Xapian Benchmark results

Krishna Bharadwaj <[email protected]> Fri, 30 Nov 2018 14:28:19 -0600

Newsgroups	gmane.comp.search.xapian.general
Message-ID	<CAGqjorT7bOLnM=bmAdb9-BmMWmcAgpFnjemJ22GYWTi6V3Arkg@mail.gmail.com>

Hi,

I am currently trying to benchmark a multithreaded xapian implementation on
a chameleon baremetal instance written in C++. My workload is a 3 Gig
wikipedia xml dump consisting of ~286 file of different sizes. My results
are showing me that indexing on xapian is an order of magnitude faster than
my lucene and lucene plusplus implementations. This is a result that I did
not expect. Just want to confirm with you guys if my
implementation(attached below) is correct. I notice that I am getting the
search results correctly but I am just not able to get over the fact that
xapian is performing so much faster than my other implementations. There
have been no optimizations in my code. Awaiting your response.

Regards,
Krishna Bharadwaj

xapianBenchmark.cpp (text/plain, 4.2 KB)


#define NUM_THREADS 4

void parallel_index(int tid, int num_threads,vector<string> documents,Xapian::WritableDatabase *db )
{
    int i =0;
     char line[2048];
     string line_string;
     Xapian::TermGenerator indexer;
	Xapian::Stem stemmer("english");
	
    indexer.set_stemmer(stemmer);
  //  doc.set_data("content");
    indexer.set_stemming_strategy(Xapian::TermGenerator::STEM_SOME);
   Xapian::WritableDatabase database = *db;
   try{
   for(i=0;i<documents.size();++i)
   {    
      
       ifstream file;
       file.open(documents[i]);
       line_string = "";
       while(file.good())
       {
            
                       file.getline(line, sizeof(line));
                       string str(line);
                       if(!str.empty())
                       {
                        line_string += string(line);
                       }
       }
       
       Xapian::Document doc;
		doc.set_data(line_string);

		indexer.set_document(doc);
		indexer.index_text(line_string);
       database.add_document(doc); 

       file.close();
   }
   database.commit();
   cout <<"count::" <<database.get_doccount() <<endl;
   }
   catch (const Xapian::Error &e) {
    cout << e.get_description() << endl;
    exit(1);
}
   
   
}


int main(int argc, char **argv)
{
    vector<thread> threads;
    int doc_ct = 0;
    char queries[][20] = {"Probation","Volleyball"};
   int th = 0;
   vector<string> documents[NUM_THREADS];
    Xapian::WritableDatabase dbs[NUM_THREADS];
   for(th=0;th<NUM_THREADS;++th)
   {
       //string s = argv[1] + to_string("\\")+to_string(th);
       char s[30] ;
       strcpy(s,argv[2]);
       strcat(s,"/DB");
       string idx = to_string(th);
       strcat(s,idx.c_str());
       
       cout << s << endl;
        dbs[th] =  Xapian::WritableDatabase(s,Xapian::DB_CREATE_OR_OVERWRITE, 512);
   }
   

    Xapian::TermGenerator indexer;
	Xapian::Stem stemmer("english");
	Xapian::Document doc;
	
   
    string line_string;
    char filename[100];
     ifstream infs;
    char *tok, filepath[1536], content[2048], line[2048];
    int i;
	high_resolution_clock::time_point start, end;
	infs.open(argv[1], ifstream::in);
	indexer.set_stemmer(stemmer);
    
    start = high_resolution_clock::now();  
    doc.set_data("content");
                indexer.set_document(doc);
                line_string = "";
    
    while(infs.good())
    {
        doc_ct = doc_ct + 1;
        infs.getline(filename, sizeof(filename));
        documents[doc_ct%NUM_THREADS].push_back(filename);
        
    }
    
      for (int tid = 0; tid < NUM_THREADS; tid++) {
            threads.push_back(thread(parallel_index, tid, NUM_THREADS, 
                    documents[tid],&dbs[tid]));
        }
        
        for (auto& th : threads) {
            th.join();
        }

   
    infs.close();
    Xapian::Database final_db;
     for(const Xapian::WritableDatabase &d : dbs)
    {
        final_db.add_database(d);
    } 
    Xapian::Enquire enquire(final_db);
	Xapian::QueryParser parser;
    end = high_resolution_clock::now();  
    auto diff1 = duration_cast<milliseconds>(end - start).count();
	cout << "Indexing time: " << diff1 << "ms" << endl;
    // search
    parser.set_stemmer(stemmer);
		parser.set_database(final_db);
		parser.set_stemming_strategy(Xapian::QueryParser::STEM_SOME);

		start = high_resolution_clock::now();
        ifstream searchFile;
        searchFile.open("test.txt");
        char word[40];
		while(searchFile.good()) {
            searchFile.getline(word, sizeof(word));
			cout << "Searching : " <<  word<< endl ;
			Xapian::Query query = parser.parse_query(word);
			enquire.set_query(query);
			
			
			Xapian::MSet hits = enquire.get_mset(0, 10);
			
	//cout << "Parsed query is: " << query.get_description() << endl;

  
    
    // Display the results.
    cout << hits.get_matches_estimated() << " results found.\n";
    cout << "Matches 1-" << hits.size() << ":\n" << endl;
			
		}
		
		end = high_resolution_clock::now();
		
		auto diff = duration_cast<milliseconds>(end - start).count();
		cout << "Average search time: " << diff*1000 << "ns" << endl;
	
    }