Improve parsing performance by sharing a global parser

serpapi · Sep 1, 2023 · ed38fb7 · ed38fb7
1 parent 18f1c26
commit ed38fb7
Show file tree

Hide file tree

Showing 2 changed files with 29 additions and 12 deletions.
diff --git a/bench/bench.rb b/bench/bench.rb
@@ -15,20 +15,34 @@
 
 html_file = File.join(current_dir, 'coffee.html')
 html = File.open(html_file, "r:UTF-8", &:read)
+html_small = "<div>Hello world</div>" * 50
 
 Benchmark.ips do |x|
   x.warmup = 5
   x.time = 20
-    
-  x.report("Nokolexbor parse") do
+
+  x.report("Nokolexbor parse (#{html.size / 1024} KB)") do
     Nokolexbor::HTML(html)
   end
-  x.report("Nokogiri parse") do
+  x.report("Nokogiri parse (#{html.size / 1024} KB)") do
     Nokogiri::HTML(html)
   end
   x.compare!
 end
 
+Benchmark.ips do |x|
+  x.warmup = 5
+  x.time = 20
+
+  x.report("Nokolexbor parse (#{html_small.size} B)") do
+    Nokolexbor::HTML(html_small)
+  end
+  x.report("Nokogiri parse (#{html_small.size} B)") do
+    Nokogiri::HTML(html_small)
+  end
+  x.compare!
+end
+
 nokolex = Nokolexbor::HTML(html)
 nokogiri = Nokogiri::HTML(html)
 css_selector = 'div.g div[data-ved] a[href]:not([href="#"])'

diff --git a/ext/nokolexbor/nl_document.c b/ext/nokolexbor/nl_document.c
@@ -4,6 +4,8 @@ extern VALUE mNokolexbor;
 extern VALUE cNokolexborNode;
 VALUE cNokolexborDocument;
 
+lxb_html_parser_t *g_parser;
+
 static void
 free_nl_document(lxb_html_document_t *document)
 {
@@ -44,18 +46,19 @@ nl_document_parse(VALUE self, VALUE rb_string_or_io)
   const char *html_c = StringValuePtr(rb_html);
   size_t html_len = RSTRING_LEN(rb_html);
 
-  lxb_html_document_t *document;
-
-  document = lxb_html_document_create();
-  if (document == NULL) {
-    rb_raise(rb_eRuntimeError, "Error creating document");
+  if (g_parser == NULL) {
+    g_parser = lxb_html_parser_create();
+    lxb_status_t status = lxb_html_parser_init(g_parser);
+    if (status != LXB_STATUS_OK) {
+      nl_raise_lexbor_error(status);
+    }
+    g_parser->tree->scripting = true;
   }
 
-  lxb_dom_document_scripting_set(lxb_dom_interface_document(document), true);
+  lxb_html_document_t *document = lxb_html_parse(g_parser, (const lxb_char_t *)html_c, html_len);
 
-  lxb_status_t status = lxb_html_document_parse(document, (const lxb_char_t *)html_c, html_len);
-  if (status != LXB_STATUS_OK) {
-    nl_raise_lexbor_error(status);
+  if (document == NULL) {
+    rb_raise(rb_eRuntimeError, "Error parsing document");
   }
 
   return TypedData_Wrap_Struct(cNokolexborDocument, &nl_document_type, document);