nidhaloff / deep-translator

A flexible free and unlimited python tool to translate between different languages in a simple way using multiple translators.
https://deep-translator.readthedocs.io/en/latest/?badge=latest
Apache License 2.0
1.61k stars 185 forks source link

Please support image translate #257

Open karim23657 opened 10 months ago

karim23657 commented 10 months ago

Thank you because of your amazing work. Google translate can translate image to new image and also return text in it, please add this feature. RPCIDS value for image translate is "WqWDPb"

Screenshot 2024-01-17 at 16-24-37 Google Translate

karim23657 commented 9 months ago

@nidhaloff @XelK @wenbindu @gleidsonh @yutkat , I found this code , good starter idea to develop it :

Everything looks right , but i don't know why not working.

import json
import requests
import base64
import os

def guess_type(image_path):
    # Get the file extension from the image path
    ext = os.path.splitext(image_path)[1].lower()

    # Map the extensions to the mimetypes
    mimetypes = {
        ".jpg": "image/jpeg",
        ".jpeg": "image/jpeg",
        ".png": "image/png"
    }

    # Return the mimetype if it exists, otherwise return None
    return mimetypes.get(ext, None)
def load_image_as_base64(image_path):
    with open(image_path, 'rb') as f:
        image_bytes = f.read()

    # Convert image bytes to base64 string
    base64_string = base64.b64encode(image_bytes).decode('utf-8')
    mimetype = guess_type(image_path)

    # Return image data in the specified format
    return [base64_string, mimetype]

_imageObj = load_image_as_base64('124+958.png')

_RPCIDS = "WqWDPb"
params = {
    'rpcids': _RPCIDS,
    'bl': 'boq_translate-webserver_20240115.08_p0',
    'soc-app': 1,
    'soc-platform': 1,
    'soc-device': 1,
    'rt': 'c'
    }

data = {'f.req': json.dumps([[
    [_RPCIDS,
    json.dumps([_imageObj,"auto","fr"], separators=(',', ':')),
     None,
     "generic"]
    ]], separators=(',', ':')),
 }

TRANSLATEURL = 'https://translate.google.com/_/TranslateWebserverUi/data/batchexecute'

response = requests.request("POST", TRANSLATEURL,
                            data=data,
                            params=params,
                            )

then i use bellow code to get translated image :

import base64

def convert_base64_to_image(base64_string, image_path):
    # Convert base64 string to image bytes
    image_bytes = base64.b64decode(base64_string)

    # Write the image bytes to a file
    with open(image_path, 'wb') as image_file:
        image_file.write(image_bytes)

b = response.text.split('\n')
li_filter = []
flag = False
for _b in b:
  if _b.isnumeric():
    flag = not flag
    _b = '==='
  if flag:
    li_filter.append(_b)

li_data = json.loads(li_filter[1])

resp_img_data =json.loads(li_data[0][2])
res_img_data=resp_img_data[0]
f_mimetypes = {
       "image/jpg": ".jpg" ,
        "image/jpeg" :".jpeg",
       "image/png" : ".png"
}
convert_base64_to_image(res_img_data[0], 'out'+f_mimetypes[res_img_data[1]])
StarkGang commented 7 months ago

I tried something similar, but google is returning same image as input and is not doing any translation

zeneisis commented 5 months ago

Here you are https://github.com/zeneisis/image-ocr-translate It's a part of my project

karim23657 commented 5 months ago

@zeneisis , It's good but its not exactly what i mean,I want just use deep-translator with no dependencies to translate images