updated docuemntation and added readme

LICENSE

@@ -0,0 +1,12 @@
+Copyright (c) 2016, Redis Labs, Inc.
+All rights reserved.
+
+Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met:
+
+1. Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer.
+
+2. Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution.
+
+3. Neither the name of the copyright holder nor the names of its contributors may be used to endorse or promote products derived from this software without specific prior written permission.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.

README.md

@@ -0,0 +1,68 @@
+# RediSearchBenchmarks
+
+Source code for benchmarking the RediSearch module, providing scalable high performance full-text search.
+
+## What's in here
+
+This is a Go application that can ingest data into the search engine, and benchmark the throughput and latency of running these benchmarks.
+
+It supports reading [Wikipedia Abstract Data Dumps](https://dumps.wikimedia.org/enwiki/latest/enwiki-latest-abstract.xml) and indexing them, in three search engines: 
+
+* [RediSearch](https://github.com/RedisLabsModules/RediSearch)
+* [ElasticSearch](https://www.elastic.co/)
+* [Solr](http://lucene.apache.org/solr/)
+
+## Benchmark output
+
+For each benchmark, we append a single line to a CSV file, with the engine used, benchmark type, query, concurrency, throughput and latency.
+
+The default file name is `benchmark.csv`, and running the app with `-o -` will result in the result printed to stdout.
+
+The output for running a benchmark on the queries "foo,bar,baz" with 4 concurrent clients, looks like this:
+
+```
+redis,"search: foo,bar,baz",4,14997.81,0.27
+``` 
+
+## Usage
+
+```
+Usage of ./RediSearchBenchmark:
+  -benchmark string
+    	[search|suggest] - if set, we run the given benchmark
+  -c int
+    	benchmark concurrency (default 4)
+  -duration int
+    	number of seconds to run the benchmark (default 5)
+  -engine string
+        [redis|elastic|solr] The search backend to run (default "redis")
+  -file string
+    	Input file to ingest data from (wikipedia abstracts)
+  -fuzzy
+    	For redis only - benchmark fuzzy auto suggest
+  -hosts string
+    	comma separated list of host:port to redis nodes (default "localhost:6379")
+  -o string
+    	results output file. set to - for stdout (default "benchmark.csv")
+  -queries string
+    	comma separated list of queries to benchmark (default "hello world")
+  -scores string
+    	read scores of documents CSV for indexing
+  -shards int
+    	the number of partitions we want (AT LEAST the number of cluster shards) (default 1)
+```
+
+## Example: Indexing documents into RediSearch
+
+```
+./RediSearchBenchmark -engine redis -shards 4 -hosts "localhost:6379,localhost:6380,localhost:6381,localhost:6382" \
+    -file ~/wiki/enwiki-20160305-abstract.xml -scores ~/wiki/scores.csv
+```
+
+## Example: Benchmarking RediSearch with 32 concurrent clients
+
+```
+./RediSearchBenchmark -engine redis -shards 4 -hosts "localhost:6379,localhost:6380,localhost:6381,localhost:6382" \
+    -benchmark search -queries "hello world,foo bar" -c 32 -o out.csv
+```
+

benchmark.go

@@ -7,12 +7,15 @@ import (
 	"math/rand"
 	"os"
 	"sync"
+	"sync/atomic"
 	"time"
 
 	"github.com/RedisLabs/RediSearchBenchmark/index"
 	"github.com/RedisLabs/RediSearchBenchmark/query"
 )
 
+// SearchBenchmark returns a closure of a function for the benchmarker to run, using a given index
+// and options, on a set of queries
 func SearchBenchmark(queries []string, idx index.Index, opts interface{}) func() error {
 
 	counter := 0
@@ -25,6 +28,8 @@ func SearchBenchmark(queries []string, idx index.Index, opts interface{}) func()
 
 }
 
+// AutocompleteBenchmark returns a configured autocomplete benchmarking function to be run by
+// the benchmarker
 func AutocompleteBenchmark(ac index.Autocompleter, fuzzy bool) func() error {
 	counter := 0
 	sz := len(prefixes)
@@ -34,6 +39,13 @@ func AutocompleteBenchmark(ac index.Autocompleter, fuzzy bool) func() error {
 		return err
 	}
 }
+
+// Benchmark runs a given function f for the given duration, and outputs the throughput and latency of the function.
+//
+// It receives metadata like the engine we are running and the title of the specific benchmark, and writes these along
+// with the results to a CSV file given by outfile.
+//
+// If outfile is "-" we write the result to stdout
 func Benchmark(concurrency int, duration time.Duration, engine, title string, outfile string, f func() error) {
 
 	var out io.WriteCloser
@@ -48,14 +60,12 @@ func Benchmark(concurrency int, duration time.Duration, engine, title string, ou
 		defer out.Close()
 	}
 
