refactoring

.gitignore

@@ -1,6 +1,6 @@
 .idea/
 vendor/
-example/*.txt
+*.txt
 
 text8
 text8.zip

README.md

@@ -68,7 +68,7 @@ Wego outputs a .txt file that is described word vector is subject to the followi
 <word> <value1> <value2> ...
 ```
 
-## Example
+## API
 
 It's also able to train word vectors using wego APIs. Examples are as follows.
 
@@ -78,41 +78,28 @@ package main
 import (
 	"os"
 
-	"github.com/ynqa/wego/pkg/builder"
+	"github.com/ynqa/wego/pkg/model/modelutil/save"
 	"github.com/ynqa/wego/pkg/model/word2vec"
 )
 
 func main() {
-	b := builder.NewWord2vecBuilder()
-
-	b.Dimension(10).
-		Window(5).
-		Model(word2vec.CBOW).
-		Optimizer(word2vec.NEGATIVE_SAMPLING).
-		NegativeSampleSize(5).
-		Verbose()
-
-	m, err := b.Build()
+	model, err := word2vec.New(
+		word2vec.WithWindow(5),
+		word2vec.WithModel(word2vec.Cbow),
+		word2vec.WithOptimizer(word2vec.NegativeSampling),
+		word2vec.WithNegativeSampleSize(5),
+		word2vec.Verbose(),
+	)
 	if err != nil {
-		// Failed to build word2vec.
+		// failed to create word2vec.
 	}
 
 	input, _ := os.Open("text8")
-
-	// Start to Train.
-	if err = m.Train(input); err != nil {
-		// Failed to train by word2vec.
+	if err = model.Train(input); err != nil {
+		// failed to train.
 	}
 
-	output, err := os.Create("example.txt")
-	if err != nil {
-		// Failed to create output file.
-	}
-
-	defer func() {
-		output.Close()
-	}()
-
-	m.Save(output)
+	// write word vector.
+	model.Save(os.Stdin, save.AggregatedVector)
 }
 ```

pkg/model/README.md → cmd/README.md

@@ -102,3 +102,86 @@ Flags:
       --verbose             verbose mode
   -w, --window int          context window size (default 5)
 ```
+
+# Search
+
+Similarity search between word vectors.
+
+## Usage
+
+```
+Search similar words
+
+Usage:
+  wego search [flags]
+
+Examples:
+  wego search -i example/word_vectors.txt microsoft
+
+Flags:
+  -h, --help               help for search
+  -i, --inputFile string   input file path for trained word vector (default "example/input.txt")
+  -r, --rank int           how many the most similar words will be displayed (default 10)
+```
+
+## Example
+
+```
+$ go run wego.go search -i example/word_vectors_sg.txt microsoft
+  RANK |    WORD    | SIMILARITY
++------+------------+------------+
+     1 | apple      |   0.994008
+     2 | operating  |   0.992855
+     3 | versions   |   0.992800
+     4 | ibm        |   0.992232
+     5 | os         |   0.989174
+     6 | computers  |   0.988998
+     7 | machines   |   0.988804
+     8 | dvd        |   0.988732
+     9 | cd         |   0.988259
+    10 | compatible |   0.988200
+```
+
+# REPL for search
+
+Similarity search between word vectors with REPL mode.
+
+## Usage
+
+```
+Search similar words with REPL mode
+
+Usage:
+  wego repl [flags]
+
+Examples:
+  wego repl -i example/word_vectors.txt
+  >> apple + banana
+  ...
+
+Flags:
+  -h, --help               help for repl
+  -i, --inputFile string   input file path for trained word vector (default "example/word_vectors.txt")
+  -r, --rank int           how many the most similar words will be displayed (default 10)
+```
+
+## Example
+
+Now, it is able to use `+`, `-` for arithmetic operations.
+
+```
+$ go run wego.go repl -i example/word_vectors_sg.txt
+>> a + b
+  RANK |  WORD   | SIMILARITY
++------+---------+------------+
+     1 | phi     |   0.907975
+     2 | q       |   0.904593
+     3 | mathbf  |   0.903066
+     4 | cdot    |   0.902205
+     5 | b       |   0.901952
+     6 | becomes |   0.900346
+     7 | int     |   0.898680
+     8 | z       |   0.897895
+     9 | named   |   0.896480
+    10 | v       |   0.895456
+```