Gitattributes

Makefile

@@ -26,7 +26,7 @@ test-coverage: $(LINGUIST_PATH)
 				tail -n +2 $(COVERAGE_PROFILE) >> $(COVERAGE_REPORT); \
 				rm $(COVERAGE_PROFILE); \
 			fi; \
-	done;
+  done;
 
 code-generate: $(LINGUIST_PATH)
 	mkdir -p data

README.md

@@ -118,7 +118,7 @@ Note that even if enry's CLI is compatible with linguist's, its main point is th
 Development
 ------------
 
-*enry* re-uses parts of original [linguist](https://github.com/github/linguist) especially data in `languages.yml` to generate internal data structures. In oreder to update to latest upstream run
+*enry* re-uses parts of original [linguist](https://github.com/github/linguist) especially data in `languages.yml` to generate internal data structures. In order to update to latest upstream run
 
     make clean code-generate
 
@@ -140,6 +140,7 @@ Using [linguist/samples](https://github.com/github/linguist/tree/master/samples)
 * all files for SQL language fall to the classifier because we don't parse this [disambiguator expresion](https://github.com/github/linguist/blob/master/lib/linguist/heuristics.rb#L433) for `*.sql` files right. This expression doesn't comply with the pattern for the rest of [heuristics.rb](https://github.com/github/linguist/blob/master/lib/linguist/heuristics.rb) file.
 
 
+
 Benchmarks
 ------------
 
@@ -172,6 +173,38 @@ to get time averages for main detection function and strategies for the whole sa
 if you want see measures by sample file
 
 
+
+.gitattributes
+--------------
+
+Like in linguist you can override the strategies via `.gitattributes` file.
+Add a `.gitattributes` file to the directory and use the same matchers that you would use in linguist `linguist-documentation`,`linguist-language` or `linguist-vendored` to do the override.
+
+#### Vendored code
+
+Use the `linguist-vendored` attribute to vendor or un-vendor paths.
+
+```
+$ cat .gitattributes
+this-is-a-vendor-directory/ linguist-vendored
+this-is-not/ linguist-vendored=false
+```
+#### Documentation
+
+Documentation works the same way as vendored code but using `linguist-documentation` and `linguist-documentation=false`.
+
+#### Language assignation
+
+If you want some files to be classified according to certain language use `linguist-language=[language]`.
+
+```
+$ cat .gitattributes
+.*\.go linguist-language=MyFavouriteLanguage
+```
+
+Note that the regular expression that matches the file name should be compatible with go, see: [Golang regexp](https://golang.org/pkg/regexp/).
+
+
 Why Enry?
 ------------
 

cli/enry/main.go

@@ -29,6 +29,12 @@ func main() {
 		log.Fatal(err)
 	}
 
+	gitAttributes := enry.NewGitAttributes()
+	reader, err := os.Open(".gitattributes")
+	if err == nil {
+		gitAttributes.LoadGitAttributes("", reader)
+	}
+
 	errors := false
 	out := make(map[string][]string, 0)
 	err = filepath.Walk(root, func(path string, f os.FileInfo, err error) error {
@@ -53,8 +59,9 @@ func main() {
 			relativePath = relativePath + "/"
 		}
 
-		if enry.IsVendor(relativePath) || enry.IsDotFile(relativePath) ||
-			enry.IsDocumentation(relativePath) || enry.IsConfiguration(relativePath) {
+		if gitAttributes.IsVendor(relativePath) || enry.IsDotFile(relativePath) ||
+			gitAttributes.IsDocumentation(relativePath) || enry.IsConfiguration(relativePath) ||
+			gitAttributes.IsGenerated(path) {
 			if f.IsDir() {
 				return filepath.SkipDir
 			}
@@ -66,20 +73,18 @@ func main() {
 			return nil
 		}
 
-		language, ok := enry.GetLanguageByExtension(path)
-		if !ok {
-			if language, ok = enry.GetLanguageByFilename(path); !ok {
-				content, err := ioutil.ReadFile(path)
-				if err != nil {
-					errors = true
-					log.Println(err)
-					return nil
-				}
-
-				language = enry.GetLanguage(filepath.Base(path), content)
-				if language == enry.OtherLanguage {
-					return nil
-				}
+		content, err := ioutil.ReadFile(path)
+		if err != nil {
+			errors = true
+			log.Println(err)
+			return nil
+		}
+
+		language := gitAttributes.GetLanguage(filepath.Base(path))
+		if language == enry.OtherLanguage {
+			language = enry.GetLanguage(filepath.Base(path), content)
+			if language == enry.OtherLanguage {
+				return nil
 			}
 		}
 

common.go

@@ -3,6 +3,7 @@ package enry
 import (
 	"bufio"
 	"bytes"
+	"os"
 	"path/filepath"
 	"regexp"
 	"strings"
@@ -95,6 +96,12 @@ func GetLanguageByClassifier(content []byte, candidates []string) (language stri
 	return getLanguageByStrategy(GetLanguagesByClassifier, "", content, candidates)
 }
 
+// GetLanguageByGitattributes returns the language assigned to a file for a given regular expresion in .gitattributes.
+// This strategy needs to be initialized calling LoadGitattributes
+func GetLanguageByGitattributes(filename string) (language string, safe bool) {
+	return getLanguageByStrategy(GetLanguagesByGitAttributes, filename, nil, nil)
+}
+
 func getLanguageByStrategy(strategy Strategy, filename string, content []byte, candidates []string) (string, bool) {
 	languages := strategy(filename, content, candidates)
 	return getFirstLanguageAndSafe(languages)
@@ -407,6 +414,25 @@ func GetLanguagesBySpecificClassifier(content []byte, candidates []string, class
 	return classifier.Classify(content, mapCandidates)
 }
 
+// GetLanguagesByGitAttributes returns either a string slice with the language
+// if the filename matches with a regExp in .gitattributes or returns an empty slice
+// in case no regExp matches the filename. It complies with the signature to be a Strategy type.
+func GetLanguagesByGitAttributes(filename string, content []byte, candidates []string) []string {
+	gitAttributes := NewGitAttributes()
+	reader, err := os.Open(".gitattributes")
+	if err != nil {
+		return nil
+	}
+
+	gitAttributes.LoadGitAttributes("", reader)
+	lang := gitAttributes.GetLanguage(filename)
+	if lang != OtherLanguage {
+		return []string{}
+	}
+
+	return []string{lang}
+}
+
 // GetLanguageExtensions returns the different extensions being used by the language.
 func GetLanguageExtensions(language string) []string {
 	return data.ExtensionsByLanguage[language]