-	//	queries = []string{"weezer", "germany", "a", "music", "music of the spheres", "abba", "queen",
-	//		"nirvana", "benjamin netanyahu", "redis", "redis labs", "german history"} // "computer science", "machine learning"}
-	//queries := []string{"earth Though is", "school etc"}
 	startTime := time.Now()
-	totalTime := time.Duration(0)
+	// the total time it took to run the functions, to measure average latency, in nanoseconds
+	var totalTime uint64
+	var total uint64
 	wg := sync.WaitGroup{}
 
-	total := 0
 	end := time.Now().Add(duration)
 
 	for i := 0; i < concurrency; i++ {
@@ -68,9 +78,10 @@ func Benchmark(concurrency int, duration time.Duration, engine, title string, ou
 				if err = f(); err != nil {
 					panic(err)
 				}
-				total++
 
-				totalTime += time.Since(tst)
+				// update the total requests performed and total time
+				atomic.AddUint64(&total, 1)
+				atomic.AddUint64(&totalTime, uint64(time.Since(tst)))
 
 			}
 			wg.Done()
@@ -81,6 +92,7 @@ func Benchmark(concurrency int, duration time.Duration, engine, title string, ou
 	avgLatency := (float64(totalTime) / float64(total)) / float64(time.Millisecond)
 	rate := float64(total) / (float64(time.Since(startTime)) / float64(time.Second))
 
+	// Output the results to CSV
 	w := csv.NewWriter(out)
 
 	err = w.Write([]string{engine, title,

index/document.go

@@ -4,12 +4,15 @@ import (
 	"sort"
 )
 
+// Document represents a single document to be indexed or returned from a query.
+// Besides a score and id, the Properties are completely arbitrary
 type Document struct {
 	Id         string
 	Score      float32
 	Properties map[string]interface{}
 }
 
+// NewDocument creates a document with the specific id and score
 func NewDocument(id string, score float32) Document {
 	return Document{
 		Id:         id,
@@ -18,6 +21,7 @@ func NewDocument(id string, score float32) Document {
 	}
 }
 
+// Set sets a property and its value in the document
 func (d Document) Set(name string, value interface{}) Document {
 	d.Properties[name] = value
 	return d
@@ -30,6 +34,7 @@ func (l DocumentList) Len() int           { return len(l) }
 func (l DocumentList) Swap(i, j int)      { l[i], l[j] = l[j], l[i] }
 func (l DocumentList) Less(i, j int) bool { return l[i].Score > l[j].Score } //reverse sorting
 
+// Sort the DocumentList
 func (l DocumentList) Sort() {
 	sort.Sort(l)
 }

index/elastic/elastic.go

@@ -2,6 +2,7 @@ package elastic
 
 import (
 	"encoding/json"
+	"errors"
 	"net/http"
 	"time"
 
@@ -10,18 +11,21 @@ import (
 	"gopkg.in/olivere/elastic.v3"
 )
 
+// Index is an ElasticSearch index
 type Index struct {
 	conn *elastic.Client
 
 	md   *index.Metadata
 	name string
+	typ  string
 }
 
-func NewIndex(addr, name string, md *index.Metadata) (*Index, error) {
+// NewIndex creates a new elasticSearch index with the given address and name. typ is the entity type
+func NewIndex(addr, name, typ string, md *index.Metadata) (*Index, error) {
 
 	client := &http.Client{
 		Transport: &http.Transport{
-			MaxIdleConnsPerHost: 500,
+			MaxIdleConnsPerHost: 200,
 		},
 		Timeout: 250 * time.Millisecond,
 	}
@@ -33,47 +37,65 @@ func NewIndex(addr, name string, md *index.Metadata) (*Index, error) {
 		conn: conn,
 		md:   md,
 		name: name,
+		typ:  typ,
 	}
 
 	return ret, nil
 
 }
 
+type mappingProperty map[string]interface{}
+
+type mapping struct {
+	Properties map[string]mappingProperty `json:"properties"`
+}
+
+// convert a fieldType to elastic mapping type string
+func fieldTypeString(f index.FieldType) (string, error) {
+	switch f {
+	case index.TextField:
+		return "string", nil
+	case index.NumericField:
+		return "double", nil
+	default:
+		return "", errors.New("Unsupported field type")
+	}
+}
+
+// Create creates the index and posts a mapping corresponding to our Metadata
 func (i *Index) Create() error {
 
