From d0c5337c0196681961d50990e0dd448dd270d430 Mon Sep 17 00:00:00 2001
From: Alex Yatskov <alex@foosoft.net>
Date: Thu, 28 Jul 2016 21:39:35 -0700
Subject: [PATCH] WIP

---
 edict.go | 58 ++++++++++++++++++++++++++++++++++++++++++++++++++++++--
 main.go  | 18 +++++++++++-------
 2 files changed, 67 insertions(+), 9 deletions(-)

diff --git a/edict.go b/edict.go
index 80b3115..ad58e43 100644
--- a/edict.go
+++ b/edict.go
@@ -22,7 +22,12 @@
 
 package main
 
-import "io"
+import (
+	"encoding/xml"
+	"io"
+	"log"
+	"regexp"
+)
 
 type edictKanji struct {
 	// This element will contain a word or short phrase in Japanese
@@ -228,5 +233,54 @@ type edictEntry struct {
 }
 
 func processEdict(reader io.Reader, writer io.Writer) error {
-	return nil
+	entries, err := loadEdict(reader)
+	log.Print(entries)
+	return err
+}
+
+func loadEdict(reader io.Reader) ([]edictEntry, error) {
+	var (
+		err     error
+		entries []edictEntry
+	)
+
+	decoder := xml.NewDecoder(reader)
+
+	for {
+		token, _ := decoder.Token()
+		if token == nil {
+			break
+		}
+
+		switch startElement := token.(type) {
+		case xml.Directive:
+			directive := token.(xml.Directive)
+			if decoder.Entity, err = parseEntities(&directive); err != nil {
+				return nil, err
+			}
+		case xml.StartElement:
+			if startElement.Name.Local == "entry" {
+				var entry edictEntry
+				if err := decoder.DecodeElement(&entry, &startElement); err != nil {
+					return nil, err
+				}
+
+				entries = append(entries, entry)
+			}
+		}
+	}
+
+	return entries, nil
+}
+
+func parseEntities(d *xml.Directive) (map[string]string, error) {
+	re := regexp.MustCompile("<!ENTITY\\s([0-9\\-A-z]+)\\s\"(.+)\">")
+	matches := re.FindAllStringSubmatch(string(*d), -1)
+
+	entities := make(map[string]string)
+	for _, match := range matches {
+		entities[match[1]] = match[2]
+	}
+
+	return entities, nil
 }
diff --git a/main.go b/main.go
index d185550..feba954 100644
--- a/main.go
+++ b/main.go
@@ -25,10 +25,10 @@ package main
 import (
 	"errors"
 	"fmt"
+	"io"
 	"log"
 	"os"
 	"path"
-	"strings"
 )
 
 func usage() {
@@ -36,6 +36,15 @@ func usage() {
 }
 
 func process(fileFormat, inputFile, outputFile string) error {
+	handlers := map[string]func(io.Reader, io.Writer) error{
+		"edict": processEdict,
+	}
+
+	handler, ok := handlers[fileFormat]
+	if !ok {
+		return errors.New("unrecognized file format")
+	}
+
 	input, err := os.Open(inputFile)
 	if err != nil {
 		return err
@@ -46,12 +55,7 @@ func process(fileFormat, inputFile, outputFile string) error {
 		return err
 	}
 
-	switch strings.ToLower(fileFormat) {
-	case "edict":
-		return processEdict(input, output)
-	default:
-		return errors.New("unrecognized file format")
-	}
+	return handler(input, output)
 }
 
 func main() {