You cannot select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
gpt4free/g4f/Provider/bing/upload_image.py

150 lines
4.6 KiB
Python

"""
Module to handle image uploading and processing for Bing AI integrations.
"""
from __future__ import annotations
import json
import math
from aiohttp import ClientSession, FormData
from ...typing import ImageType, Tuple
from ...image import to_image, process_image, to_base64_jpg, ImageRequest, Image
from ...requests import raise_for_status
IMAGE_CONFIG = {
"maxImagePixels": 360000,
"imageCompressionRate": 0.7,
"enableFaceBlurDebug": False,
}
async def upload_image(
session: ClientSession,
image_data: ImageType,
tone: str,
headers: dict
) -> ImageRequest:
"""
Uploads an image to Bing's AI service and returns the image response.
Args:
session (ClientSession): The active session.
image_data (bytes): The image data to be uploaded.
tone (str): The tone of the conversation.
proxy (str, optional): Proxy if any. Defaults to None.
Raises:
RuntimeError: If the image upload fails.
Returns:
ImageRequest: The response from the image upload.
"""
image = to_image(image_data)
new_width, new_height = calculate_new_dimensions(image)
image = process_image(image, new_width, new_height)
img_binary_data = to_base64_jpg(image, IMAGE_CONFIG['imageCompressionRate'])
data = build_image_upload_payload(img_binary_data, tone)
async with session.post("https://www.bing.com/images/kblob", data=data, headers=prepare_headers(headers)) as response:
await raise_for_status(response, "Failed to upload image")
return parse_image_response(await response.json())
def calculate_new_dimensions(image: Image) -> Tuple[int, int]:
"""
Calculates the new dimensions for the image based on the maximum allowed pixels.
Args:
image (Image): The PIL Image object.
Returns:
Tuple[int, int]: The new width and height for the image.
"""
width, height = image.size
max_image_pixels = IMAGE_CONFIG['maxImagePixels']
if max_image_pixels / (width * height) < 1:
scale_factor = math.sqrt(max_image_pixels / (width * height))
return int(width * scale_factor), int(height * scale_factor)
return width, height
def build_image_upload_payload(image_bin: str, tone: str) -> FormData:
"""
Builds the payload for image uploading.
Args:
image_bin (str): Base64 encoded image binary data.
tone (str): The tone of the conversation.
Returns:
Tuple[str, str]: The data and boundary for the payload.
"""
data = FormData()
knowledge_request = json.dumps(build_knowledge_request(tone), ensure_ascii=False)
data.add_field('knowledgeRequest', knowledge_request, content_type="application/json")
data.add_field('imageBase64', image_bin)
return data
def build_knowledge_request(tone: str) -> dict:
"""
Builds the knowledge request payload.
Args:
tone (str): The tone of the conversation.
Returns:
dict: The knowledge request payload.
"""
return {
"imageInfo": {},
"knowledgeRequest": {
'invokedSkills': ["ImageById"],
'subscriptionId': "Bing.Chat.Multimodal",
'invokedSkillsRequestData': {
'enableFaceBlur': True
},
'convoData': {
'convoid': "",
'convotone': tone
}
}
}
def prepare_headers(headers: dict) -> dict:
"""
Prepares the headers for the image upload request.
Args:
session (ClientSession): The active session.
boundary (str): The boundary string for the multipart/form-data.
Returns:
dict: The headers for the request.
"""
headers["Referer"] = 'https://www.bing.com/search?q=Bing+AI&showconv=1&FORM=hpcodx'
headers["Origin"] = 'https://www.bing.com'
return headers
def parse_image_response(response: dict) -> ImageRequest:
"""
Parses the response from the image upload.
Args:
response (dict): The response dictionary.
Raises:
RuntimeError: If parsing the image info fails.
Returns:
ImageRequest: The parsed image response.
"""
if not response.get('blobId'):
raise RuntimeError("Failed to parse image info.")
result = {'bcid': response.get('blobId', ""), 'blurredBcid': response.get('processedBlobId', "")}
result["imageUrl"] = f"https://www.bing.com/images/blob?bcid={result['blurredBcid'] or result['bcid']}"
result['originalImageUrl'] = (
f"https://www.bing.com/images/blob?bcid={result['blurredBcid']}"
if IMAGE_CONFIG["enableFaceBlurDebug"] else
f"https://www.bing.com/images/blob?bcid={result['bcid']}"
)
return ImageRequest(result)