Merge pull request #268 from zachfeldman/zfeldman/cbruckmayer/impleme…

…nt-file-cache

[Feature] Implement file cache

README.md

@@ -612,6 +612,23 @@ app/views/users/_graph.html.erb:27:37: Extra space detected where there should b
 2 error(s) were found in ERB files
 ```
 
+## Caching
+
+The cache is currently opt-in - to turn it on, use the --cache option:
+
+```sh
+erblint --cache ./app
+Cache mode is on
+Linting 413 files with 15 linters...
+File names pruned from the cache will be logged
+
+No errors were found in ERB files
+```
+
+When the cache is on, lint results are stored in the `.erb-lint-cache` directory, in files with a filename computed with a hash of information about the file and `erb-lint` that should change when necessary. These files store instance attributes of the `CachedOffense` object, which only contain the `Offense` attributes necessary to restore the results of running `erb-lint` for output. The cache also automatically prunes outdated files each time it's run.
+
+You can also use the --clear-cache option to delete the cache file directory.
+
 ## License
 
 This project is released under the [MIT license](LICENSE.txt).

lib/erb_lint/all.rb

@@ -3,6 +3,8 @@
 require "rubocop"
 
 require "erb_lint"
+require "erb_lint/cache"
+require "erb_lint/cached_offense"
 require "erb_lint/corrector"
 require "erb_lint/file_loader"
 require "erb_lint/linter_config"

lib/erb_lint/cache.rb

@@ -0,0 +1,88 @@
+# frozen_string_literal: true
+
+module ERBLint
+  class Cache
+    CACHE_DIRECTORY = ".erb-lint-cache"
+
+    def initialize(config, file_loader = nil)
+      @config = config
+      @file_loader = file_loader
+      @hits = []
+      @new_results = []
+      puts "Cache mode is on"
+    end
+
+    def get(filename, file_content)
+      file_checksum = checksum(filename, file_content)
+      begin
+        cache_file_contents_as_offenses = JSON.parse(
+          File.read(File.join(CACHE_DIRECTORY, file_checksum))
+        ).map do |offense_hash|
+          ERBLint::CachedOffense.new(offense_hash)
+        end
+      rescue Errno::ENOENT
+        return false
+      end
+      @hits.push(file_checksum)
+      cache_file_contents_as_offenses
+    end
+
+    def set(filename, file_content, offenses_as_json)
+      file_checksum = checksum(filename, file_content)
+      @new_results.push(file_checksum)
+
+      FileUtils.mkdir_p(CACHE_DIRECTORY)
+
+      File.open(File.join(CACHE_DIRECTORY, file_checksum), "wb") do |f|
+        f.write(offenses_as_json)
+      end
+    end
+
+    def close
+      prune_cache
+    end
+
+    def prune_cache
+      if hits.empty?
+        puts "Cache being created for the first time, skipping prune"
+        return
+      end
+
+      cache_files = Dir.new(CACHE_DIRECTORY).children
+      cache_files.each do |cache_file|
+        next if hits.include?(cache_file) || new_results.include?(cache_file)
+
+        File.delete(File.join(CACHE_DIRECTORY, cache_file))
+      end
+    end
+
+    def cache_dir_exists?
+      File.directory?(CACHE_DIRECTORY)
+    end
+
+    def clear
+      return unless cache_dir_exists?
+
+      puts "Clearing cache by deleting cache directory"
+      FileUtils.rm_r(CACHE_DIRECTORY)
+    end
+
+    private
+
+    attr_reader :config, :hits, :new_results
+
+    def checksum(filename, file_content)
+      digester = Digest::SHA1.new
+      mode = File.stat(filename).mode
+
+      digester.update(
+        "#{mode}#{config.to_hash}#{ERBLint::VERSION}#{file_content}"
+      )
+      digester.hexdigest
+    rescue Errno::ENOENT
+      # Spurious files that come and go should not cause a crash, at least not
+      # here.
+      "_"
+    end
+  end
+end

lib/erb_lint/cached_offense.rb

@@ -0,0 +1,58 @@
+# frozen_string_literal: true
+
+module ERBLint
+  # A Cached version of an Offense with only essential information represented as strings
+  class CachedOffense
+    attr_reader(
+      :message,
+      :line_number,
+      :severity,
+      :column,
+      :simple_name,
+      :last_line,
+      :last_column,
+      :length,
+    )
+
+    def initialize(params)
+      params = params.transform_keys(&:to_sym)
+
+      @message = params[:message]
+      @line_number = params[:line_number]
+      @severity = params[:severity]&.to_sym
+      @column = params[:column]
+      @simple_name = params[:simple_name]
+      @last_line = params[:last_line]
+      @last_column = params[:last_column]
+      @length = params[:length]
+    end
+
+    def self.new_from_offense(offense)
+      new(
+        {
+          message: offense.message,
+          line_number: offense.line_number,
+          severity: offense.severity,
+          column: offense.column,
+          simple_name: offense.simple_name,
+          last_line: offense.last_line,
+          last_column: offense.last_column,
+          length: offense.length,
+        }
+      )
+    end
+
+    def to_h
+      {
+        message: message,
+        line_number: line_number,
+        severity: severity,
+        column: column,
+        simple_name: simple_name,
+        last_line: last_line,
+        last_column: last_column,
+        length: length,
+      }
+    end
+  end
+end

lib/erb_lint/cli.rb

@@ -30,10 +30,26 @@ def initialize
     def run(args = ARGV)
       dupped_args = args.dup
       load_options(dupped_args)
+
+      if cache? && autocorrect?
+        failure!("cannot run autocorrect mode with cache")
