Rework to embed binary data instead of reading wordlist from storage on start-up

anagram/anagram.go

@@ -8,18 +8,13 @@ import (
 	"github.com/ray1729/wordsearch/util"
 )
 
-type DB interface {
-	FindAnagrams(s string) []string
-	Add(s string)
+type DB map[string][]string
+
+func New() DB {
+	return make(DB)
 }
 
-type HashDBImpl map[string][]string
-
-func New() HashDBImpl {
-	return make(HashDBImpl)
-}
-
-func (db HashDBImpl) Add(s string) {
+func (db DB) Add(s string) {
 	k := toKey(s)
 	db[k] = append(db[k], s)
 }
@@ -37,12 +32,9 @@ func Load(r io.Reader) (DB, error) {
 		db.Add(sc.Text())
 
 	}
-	if err := sc.Err(); err != nil {
-		return nil, err
-	}
-	return db, nil
+	return db, sc.Err()
 }
 
-func (d HashDBImpl) FindAnagrams(s string) []string {
+func (d DB) FindAnagrams(s string) []string {
 	return d[toKey(s)]
 }

cloudfn.go

@@ -1,16 +1,14 @@
 package wordsearch
 
 import (
-	"bufio"
-	"context"
+	"embed"
+	"encoding/gob"
 	"fmt"
 	"log"
 	"net/http"
-	"os"
 	"sort"
 	"text/template"
 
-	"cloud.google.com/go/storage"
 	"github.com/GoogleCloudPlatform/functions-framework-go/functions"
 	"github.com/ray1729/wordsearch/anagram"
 	"github.com/ray1729/wordsearch/match"
@@ -20,37 +18,30 @@ import (
 var anagramDB anagram.DB
 var matchDB match.DB
 
-func initializeDB(ctx context.Context, bucketName, objectName string) error {
-	client, err := storage.NewClient(ctx)
+func initializeDB() error {
+	anagrams, err := fs.Open("data/anagram.bin")
 	if err != nil {
-		return fmt.Errorf("error creating storage client: %v", err)
+		return err
 	}
-	r, err := client.Bucket(bucketName).Object(objectName).NewReader(ctx)
+	defer anagrams.Close()
+	if err := gob.NewDecoder(anagrams).Decode(&anagramDB); err != nil {
+		return err
+	}
+	matches, err := fs.Open("data/match.bin")
 	if err != nil {
-		return fmt.Errorf("error opening gs://%s/%s: %v", bucketName, objectName, err)
+		return err
 	}
-	defer r.Close()
-	anagramDB = anagram.New()
-	matchDB = match.New()
-	sc := bufio.NewScanner(r)
-	for sc.Scan() {
-		s := sc.Text()
-		anagramDB.Add(s)
-		matchDB.Add(s)
-	}
-	if err := sc.Err(); err != nil {
-		return fmt.Errorf("error reading gs://%s/%s: %v", bucketName, objectName, err)
+	defer matches.Close()
+	if err := gob.NewDecoder(matches).Decode(&matchDB); err != nil {
+		return err
 	}
 	return nil
 }
 
 func init() {
-	ctx := context.Background()
-	bucketName := mustGetenv("WORDLIST_BUCKET")
-	objectName := mustGetenv("WORDLIST_PATH")
 	log.Println("Initializing databases")
-	if err := initializeDB(ctx, bucketName, objectName); err != nil {
-		panic(err)
+	if err := initializeDB(); err != nil {
+		log.Fatal(err)
 	}
 	corsHandler := cors.New(cors.Options{
 		AllowedOrigins:   []string{"*"},
@@ -65,14 +56,6 @@ func init() {
 	})
 }
 
-func mustGetenv(s string) string {
-	v := os.Getenv(s)
-	if len(v) == 0 {
-		panic(fmt.Sprintf("environment variable %s not set", s))
-	}
-	return v
-}
-
 func handleFormSubmission(w http.ResponseWriter, r *http.Request) {
 	if err := r.ParseForm(); err != nil {
 		log.Printf("error parsing form: %v", err)
@@ -145,3 +128,6 @@ var resultsTmpl = template.Must(template.New("results").Parse(`
 {{ end }}
 </ul>
 `))
+
+//go:embed data/*
+var fs embed.FS

cmd/generate-data/main.go

@@ -0,0 +1,50 @@
+package main
+
+import (
+	"encoding/gob"
+	"flag"
+	"log"
+	"os"
+	"path/filepath"
+
+	"github.com/ray1729/wordsearch/anagram"
+	"github.com/ray1729/wordsearch/match"
+)
+
+var wordlist = flag.String("wordlist", "", "Path to wordlist")
+var dataDir = flag.String("data", "", "Path to output directory")
+
+func main() {
+	flag.Parse()
+	os.MkdirAll(*dataDir, 0755)
+	words, err := os.Open(*wordlist)
+	if err != nil {
+		log.Fatal(err)
+	}
+	anagramDB, err := anagram.Load(words)
+	if err != nil {
+		log.Fatal(err)
+	}
+	if err := writeData(*dataDir, "anagram.bin", anagramDB); err != nil {
+		log.Fatal(err)
+	}
+
+	words.Seek(0, 0)
+	matchDB, err := match.Load(words)
+	if err != nil {
+		log.Fatal(err)
+	}
+	if err := writeData(*dataDir, "match.bin", matchDB); err != nil {
+		log.Fatal(err)
+	}
+}
+
+func writeData(dirName string, fileName string, data interface{}) error {
+	path := filepath.Join(dirName, fileName)
+	f, err := os.OpenFile(path, os.O_CREATE|os.O_RDWR|os.O_TRUNC, 0644)
+	if err != nil {
+		return err
+	}
+	defer f.Close()
+	return gob.NewEncoder(f).Encode(data)
+}

deploy.sh

@@ -4,6 +4,6 @@ set -e
 
 V=$(git describe --tags)
 
-zip function.zip go.mod go.sum cloudfn.go match/match.go anagram/anagram.go util/util.go
+zip function.zip go.mod go.sum cloudfn.go match/match.go anagram/anagram.go util/util.go data/*.bin
 
 gsutil cp function.zip gs://word-search-1729-assets/cloudfn/${V}/

match/match.go

@@ -7,20 +7,15 @@ import (
 	"github.com/ray1729/wordsearch/util"
 )
 
-type DB interface {
-	FindMatches(s string) []string
-	Add(s string)
-}
-
-type PrefixTreeImpl struct {
+type DB struct {
 	Root *Node
 }
 
-func New() PrefixTreeImpl {
-	return PrefixTreeImpl{Root: &Node{}}
+func New() DB {
+	return DB{Root: &Node{}}
 }
 
-func (db PrefixTreeImpl) Add(s string) {
+func (db DB) Add(s string) {
 	xs := util.LowerCaseAlpha(s)
 	db.Root.add(xs, s)
 }
@@ -57,13 +52,10 @@ func Load(r io.Reader) (DB, error) {
 	for sc.Scan() {
 		db.Add(sc.Text())
 	}
-	if err := sc.Err(); err != nil {
-		return nil, err
-	}
-	return db, nil
+	return db, sc.Err()
 }
 
-func (db PrefixTreeImpl) FindMatches(s string) []string {
+func (db DB) FindMatches(s string) []string {
 	return db.Root.find(util.LowerCaseAlphaOrDot(s))
 }