all: detect annotation from REUSE configuration

During scan, the program will read the REUSE.toml configuration.
File that is already annotated inside REUSE.toml will be ignored during
scan.

11 files changed, 209 insertions, 11 deletions

diff --git a/README.md b/README.md
index e4f51d3..1fdf810 100644
--- a/README.md
+++ b/README.md
@@ -13,6 +13,7 @@ This tool works in tandem with [REUSE software](https://reuse.software).
 
 Features,
 
+- Detect annotations from REUSE configuration (REUSE.toml)
 - Customizable values for default license identifier and copyright
 - Customizable pattern for setting comment syntax based on file name
 - Customizable pattern for searching and capturing existing license through
@@ -249,6 +250,9 @@ by "apply" command.
 
 The scan command work in the following way,
 
+(0) Skip the file if its ignored by git or already annotated in REUSE.toml
+configuration.
+
 (1) Check the file for SPDX-License-Identifier and SPDX-FileCopyrightText.
 If both exist, skip the file.
 
diff --git a/go.mod b/go.mod
index fdcb4b9..75e2aec 100644
--- a/go.mod
+++ b/go.mod
@@ -5,7 +5,10 @@ module git.sr.ht/~shulhan/spdxconv
 
 go 1.24.0
 
-require git.sr.ht/~shulhan/pakakeh.go v0.60.3-0.20260113140641-d18ec5d8635b
+require (
+	git.sr.ht/~shulhan/pakakeh.go v0.60.3-0.20260114074200-83aa06572a83
+	github.com/BurntSushi/toml v1.6.0
+)
 
