Merge pull request #140 from rubycdp/137-mhtml

Implement mhtml

Author: route

README.md

@@ -372,6 +372,18 @@ browser.go_to("https://google.com/")
 browser.pdf(path: "google.pdf", paper_width: 1.0, paper_height: 1.0) # => 14983
 ```
 
+#### mhtml(\*\*options) : `String` | `Integer`
+
+Saves MHTML on a disk or returns it as a string.
+
+* options `Hash`
+  * :path `String` to save a file on the disk.
+
+```ruby
+browser.go_to("https://google.com/")
+browser.mhtml(path: "google.mhtml") # => 87742
+```
+
 
 ## Network
 

lib/ferrum/browser.rb

@@ -22,7 +22,7 @@ class Browser
                 body doctype set_content
                 headers cookies network
                 mouse keyboard
-                screenshot pdf viewport_size
+                screenshot pdf mhtml viewport_size
                 frames frame_by main_frame
                 evaluate evaluate_on evaluate_async execute
                 add_script_tag add_style_tag bypass_csp

lib/ferrum/page/screenshot.rb

@@ -29,14 +29,24 @@ def screenshot(**opts)
         options = screenshot_options(path, **opts)
         data = capture_screenshot(options, opts[:full])
         return data if encoding == :base64
-        save_file(path, data)
+
+        bin = Base64.decode64(data)
+        save_file(path, bin)
       end
 
       def pdf(**opts)
         path, encoding = common_options(**opts)
         options = pdf_options(**opts)
         data = command("Page.printToPDF", **options).fetch("data")
         return data if encoding == :base64
+
+        bin = Base64.decode64(data)
+        save_file(path, bin)
+      end
+
+      def mhtml(path: nil)
+        data = command("Page.captureSnapshot", format: :mhtml).fetch("data")
+        return data if path.nil?
         save_file(path, data)
       end
 
@@ -56,9 +66,8 @@ def document_size
       private
 
       def save_file(path, data)
-        bin = Base64.decode64(data)
-        return bin unless path
-        File.open(path.to_s, "wb") { |f| f.write(bin) }
+        return data unless path
+        File.open(path.to_s, "wb") { |f| f.write(data) }
       end
 
       def common_options(encoding: :base64, path: nil, **_)

spec/screenshot_spec.rb

@@ -87,7 +87,6 @@ def create_screenshot(**options)
         end
 
         after do
-          FileUtils.rm_f("#{PROJECT_ROOT}/spec/tmp/screenshot.pdf")
           FileUtils.rm_f("#{PROJECT_ROOT}/spec/tmp/screenshot.png")
         end
 
@@ -206,6 +205,56 @@ def create_screenshot(**options)
           include_examples "when scale is set"
         end
 
+        include_examples "screenshot screen"
+
+        context "when encoding is base64" do
+          let(:file) { "#{PROJECT_ROOT}/spec/tmp/screenshot.#{format}" }
+
+          def create_screenshot(path: file, **options)
+            image = browser.screenshot(format: format, encoding: :base64, **options)
+            File.open(file, "wb") { |f| f.write Base64.decode64(image) }
+          end
+
+          it "defaults to base64 when path isn't set" do
+            browser.go_to
+
+            screenshot = browser.screenshot(format: format)
+
+            expect(screenshot.length).to be > 100
+          end
+
+          it "supports screenshotting the page in base64" do
+            browser.go_to
+
+            screenshot = browser.screenshot(encoding: :base64)
+
+            expect(screenshot.length).to be > 100
+          end
+
+          context "png" do
+            let(:format) { :png }
+            after { FileUtils.rm_f(file) }
+
+            include_examples "screenshot screen"
+          end
+
+          context "jpeg" do
+            let(:format) { :jpeg }
+            after { FileUtils.rm_f(file) }
+
+            include_examples "screenshot screen"
+          end
+        end
+      end
+
+      describe "#pdf" do
+        let(:format) { :pdf }
+        let(:file) { "#{PROJECT_ROOT}/spec/tmp/screenshot.#{format}" }
+
+        after do
+          FileUtils.rm_f("#{PROJECT_ROOT}/spec/tmp/screenshot.pdf")
+        end
+
         context "when :paper_width and :paper_height are set" do
           it "changes pdf size" do
             browser.go_to("/ferrum/long_page")
@@ -280,46 +329,37 @@ def create_screenshot(**options)
                                     transfer_mode: "ReturnAsBase64")
           end
         end
+      end
 
-        include_examples "screenshot screen"
-
-        context "when encoding is base64" do
-          let(:file) { "#{PROJECT_ROOT}/spec/tmp/screenshot.#{format}" }
-
-          def create_screenshot(path: file, **options)
-            image = browser.screenshot(format: format, encoding: :base64, **options)
-            File.open(file, "wb") { |f| f.write Base64.decode64(image) }
-          end
-
-          it "defaults to base64 when path isn't set" do
-            browser.go_to
-
-            screenshot = browser.screenshot(format: format)
-
-            expect(screenshot.length).to be > 100
-          end
+      describe "#mhtml" do
+        let(:format) { :mhtml }
+        let(:file) { "#{PROJECT_ROOT}/spec/tmp/screenshot.#{format}" }
 
-          it "supports screenshotting the page in base64" do
-            browser.go_to
+        after do
+          FileUtils.rm_f("#{PROJECT_ROOT}/spec/tmp/screenshot.mhtml")
+        end
 
-            screenshot = browser.screenshot(encoding: :base64)
+        it "returns data" do
+          browser.go_to("/ferrum/simple")
 
-            expect(screenshot.length).to be > 100
-          end
+          data = browser.mhtml
 
-          context "png" do
-            let(:format) { :png }
-            after { FileUtils.rm_f(file) }
+          expect(data).to match(/\/ferrum\/simple/)
+          expect(data).to match(/mhtml.blink/)
+          expect(data).to match(/\<\!DOCTYPE html\>/)
+          expect(data).to match(/Foo\<br\>Bar/)
+        end
 
-            include_examples "screenshot screen"
-          end
+        it "saves a file" do
+          browser.go_to("/ferrum/simple")
 
-          context "jpeg" do
-            let(:format) { :jpeg }
-            after { FileUtils.rm_f(file) }
+          browser.mhtml(path: file)
 
-            include_examples "screenshot screen"
-          end
+          content = File.read(file)
+          expect(content).to match(/\/ferrum\/simple/)
+          expect(content).to match(/mhtml.blink/)
+          expect(content).to match(/\<\!DOCTYPE html\>/)
+          expect(content).to match(/Foo\<br\>Bar/)
         end
       end
     end