support msgpack

[Ehekatl]

@Transfusion

Java Example for get bow_voc:

import java.io.BufferedReader;
import java.io.DataOutputStream;
import java.io.InputStreamReader;
import java.net.HttpURLConnection;
import java.net.URL;
import java.util.Arrays;
import java.util.Base64;

import javax.net.ssl.HttpsURLConnection;

import com.google.gson.Gson;
import org.msgpack.MessagePack;

public class Main {

    MessagePack msgpack = new MessagePack();
    Gson gson = new Gson();

    public static void main(String[] args) throws Exception {

        Main http = new Main();

        http.sendGet();
    }

    private void sendGet() throws Exception {

        String url = "http://127.0.0.1:8000/vocabulary/";

        URL obj = new URL(url);
        HttpURLConnection con = (HttpURLConnection) obj.openConnection();

        // optional default is GET
        con.setRequestMethod("GET");

        int responseCode = con.getResponseCode();
        System.out.println("\nSending 'GET' request to URL : " + url);
        System.out.println("Response Code : " + responseCode);

        BufferedReader in = new BufferedReader(
                new InputStreamReader(con.getInputStream()));
        String inputLine;
        StringBuffer response = new StringBuffer();

        while ((inputLine = in.readLine()) != null) {
            response.append(inputLine);
        }
        in.close();

        //print result
        byte[] bytes = Base64.getDecoder().decode(response.toString());
        Float[][] fff = msgpack.read(bytes, Float[][].class);
        System.out.println(Arrays.deepToString(fff));
    }
}

Python Example for search by histogram:

import time
import requests
import base64
import cv2
import numpy as np
import msgpack
from algolib.populator import Populator

url = 'http://host/vocabulary/'
s = base64.b64decode(requests.get(url).text)
voc = np.array(msgpack.unpackb(s), dtype=np.float32)

pop = Populator()
img = cv2.imread('test/6.jpg')
bow_hist = pop.bow_hist(img, voc)
color_hist = pop.color_hist(img)

bow_hist = base64.b64encode(msgpack.packb(bow_hist.tolist()))
color_hist = base64.b64encode(msgpack.packb(color_hist.tolist()))

url = 'http://host/images/'
args = {'limit': 20, 'method': 'chisqr_alt'}
data = {'bow_hist':bow_hist, 'color_hist':color_hist}
start_time = time.clock()
r = requests.post(url, data=data, params=args)
print(time.clock() - start_time, "seconds")
print(r.status_code)
print(r.text)

Python Example for search by file (sketched image with alpha channel):

import requests
import time

url = 'http://host/images/'
files = {'image': open('test/Drawing.png', 'rb')}
args = {'limit': 20, 'method': 'chisqr_alt', 'sketch': 'putanythinghere'}
start_time = time.clock()
r = requests.post(url, files=files, params=args)
print(time.clock() - start_time, "seconds")
print(r.status_code)
print(r.text)

[Ehekatl]

@Transfusion
API Post result example:

[{"img_path":"http://7xnyxq.com1.z0.glb.clouddn.com/07389.JPEG"},{"img_path":"http://7xnyxq.com1.z0.glb.clouddn.com/06273.JPEG"},{"img_path":"http://7xnyxq.com1.z0.glb.clouddn.com/07201.JPEG"},{"img_path":"http://7xnyxq.com1.z0.

glb.clouddn.com/01219.JPEG"},{"img_path":"http://7xnyxq.com1.z0.glb.clouddn.com/00287.JPEG"},{"img_path":"http://7xnyxq.com1.z0.glb.clouddn.com/05264.JPEG"},{"img_path":"http://7xnyxq.com1.z0.glb.clouddn.com/06288.JPEG"},{"img_p
ath":"http://7xnyxq.com1.z0.glb.clouddn.com/08774.JPEG"},{"img_path":"http://7xnyxq.com1.z0.glb.clouddn.com/00863.JPEG"},{"img_path":"http://7xnyxq.com1.z0.glb.clouddn.com/00059.JPEG"},{"img_path":"http://7xnyxq.com1.z0.glb.clou
ddn.com/08711.JPEG"},{"img_path":"http://7xnyxq.com1.z0.glb.clouddn.com/00424.JPEG"},{"img_path":"http://7xnyxq.com1.z0.glb.clouddn.com/08316.JPEG"},{"img_path":"http://7xnyxq.com1.z0.glb.clouddn.com/07062.JPEG"},{"img_path":"ht
tp://7xnyxq.com1.z0.glb.clouddn.com/03833.JPEG"},{"img_path":"http://7xnyxq.com1.z0.glb.clouddn.com/03966.JPEG"},{"img_path":"http://7xnyxq.com1.z0.glb.clouddn.com/07575.JPEG"},{"img_path":"http://7xnyxq.com1.z0.glb.clouddn.com/
03178.JPEG"},{"img_path":"http://7xnyxq.com1.z0.glb.clouddn.com/00745.JPEG"},{"img_path":"http://7xnyxq.com1.z0.glb.clouddn.com/07341.JPEG"}]

I disabled private url, now the image can have direct access, for example:
http://7xnyxq.com1.z0.glb.clouddn.com/03979.JPEG

If you need small size image:
http://7xnyxq.com1.z0.glb.clouddn.com/03979.JPEG-middle

http://7xnyxq.com1.z0.glb.clouddn.com/03979.JPEG-small

http://7xnyxq.com1.z0.glb.clouddn.com/03979.JPEG-tiny

[Ehekatl]

Change:
Now the maximum number of results will be 100, you can add optional args to disable BoW compare

import requests
import time

url = 'http://host/images/'
files = {'image': open('test/Drawing.png', 'rb')}
args = {'limit': 20, 'method': 'chisqr_alt', 'sketch': 'putanythinghere', 'disableBoW': 'putanythinghere'}
start_time = time.clock()
r = requests.post(url, files=files, params=args)
print(time.clock() - start_time, "seconds")
print(r.status_code)
print(r.text)

As a result, the image search performance have great improvement, on local it deduce from 600+ ms to 100ms, on server it drop from 1400+ms to 400ms

Example result:
without BoW,

with BoW,