open-uri-cached 0.0.1 → 1.0.0

checksums.yaml

@@ -0,0 +1,7 @@
+---
+SHA256:
+  metadata.gz: 1cf2ef137bca3a2593b5adc10d8baee89ec92bd28fec595f69dde388e6fbc0d5
+  data.tar.gz: c22f7c6d400194f783623f3b38cb38c45b10c3c88586b9fcf866339eca1d2b13
+SHA512:
+  metadata.gz: 5d6511c368e3d2be307f5aa9ee64b49750437e0771e5f7e0649b381d6022a20733b1e9f03bea86a388f08e3c5d243ccc2d0fd8fbb8c28a5e0b47cada6f15404a
+  data.tar.gz: b0848773a410b08fa05e6546b12b9316ef45658f71b0b0f4558dff734cc5bbb586a0bf80dd785571e0a6b70ffc14e8cea02b35329ece7f4d3c51f6d0715e4019

data/{README.markdown → README.md}

@@ -11,4 +11,10 @@ Require the library
 
 ## Configuring
 
-OpenURI::Cache.cache_path = '/tmp/open-uri'
+`OpenURI::Cache.cache_path = '/tmp/open-uri'`
+
+## Invalidating
+
+`OpenURI::Cache.invalidate('https://example.com/')`
+
+`OpenURI::Cache.invalidate_all!`

data/lib/open-uri/cached.rb

@@ -1,18 +1,14 @@
-require 'open-uri'
 require 'digest/sha1'
+require 'fileutils'
+require 'open-uri'
+require 'yaml'
 
 module OpenURI
   class << self
     alias original_open_uri open_uri #:nodoc:
     def open_uri(uri, *rest, &block)
-      response = Cache.get(uri.to_s)
-
-      unless response
-        response = original_open_uri(uri, *rest).read
-        Cache.set(uri.to_s, response)
-      end
-
-      response = StringIO.new(response)
+      response = Cache.get(uri.to_s) ||
+                 Cache.set(uri.to_s, original_open_uri(uri, *rest))
 
       if block_given?
         begin
@@ -27,19 +23,104 @@ module OpenURI
   end
 
   class Cache
-    @cache_path = '/tmp/open-uri'
+    @cache_path = "/tmp/open-uri-#{Process.uid}"
 
     class << self
+      attr_accessor :cache_path
+
+      ##
+      # Retrieve file content and meta data from cache
+      # @param [String] key
+      # @return [StringIO]
       def get(key)
         filename = filename_from_url(key)
         # TODO: head request to determine last_modified vs file modtime
-        File.exists?(filename) ? File.read(filename) : nil
+
+        # Read metadata, if it exists
+        if File.exist?("#{filename}.meta")
+          if YAML.respond_to?(:unsafe_load)
+            meta = YAML.unsafe_load(File.read("#{filename}.meta"))
+          else
+            meta = YAML.load(File.read("#{filename}.meta"))
+          end
+        end
+
+        f = File.exist?(filename) ? StringIO.new(File.open(filename, "rb") { |fd| fd.read }) : nil
+
+        # Add meta accessors
+        if meta && f
+          f.instance_variable_set(:"@meta", meta)
+
+          def f.meta
+            @meta
+          end
+          def f.base_uri
+            @meta[:base_uri]
+          end
+          def f.content_type
+            @meta[:content_type]
+          end
+          def f.charset
+            @meta[:charset]
+          end
+          def f.content_encoding
+            @meta[:content_encoding]
+          end
+          def f.last_modified
+            @meta[:last_modified]
+          end
+          def f.status
+            @meta[:status]
+          end
+        end
+
+        f
       end
 
