Try to guess language by file content

pkg/language/chroma.go

@@ -97,9 +97,50 @@ func detectChromaCustomized(filepath string) (heartbeat.Language, float32, bool)
 		return language, weight, true
 	}
 
+	// Finally, try matching by file content.
+	head, err := fileHead(filepath)
+	if err != nil {
+		log.Warnf("failed to load head from file %q: %s", filepath, err)
+		return heartbeat.LanguageUnknown, 0, false
+	}
+
+	if len(head) == 0 {
+		return heartbeat.LanguageUnknown, 0, false
+	}
+
+	if lexer := analyse(string(head)); lexer != nil {
+		language, ok := heartbeat.ParseLanguageFromChroma(lexer.Config().Name)
+		if !ok {
+			log.Warnf("failed to parse language from chroma lexer name %q", lexer.Config().Name)
+			return heartbeat.LanguageUnknown, 0, false
+		}
+
+		return language, 0, true
+	}
+
 	return heartbeat.LanguageUnknown, 0, false
 }
 
+// analyse text content and return the "best" lexer.
+// This is a copy of chroma.lexers.internal.api:Analyse().
+func analyse(text string) chroma.Lexer {
+	var picked chroma.Lexer
+
+	highest := float32(0.0)
+
+	for _, lexer := range lexers.Registry.Lexers {
+		if analyser, ok := lexer.(chroma.Analyser); ok {
+			weight := analyser.AnalyseText(text)
+			if weight > highest {
+				picked = lexer
+				highest = weight
+			}
+		}
+	}
+
+	return picked
+}
+
 // weightedLexer is a lexer with priority and weight.
 type weightedLexer struct {
 	chroma.Lexer
@@ -125,7 +166,7 @@ func selectByCustomizedPriority(filepath string, lexers chroma.PrioritisedLexers
 
 	extensions, err := loadFolderExtensions(dir)
 	if err != nil {
-		log.Warnf("failed to load folder extensions: %s", err)
+		log.Warnf("failed to load folder files extensions: %s", err)
 		return lexers[0], 0
 	}
 
@@ -214,9 +255,7 @@ func fileHead(filepath string) ([]byte, error) {
 		}
 	}()
 
-	data := make([]byte, maxFileSize)
-
-	_, err = f.ReadAt(data, 0)
+	data, err := io.ReadAll(io.LimitReader(f, maxFileSize))
 	if err != nil && err != io.EOF {
 		return nil, fmt.Errorf("failed to read bytes from file: %s", err)
 	}

pkg/language/language.go

@@ -175,7 +175,7 @@ func correspondingFileExists(fp string, extension string) bool {
 	return false
 }
 
-// loadFolderExtensions loads all existing from a folder.
+// loadFolderExtensions loads all existing file extensions from a folder.
 func loadFolderExtensions(dir string) ([]string, error) {
 	files, err := os.ReadDir(dir)
 	if err != nil {

pkg/language/language_test.go

@@ -305,6 +305,7 @@ func TestDetect_ChromaTopLanguagesRetrofit(t *testing.T) {
 				"path/to/zshrc",
 				"path/to/.zshrc",
 				"path/to/PKGBUILD",
+				"testdata/bash",
 			},
 			Expected: heartbeat.LanguageBash,
 		},
@@ -469,8 +470,11 @@ func TestDetect_ChromaTopLanguagesRetrofit(t *testing.T) {
 			Expected: heartbeat.LanguageINI,
 		},
 		"java": {
-			Filepaths: []string{"path/to/java.java"},
-			Expected:  heartbeat.LanguageJava,
+			Filepaths: []string{
+				"path/to/java.java",
+				"testdata/java",
+			},
+			Expected: heartbeat.LanguageJava,
 		},
 		"javascript": {
 			Filepaths: []string{
@@ -586,8 +590,11 @@ func TestDetect_ChromaTopLanguagesRetrofit(t *testing.T) {
 			Expected:  heartbeat.LanguagePawn,
 		},
 		"perl not prolog": {
-			Filepaths: []string{"testdata/codefiles/chroma_unsupported_top/perl.pl"},
-			Expected:  heartbeat.LanguagePerl,
+			Filepaths: []string{
+				"testdata/codefiles/chroma_unsupported_top/perl.pl",
+				"testdata/perl",
+			},
+			Expected: heartbeat.LanguagePerl,
 		},
 		"php": {
 			Filepaths: []string{
@@ -655,6 +662,7 @@ func TestDetect_ChromaTopLanguagesRetrofit(t *testing.T) {
 				"path/to/BUILD.bazel",
 				"path/to/WORKSPACE",
 				"path/to/file.tac",
+				"testdata/python3",
 			},
 			Expected: heartbeat.LanguagePython,
 		},

pkg/language/testdata/bash

@@ -0,0 +1,19 @@
+#!/bin/bash
+
+# Greet user and request their name
+echo "The activity generator"
+read -p "What is your name? " name
+
+# Create an array of activities
+activity[0]="Football"
+activity[1]="Table Tennis"
+activity[2]="8 Ball Pool"
+activity[3]="PS5"
+activity[4]="Blackjack"
+
+array_length=${#activity[@]} # Store the length of the array
+index=$(($RANDOM % $array_length)) # Randomly select an index from 0 to array_length
+
+# Invite the user to join you participate in an activity
+echo "Hi" $name, "would you like to play" ${activity[$index]}"?"
+read -p "Answer: " answer

pkg/language/testdata/perl

@@ -0,0 +1,18 @@
+#!/usr/bin/perl
+use strict;
+use warnings;
+
+use Path::Tiny;
+
+my $dir = path('foo','bar'); # foo/bar
+
+# Iterate over the content of foo/bar
+my $iter = $dir->iterator;
+while (my $file = $iter->()) {
+
+    # See if it is a directory and skip
+    next if $file->is_dir();
+
+    # Print out the file name and path
+    print "$file\n";
+}

pkg/language/testdata/python3

@@ -0,0 +1,4 @@
+#!/usr/bin/python3
+
+if __name__ == "__main__":
+    print("Hello, World!")