+      end
+
       @files = @options[:stdin] || dupped_args
 
       load_config
 
+      @cache = Cache.new(@config, file_loader) if cache? || clear_cache?
+
+      if clear_cache?
+        if cache.cache_dir_exists?
+          cache.clear
+          success!("cache directory cleared")
+        else
+          failure!("cache directory doesn't exist, skipping deletion.")
+        end
+      end
+
       if !@files.empty? && lint_files.empty?
         if allow_no_files?
           success!("no files found...\n")
@@ -65,7 +81,7 @@ def run(args = ARGV)
       lint_files.each do |filename|
         runner.clear_offenses
         begin
-          file_content = run_with_corrections(runner, filename)
+          file_content = run_on_file(runner, filename)
         rescue => e
           @stats.exceptions += 1
           puts "Exception occurred when processing: #{relative_filename(filename)}"
@@ -77,6 +93,8 @@ def run(args = ARGV)
         end
       end
 
+      cache&.close
+
       reporter.show
 
       if stdin? && autocorrect?
@@ -99,13 +117,43 @@ def run(args = ARGV)
 
     private
 
+    attr_reader :cache, :config
+
+    def run_on_file(runner, filename)
+      file_content = read_content(filename)
+
+      if cache? && !autocorrect?
+        run_using_cache(runner, filename, file_content)
+      else
+        file_content = run_with_corrections(runner, filename, file_content)
+      end
+
+      log_offense_stats(runner, filename)
+      file_content
+    end
+
+    def run_using_cache(runner, filename, file_content)
+      if (cache_result_offenses = cache.get(filename, file_content))
+        runner.restore_offenses(cache_result_offenses)
+      else
+        run_with_corrections(runner, filename, file_content)
+        cache.set(filename, file_content, runner.offenses.map(&:to_cached_offense_hash).to_json)
+      end
+    end
+
     def autocorrect?
       @options[:autocorrect]
     end
 
-    def run_with_corrections(runner, filename)
-      file_content = read_content(filename)
+    def cache?
+      @options[:cache]
+    end
+
+    def clear_cache?
+      @options[:clear_cache]
+    end
 
+    def run_with_corrections(runner, filename, file_content)
       7.times do
         processed_source = ERBLint::ProcessedSource.new(filename, file_content)
         runner.run(processed_source)
@@ -127,6 +175,11 @@ def run_with_corrections(runner, filename)
         file_content = corrector.corrected_content
         runner.clear_offenses
       end
+
+      file_content
+    end
+
+    def log_offense_stats(runner, filename)
       offenses_filename = relative_filename(filename)
       offenses = runner.offenses || []
 
@@ -138,8 +191,6 @@ def run_with_corrections(runner, filename)
 
       @stats.processed_files[offenses_filename] ||= []
       @stats.processed_files[offenses_filename] |= offenses
-
-      file_content
     end
 
     def read_content(filename)
@@ -283,6 +334,14 @@ def option_parser
           @options[:enabled_linters] = known_linter_names
         end
 
+        opts.on("--cache", "Enable caching") do |config|
+          @options[:cache] = config
+        end
+
+        opts.on("--clear-cache", "Clear cache") do |config|
+          @options[:clear_cache] = config
+        end
+
         opts.on("--enable-linters LINTER[,LINTER,...]", Array,
           "Only use specified linter", "Known linters are: #{known_linter_names.join(", ")}") do |linters|
           linters.each do |linter|

lib/erb_lint/linter.rb

@@ -30,7 +30,7 @@ def support_autocorrect?
       end
     end
 
-    attr_reader :offenses
+    attr_reader :offenses, :config
 
     # Must be implemented by the concrete inheriting class.
     def initialize(file_loader, config)

lib/erb_lint/offense.rb

@@ -17,6 +17,10 @@ def initialize(linter, source_range, message, context = nil, severity = nil)
       @severity = severity
     end
 
+    def to_cached_offense_hash
+      ERBLint::CachedOffense.new_from_offense(self).to_h
+    end
+
     def inspect
       "#<#{self.class.name} linter=#{linter.class.name} "\
         "source_range=#{source_range.begin_pos}...#{source_range.end_pos} "\
@@ -43,5 +47,21 @@ def line_number
     def column
       source_range.column
     end
+
+    def simple_name
+      linter.class.simple_name
+    end
+
+    def last_line
+      source_range.last_line
+    end
+
+    def last_column
+      source_range.last_column
+    end
+
+    def length
+      source_range.length
+    end
   end
 end

lib/erb_lint/reporters/json_reporter.rb

@@ -56,14 +56,14 @@ def formatted_offenses(offenses)
 
       def format_offense(offense)
         {
-          linter: offense.linter.class.simple_name,
+          linter: offense.simple_name,
           message: offense.message.to_s,
           location: {
             start_line: offense.line_number,
             start_column: offense.column,
-            last_line: offense.source_range.last_line,
-            last_column: offense.source_range.last_column,
-            length: offense.source_range.length,
+            last_line: offense.last_line,
+            last_column: offense.last_column,
+            length: offense.length,
           },
         }
       end

lib/erb_lint/runner.rb

@@ -30,5 +30,9 @@ def clear_offenses
       @offenses = []
       @linters.each(&:clear_offenses)
     end
+
+    def restore_offenses(offenses)
+      @offenses.concat(offenses)
+    end
   end
 end