-	docMapping := `
-	{
-		"mappings": {
-			"doc":{
-				"properties":{
-					"body":{
-						"type":"string"
-					},
-					"title":{
-						"type":"string"
-					}
-					
-				}
-			},
-			"autocomplete":{
-				"properties":{
-					"sugg":{
-						"type":"completion",
-						"payloads":true
-					}
-				}
-			}
+	doc := mapping{Properties: map[string]mappingProperty{}}
+	for _, f := range i.md.Fields {
+		doc.Properties[f.Name] = mappingProperty{}
+		fs, err := fieldTypeString(f.Type)
+		if err != nil {
+			return err
 		}
+		doc.Properties[f.Name]["type"] = fs
 	}
-	`
 
-	_, err := i.conn.CreateIndex(i.name).BodyJson(docMapping).Do()
+	// we currently manually create the autocomplete mapping
+	ac := mapping{
+		Properties: map[string]mappingProperty{
+			"sugg": mappingProperty{
+				"type":     "completion",
+				"payloads": true,
+			},
+		},
+	}
+
+	mappings := map[string]mapping{
+		i.typ:          doc,
+		"autocomplete": ac,
+	}
+
+	_, err := i.conn.CreateIndex(i.name).BodyJson(map[string]interface{}{"mappings": mappings}).Do()
 
 	return err
 }
 
-// Add indexes one entry in the index.
-// TODO: Add support for multiple insertions
+// Index indexes multiple documents
 func (i *Index) Index(docs []index.Document, opts interface{}) error {
 
 	blk := i.conn.Bulk()
@@ -88,6 +110,8 @@ func (i *Index) Index(docs []index.Document, opts interface{}) error {
 	return err
 }
 
+// Search searches the index for the given query, and returns documents,
+// the total number of results, or an error if something went wrong
 func (i *Index) Search(q query.Query) ([]index.Document, int, error) {
 
 	eq := elastic.NewQueryStringQuery(q.Term)
@@ -116,12 +140,14 @@ func (i *Index) Search(q query.Query) ([]index.Document, int, error) {
 	return ret, int(res.TotalHits()), err
 }
 
+// Drop deletes the index
 func (i *Index) Drop() error {
 	i.conn.DeleteIndex(i.name).Do()
 
 	return nil
 }
 
+// AddTerms add suggestion terms to the suggester index
 func (i *Index) AddTerms(terms ...index.Suggestion) error {
 	blk := i.conn.Bulk()
 
@@ -137,6 +163,9 @@ func (i *Index) AddTerms(terms ...index.Suggestion) error {
 	return err
 
 }
+
+// Suggest gets completion suggestions for a given prefix.
+// TODO: fuzzy not supported yet
 func (i *Index) Suggest(prefix string, num int, fuzzy bool) ([]index.Suggestion, error) {
 
 	s := elastic.NewCompletionSuggester("autocomplete").Field("sugg").Text(prefix).Size(num)
@@ -152,7 +181,7 @@ func (i *Index) Suggest(prefix string, num int, fuzzy bool) ([]index.Suggestion,
 
 			ret := make([]index.Suggestion, 0, len(opts))
 			for _, op := range opts {
-				ret = append(ret, index.Suggestion{op.Text, float64(op.Score)})
+				ret = append(ret, index.Suggestion{Term: op.Text, Score: float64(op.Score)})
 			}
 			return ret, nil
 		}

index/elastic/elastic_test.go

@@ -11,12 +11,12 @@ import (
 )
 
 func TestIndex(t *testing.T) {
-	t.SkipNow()
+	//t.SkipNow()
 	// todo: run redisearch automatically
 	md := index.NewMetadata().AddField(index.NewTextField("title", 1.0)).
 		AddField(index.NewNumericField("score"))
 
-	idx, err := NewIndex("http://localhost:9200", "testung", md)
+	idx, err := NewIndex("http://localhost:9200", "testung", "doc", md)
 	assert.NoError(t, err)
 	assert.NoError(t, idx.Drop())
 	assert.NoError(t, idx.Create())
@@ -46,10 +46,11 @@ func TestIndex(t *testing.T) {
 }
 
 func TestSuggest(t *testing.T) {
+
 	md := index.NewMetadata().AddField(index.NewTextField("title", 1.0)).
 		AddField(index.NewNumericField("score"))
 
-	idx, err := NewIndex("http://localhost:9200", "testung", md)
+	idx, err := NewIndex("http://localhost:9200", "testung", "doc", md)
 	assert.NoError(t, err)
 	assert.NoError(t, idx.Drop())
 	assert.NoError(t, idx.Create())

index/index.go

@@ -4,6 +4,8 @@ import (
 	"github.com/RedisLabs/RediSearchBenchmark/query"
 )
 
+// Index is the abstract representation of a search index we're working against.
+// It is implemented for redisearch, elasticserch and solr.
 type Index interface {
 	Index(documents []Document, options interface{}) error
 	Search(query.Query) (docs []Document, total int, err error)