+      # Cache file content and metadata
+      # @param [String] key
+      #   URL of content to be cached
+      # @param [StringIO] value
+      #   value to be cached, typically StringIO returned from `original_open_uri`
+      # @return [StringIO]
+      #   Returns value
       def set(key, value)
         filename = filename_from_url(key)
         mkpath(filename)
-        File.open(filename, 'w'){|f| f.write value }
+
+        # Save metadata in a parallel file
+        if value.respond_to?(:meta)
+          filename_meta = "#{filename}.meta"
+          meta = value.meta
+          meta[:status] = value.status if value.respond_to?(:status)
+          meta[:content_type] = value.content_type if value.respond_to?(:content_type)
+          meta[:base_uri] = value.base_uri if value.respond_to?(:base_uri)
+          File.open(filename_meta, 'wb') {|f| YAML::dump(meta, f)}
+        end
+
+        # Save file contents
+        File.open(filename, 'wb'){|f| f.write value.read }
+        value.rewind
+        value
+      end
+
+      # Invalidate cache for a key, optionally if older than time givan
+      # @param [String] key
+      #   URL of content to be invalidated
+      # @param [Time] time
+      #   (optional): the maximum age at which the cached value is still acceptable
+      # @return
+      #   Returns 1 if a cached value was invalidated, false otherwise
+      def invalidate(key, time = Time.now)
+        filename = filename_from_url(key)
+        File.delete(filename) if File.stat(filename).mtime < time
+      rescue Errno::ENOENT
+        false
+      end
+
+      # Invalidate all caches we know about
+      def invalidate_all!
+        FileUtils.rm_r(@cache_path, force: true, secure: true)
       end
 
       protected
@@ -55,7 +136,7 @@ module OpenURI
             full.push(dir)
             dir = full.join('/')
             next if dir.to_s == ''
-            Dir.mkdir(dir) unless File.exists?(dir)
+            Dir.mkdir(dir) unless File.exist?(dir)
           end
         end
     end

metadata

@@ -1,57 +1,74 @@
---- !ruby/object:Gem::Specification 
+--- !ruby/object:Gem::Specification
 name: open-uri-cached
-version: !ruby/object:Gem::Version 
-  version: 0.0.1
+version: !ruby/object:Gem::Version
+  version: 1.0.0
 platform: ruby
-authors: 
+authors:
 - Danial Pearce
-autorequire: 
+autorequire:
 bindir: bin
 cert_chain: []
-
-date: 2010-02-27 00:00:00 +11:00
-default_executable: 
-dependencies: []
-
+date: 2021-12-28 00:00:00.000000000 Z
+dependencies:
+- !ruby/object:Gem::Dependency
+  name: rspec
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: 3.10.0
+  type: :development
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: 3.10.0
+- !ruby/object:Gem::Dependency
+  name: webmock
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: 3.14.0
+  type: :development
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: 3.14.0
 description: OpenURI with transparent disk caching
-email: rubygems@tigris.id.au
+email: danial.pearce@gmail.com
 executables: []
-
 extensions: []
-
 extra_rdoc_files: []
-
-files: 
-- README.markdown
+files:
 - LICENSE
+- README.md
 - lib/open-uri/cached.rb
-has_rdoc: true
-homepage: http://github.com/tigris/open-uri-cached
+homepage: https://github.com/tigris/open-uri-cached
 licenses: []
-
-post_install_message: 
+metadata: {}
+post_install_message:
 rdoc_options: []
-
-require_paths: 
+require_paths:
 - lib
-required_ruby_version: !ruby/object:Gem::Requirement 
-  requirements: 
+required_ruby_version: !ruby/object:Gem::Requirement
+  requirements:
   - - ">="
-    - !ruby/object:Gem::Version 
-      version: "0"
-  version: 
-required_rubygems_version: !ruby/object:Gem::Requirement 
-  requirements: 
+    - !ruby/object:Gem::Version
+      version: '0'
+required_rubygems_version: !ruby/object:Gem::Requirement
+  requirements:
   - - ">="
-    - !ruby/object:Gem::Version 
-      version: "0"
-  version: 
+    - !ruby/object:Gem::Version
+      version: '0'
 requirements: []
-
-rubyforge_project: 
-rubygems_version: 1.3.5
-signing_key: 
-specification_version: 3
-summary: Do a lot of site scraping but take lots of attempts at parsing the content before reaching your end result? This gem is for you. But wait, there's more... Ok, no there isn't.
+rubygems_version: 3.2.22
+signing_key:
+specification_version: 4
+summary: Do a lot of site scraping but take lots of attempts at parsing the content
+  before reaching your end result? This gem is for you. But wait, there's more...
+  Ok, no there isn't.
 test_files: []
-