fix: reject directory args in keywords CLI

.hound.yml

@@ -0,0 +1,10 @@
+# Hound bundles a RuboCop older than 0.78. That version cannot parse
+# .rubocop.yml: it expects the Metrics/LineLength namespace instead of
+# Layout/LineLength, and it rejects TargetRubyVersion 3.1. Pinning it to
+# 2019 rules would judge this codebase, and its rbs-inline annotations,
+# by cops that predate them.
+#
+# The CI lint job runs RuboCop 1.89 against .rubocop.yml on every pull
+# request. That job is the gate for Ruby style.
+ruby:
+  enabled: false

lib/classifier/keywords/cli.rb

@@ -144,12 +144,7 @@ def command_fit
           if @args.empty?
             [@stdin ? StringIO.new(@stdin.to_s) : $stdin]
           else
-            @args.flat_map do |arg|
-              matches = Dir.glob(arg)
-              raise UsageError, "No files matched #{arg.inspect}" if matches.empty?
-
-              matches.map { |f| File.expand_path(f) }
-            end.uniq
+            collect_files
           end
 
         tfidf = TFIDF.new(
@@ -165,6 +160,19 @@ def command_fit
         @output << "Saved to #{@options[:model].inspect}" unless @options[:quiet]
       end
 
+      def collect_files
+        files = @args.flat_map do |arg|
+          matches = Dir.glob(arg)
+          raise UsageError, "No files matched #{arg.inspect}" if matches.empty?
+
+          matches.select { |f| File.file?(f) }.map { |f| File.expand_path(f) }
+        end.uniq
+
+        raise UsageError, 'No files to fit' if files.empty?
+
+        files
+      end
+
       def command_extract
         @args.shift
 
@@ -174,6 +182,7 @@ def command_extract
           else
             file = File.expand_path(@args.first)
             raise UsageError, "File #{file.inspect} does not exist" unless File.exist?(file)
+            raise UsageError, "#{file.inspect} is a directory, not a file" if File.directory?(file)
 
             File.read(file)
           end

test/keywords/cli_test.rb

@@ -212,6 +212,49 @@ def test_fit_command_with_no_documents
       refute_path_exists @model_path
     end
 
+    def test_fit_command_skips_directories_matched_by_a_glob
+      make_articles
+      FileUtils.mkdir_p(File.join(@tmpdir, 'articles', 'nested'))
+
+      result = run_cli('-m', @model_path, 'fit', *Dir.glob(File.join(@tmpdir, 'articles/*')))
+
+      assert_equal 0, result[:exit_code]
+      assert_predicate File.size(@model_path), :positive?
+    end
+
+    def test_fit_command_with_directory_argument
+      dir = File.join(@tmpdir, 'articles')
+      FileUtils.mkdir_p(dir)
+
+      result = run_cli('-m', @model_path, 'fit', dir)
+
+      assert_equal 2, result[:exit_code]
+      assert_match('Error: No files to fit', result[:error])
+      refute_path_exists @model_path
+    end
+
+    def test_fit_command_with_not_exists_input_file
+      make_articles
+      missing = File.join(@tmpdir, 'articles', 'not_exists.txt')
+
+      result = run_cli('-m', @model_path, 'fit', File.join(@tmpdir, 'articles/a1.txt'), missing)
+
+      assert_equal 2, result[:exit_code]
+      assert_match("Error: No files matched #{missing.inspect}", result[:error])
+      refute_path_exists @model_path
+    end
+
+    def test_extract_command_with_directory_input
+      make_model
+      dir = File.join(@tmpdir, 'articles')
+      FileUtils.mkdir_p(dir)
+
+      result = run_cli('-m', @model_path, 'extract', dir)
+
+      assert_equal 2, result[:exit_code]
+      assert_match("Error: #{dir.inspect} is a directory, not a file", result[:error])
+    end
+
     def test_extract_command_with_not_exists_input_file
       make_model
       file = File.join(@tmpdir, 'not_exists.txt')