spaCy-Go

spacy-go is Golang interface for accessing linguistic annotations provided by spaCy using Google's gRPC. This module only supports basic functionalities like loading language models, linguistic annotation and similarity for text sentences.

Installation

Installing Golang library

spacy-go Golang library can be installed by following single command.

go get -v "github.com/yash1994/spacy-go"

Setting up python gRPC server

The $GOPATH environment variable lists places for Go to look for Go Workspaces. By default, Go assumes our GOPATH location is at $HOME/go, where $HOME is the root directory of our user account on our computer.

Before importing the golang library, these commands need to be executed (inside source package) to spin up the python gRPC server.

pip install -r $GOPATH/src/github.com/yash1994/spacy-go/requirements.txt

Install spacy language models with following command.

python3 -m spacy download en_core_web_sm

Connection between client and server is secured by TLS/SSL authentication. Use following command to generate unique pair of root certificate that is used to authenticate the server and private key that only the server has access to.

openssl req -newkey rsa:2048 -nodes -keyout $GOPATH/src/github.com/yash1994/spacy-go/server.key -x509 -days 365 -out $GOPATH/src/github.com/yash1994/spacy-go/server.crt -subj "/CN=localhost"

The following command will spin up python gRPC server at localhost:50051.

python3 $GOPATH/src/github.com/yash1994/spacy-go/api/server.py &

Usage

package main

import (
	"fmt"

	spacygo "github.com/yash1994/spacy-go"
)

func main() {

	// load language model
	var modelName string = "en_core_web_sm"
	r, err := spacygo.Load(modelName)

	if err != nil {
		return
	}

	fmt.Printf("%v \n", r.GetMessage())

	// annotate text
	var text string = "I propose to consider the question, 'Can machines think?"

	doc, err := spacygo.Nlp(text)

	// print annotated info : part-of-speech
	for i, token := range doc.GetTokens() {
		fmt.Printf("token %v '%v' part-of-speech tag: %v \n", i, token.GetText(), token.GetPos())
	}

	// calculate text similarity
	var texta string = "I like apples"
	var textb string = "I like oranges"

	textSimilarity, err := spacygo.Similarity(texta, textb)

	fmt.Printf("text similarity between %v and %v is %v", texta, textb, textSimilarity.GetSimilarity())
}

:dizzy: APIs

ToDos

• [x] Extensive Test cases
• [x] Error handling server side
• [x] Add SSL and auth
• [x] API Docs
• [x] Similarity API