 require (
 	golang.org/x/exp v0.0.0-20251219203646-944ab1f22d93 // indirect
diff --git a/go.sum b/go.sum
index 8de41cb..409617e 100644
--- a/go.sum
+++ b/go.sum
@@ -1,5 +1,7 @@
-git.sr.ht/~shulhan/pakakeh.go v0.60.3-0.20260113140641-d18ec5d8635b h1:tPRWSIjzQTWoQJ4WG2zJFP1GybZuvpkniKjjJc4gvJ8=
-git.sr.ht/~shulhan/pakakeh.go v0.60.3-0.20260113140641-d18ec5d8635b/go.mod h1:1MkKXbLZRHTcnheeSEbRpGztkym4Yxzh90ep+jCxbDc=
+git.sr.ht/~shulhan/pakakeh.go v0.60.3-0.20260114074200-83aa06572a83 h1:7sKv4zQd+o5XC0hmNV+qjAlexpkrlkC5Rseh6WRagWM=
+git.sr.ht/~shulhan/pakakeh.go v0.60.3-0.20260114074200-83aa06572a83/go.mod h1:1MkKXbLZRHTcnheeSEbRpGztkym4Yxzh90ep+jCxbDc=
+github.com/BurntSushi/toml v1.6.0 h1:dRaEfpa2VI55EwlIW72hMRHdWouJeRF7TPYhI+AUQjk=
+github.com/BurntSushi/toml v1.6.0/go.mod h1:ukJfTF/6rtPPRCnwkur4qwRxa8vTRFBF0uk2lLoLwho=
 github.com/google/go-cmp v0.7.0 h1:wk8382ETsv4JYUZwIsn6YpYiWiBsYLSJiTsyBybVuN8=
 github.com/google/go-cmp v0.7.0/go.mod h1:pXiqmnSA92OHEEa9HXL2W4E7lf9JzCmGVUdgjX3N/iU=
 golang.org/x/exp v0.0.0-20251219203646-944ab1f22d93 h1:fQsdNF2N+/YewlRZiricy4P1iimyPKZ/xwniHj8Q2a0=
diff --git a/reuse_config.go b/reuse_config.go
new file mode 100644
index 0000000..3079762
--- /dev/null
+++ b/reuse_config.go
@@ -0,0 +1,93 @@
+// SPDX-License-Identifier: GPL-3.0-only
+// SPDX-FileCopyrightText: 2026 M. Shulhan <ms@kilabit.info>
+
+package spdxconv
+
+import (
+	"fmt"
+	"os"
+
+	"git.sr.ht/~shulhan/pakakeh.go/lib/git"
+	"github.com/BurntSushi/toml"
+)
+
+// ReuseConfigFile file name of REUSE configuration.
+const ReuseConfigFile = `REUSE.toml`
+
+type annotation struct {
+	precedence        string
+	fileCopyrightText string
+	licenseIdentifier string
+
+	path []string
+
+	includePattern []git.IgnorePattern
+}
+
+func (ann *annotation) UnmarshalTOML(raw any) (err error) {
+	data, _ := raw.(map[string]any)
+	for k, v := range data {
+		switch k {
+		case `path`:
+			switch vv := v.(type) {
+			case string:
+				ann.path = append(ann.path, vv)
+			case []any:
+				for _, p := range vv {
+					path, _ := p.(string)
+					ann.path = append(ann.path, path)
+				}
+			}
+		case `precedence`:
+			ann.precedence, _ = v.(string)
+		case `SPDX-FileCopyrightText`:
+			ann.fileCopyrightText, _ = v.(string)
+		case `SPDX-License-Identifier`:
+			ann.licenseIdentifier, _ = v.(string)
+		}
+	}
+	return nil
+}
+
+// reuseConfig define the struct for storing REUSE.toml configuration.
+type reuseConfig struct {
+	ListAnnotation []*annotation `toml:"annotations"`
+
+	Version int64
+}
+
+func loadReuseConfig(path string) (reuse *reuseConfig, err error) {
+	var logp = `loadReuseConfig`
+	content, err := os.ReadFile(path)
+	if err != nil {
+		if os.IsNotExist(err) {
+			return nil, nil
+		}
+		return nil, fmt.Errorf(`%s: %w`, logp, err)
+	}
+	reuse = &reuseConfig{}
+	err = toml.Unmarshal(content, reuse)
+	if err != nil {
+		return nil, fmt.Errorf(`%s: %w`, logp, err)
+	}
+	for _, ann := range reuse.ListAnnotation {
+		for _, path := range ann.path {
+			ign := git.ParseIgnorePattern([]byte(path))
+			ann.includePattern = append(ann.includePattern, ign)
+		}
+	}
+	return reuse, nil
+}
+
+// isAnnotated return true if the path is included in one of the annotations
+// table.
+func (reuse *reuseConfig) isAnnotated(path string) bool {
+	for _, ann := range reuse.ListAnnotation {
+		for _, pat := range ann.includePattern {
+			if pat.IsMatch(path) {
+				return true
+			}
+		}
+	}
+	return false
+}
diff --git a/reuse_config_test.go b/reuse_config_test.go
new file mode 100644
index 0000000..110daae
--- /dev/null
+++ b/reuse_config_test.go
@@ -0,0 +1,52 @@
+// SPDX-License-Identifier: GPL-3.0-only
+// SPDX-FileCopyrightText: 2026 M. Shulhan <ms@kilabit.info>
+
+package spdxconv
+
+import (
+	"testing"
+
+	"git.sr.ht/~shulhan/pakakeh.go/lib/git"
+	"git.sr.ht/~shulhan/pakakeh.go/lib/test"
+)
+
+func TestLoadReuseConfig(t *testing.T) {
+	listCase := []struct {
+		path     string
+		exp      *reuseConfig
+		expError string
+	}{{
+		path: `testdata/scan/REUSE.toml`,
+		exp: &reuseConfig{
+			Version: 1,
+			ListAnnotation: []*annotation{{
+				includePattern: []git.IgnorePattern{
+					git.ParseIgnorePattern([]byte(`po/*.po`)),
+					git.ParseIgnorePattern([]byte(`po/*.pot`)),
+				},
+				path:              []string{`po/*.po`, `po/*.pot`},
+				precedence:        `aggregate`,
+				fileCopyrightText: `2026 M. Shulhan <ms@kilabit.info>`,
+				licenseIdentifier: `GPL-3.0-only`,
+			}, {
+				includePattern: []git.IgnorePattern{
+					git.ParseIgnorePattern([]byte(`tests/resources/**`)),
+				},
+				path:              []string{`tests/resources/**`},
+				precedence:        `override`,
+				fileCopyrightText: `2026 M. Shulhan <ms@kilabit.info>`,
+				licenseIdentifier: `CC0-1.0`,
+			}},
+		},
+	}, {
+		path: `not_exist`,
+	}}
+	for _, tc := range listCase {
+		got, err := loadReuseConfig(tc.path)
+		if err != nil {
+			test.Assert(t, tc.path+` error`, tc.expError, err.Error())
+			continue
+		}
+		test.Assert(t, tc.path, tc.exp, got)
+	}
+}
diff --git a/spdxconv.go b/spdxconv.go
index 466b0ee..778ec51 100644
--- a/spdxconv.go
+++ b/spdxconv.go
@@ -26,6 +26,9 @@ var suffixLicense = `.license`
 
 // SPDXConv the main type for converting files to SPDX format.
 type SPDXConv struct {
+	// reuse contains the REUSE.toml configuration.
+	reuse *reuseConfig
+
 	scm sourceCodeManagement
 
 	// curDir contains the current working directory.
@@ -65,6 +68,8 @@ func Scan(path string) (err error) {
 		return fmt.Errorf(`%s: %w`, logp, err)
 	}
 
+	conv.loadReuseConfig()
+
 	var listFile []string
 	if conv.name == `` {
 		listFile, err = conv.scanDir([]string{path})
@@ -218,6 +223,16 @@ func (conv *SPDXConv) loadConfig(dir string) (err error) {
 	return nil
 }
 
