + "/a/b/c/Hello/World/x/*.txt",
want: []result{
{xFile1, []string{"file1.txt"}},
{xFile2, []string{"file2.txt"}},
{xFile3, []string{"file3.txt"}},
},
},
// Errors
{
name: "malformed pattern",
pattern: "/a/b/c/Hello/World/x/[file1.txt",
wantErr: true,
},
}
for _, tt := range tests {
t.Run(tt.name, func(t *testing.T) {
gotIter, err := Glob(tt.pattern)
if tt.wantErr {
require.Error(t, err, "Glob")
return
}
require.NoError(t, err, "Glob")
var got []result
for file, values := range gotIter {
require.Truef(t, file.Exists(), "file %s does not exist", file)
got = append(got, result{file, values})
}
sort.Slice(got, func(i, j int) bool { return got[i].file.LocalPath() < got[j].file.LocalPath() })
require.Equal(t, tt.want, got, "file path sorted results")
})
}
}

0 comments on commit 1b6f933

Please sign in to comment.