+// loadReuseConfig load the REUSE.toml configuration from current directory.
+func (conv *SPDXConv) loadReuseConfig() {
+	path := filepath.Join(conv.curDir, ReuseConfigFile)
+	var err error
+	conv.reuse, err = loadReuseConfig(path)
+	if err != nil {
+		return
+	}
+}
+
 // scanForSCM scan for source-code management (SCM) from directory `dir` until
 // the current working directory.
 // Currently, only support git.
@@ -267,14 +282,15 @@ func (conv *SPDXConv) scanFile(dir, name string) (listFile []string, err error)
 func (conv *SPDXConv) scanDir(listDir []string) (listFile []string, err error) {
 	var logp = `scanDir`
 	var commonIgnore = map[string]struct{}{
-		`.git`:         struct{}{},
-		ConfigFile:     struct{}{},
-		`COPYING`:      struct{}{},
-		`LICENSE`:      struct{}{},
-		`LICENSES`:     struct{}{}, // Directory contains licenses.
-		`node_modules`: struct{}{},
-		ReportFile:     struct{}{},
-		`vendor`:       struct{}{},
+		`.git`:          struct{}{},
+		ConfigFile:      struct{}{},
+		`COPYING`:       struct{}{},
+		`LICENSE`:       struct{}{},
+		`LICENSES`:      struct{}{}, // Directory contains licenses.
+		`node_modules`:  struct{}{},
+		ReportFile:      struct{}{},
+		ReuseConfigFile: struct{}{},
+		`vendor`:        struct{}{},
 	}
 
 	var dir string
@@ -309,6 +325,9 @@ func (conv *SPDXConv) scanDir(listDir []string) (listFile []string, err error) {
 			if conv.scm.IsIgnored(relpath) {
 				continue
 			}
+			if conv.reuse != nil && conv.reuse.isAnnotated(relpath) {
+				continue
+			}
 			if de.IsDir() {
 				listDir = append(listDir, relpath)
 				continue
diff --git a/spdxconv_test.go b/spdxconv_test.go
index f473a2e..fd318b8 100644
--- a/spdxconv_test.go
+++ b/spdxconv_test.go
@@ -135,6 +135,7 @@ func TestScan(t *testing.T) {
 //spdxconv:regular
 //
 no_copyright_year.md,match,2,2026,default,0,<!-- ," -->"
+po/included.go,match,1,2026,match,0,// ,
 test.go,match,1,2022,match,0,// ,
 test.html,match,4,2022,match,1,<!-- ," -->"
 test.sh,default,0,2026,default,0,# ,
@@ -337,6 +338,7 @@ func TestSPDXConv_scanDir(t *testing.T) {
 			`.gitignore`,
 			`a/b/.gitignore`,
 			`no_copyright_year.md`,
+			`po/included.go`,
 			`test.go`,
 			`test.html`,
 			`test.sh`,
@@ -362,6 +364,7 @@ func TestSPDXConv_scanDir(t *testing.T) {
 			if err != nil {
 				tt.Fatal(err)
 			}
+			conv.loadReuseConfig()
 			got, err = conv.scanDir([]string{`.`})
 			if err != nil {
 				tt.Fatal(err)
diff --git a/testdata/scan/REUSE.toml b/testdata/scan/REUSE.toml
new file mode 100644
index 0000000..3621d08
--- /dev/null
+++ b/testdata/scan/REUSE.toml
@@ -0,0 +1,13 @@
+version = 1
+
+[[annotations]]
+path = ["po/*.po", "po/*.pot"]
+precedence = "aggregate"
+SPDX-FileCopyrightText = "2026 M. Shulhan <ms@kilabit.info>"
+SPDX-License-Identifier = "GPL-3.0-only"
+
+[[annotations]]
+path = "tests/resources/**"
+precedence = "override"
+SPDX-FileCopyrightText = "2026 M. Shulhan <ms@kilabit.info>"
+SPDX-License-Identifier = "CC0-1.0"
diff --git a/testdata/scan/po/included.go b/testdata/scan/po/included.go
new file mode 100644
index 0000000..d499f94
--- /dev/null
+++ b/testdata/scan/po/included.go
@@ -0,0 +1,7 @@
+// Copyright 2026, Shulhan <ms@kilabit.info>. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package po
+
+// This file not annotated by REUSE.toml.
diff --git a/testdata/scan/po/test.po b/testdata/scan/po/test.po
new file mode 100644
index 0000000..327d157
--- /dev/null
+++ b/testdata/scan/po/test.po
@@ -0,0 +1 @@
+This file is annotated by REUSE.toml.
diff --git a/testdata/scan/po/test.pot b/testdata/scan/po/test.pot
new file mode 100644
index 0000000..327d157
--- /dev/null
+++ b/testdata/scan/po/test.pot
@@ -0,0 +1 @@
+This file is annotated by REUSE.toml.