Add upload svg image support

Fix upload image in Bing Provider
This commit is contained in:
Heiner Lohaus 2024-01-23 01:47:55 +01:00
parent 38dbe4b8e5
commit 07c944ad0a
6 changed files with 51 additions and 26 deletions

View File

@ -82,13 +82,16 @@ def build_image_upload_payload(image_bin: str, tone: str) -> Tuple[str, str]:
Tuple[str, str]: The data and boundary for the payload. Tuple[str, str]: The data and boundary for the payload.
""" """
boundary = "----WebKitFormBoundary" + ''.join(random.choices(string.ascii_letters + string.digits, k=16)) boundary = "----WebKitFormBoundary" + ''.join(random.choices(string.ascii_letters + string.digits, k=16))
data = f"--{boundary}\r\n" \ data = f"""--{boundary}
f"Content-Disposition: form-data; name=\"knowledgeRequest\"\r\n\r\n" \ Content-Disposition: form-data; name="knowledgeRequest"
f"{json.dumps(build_knowledge_request(tone), ensure_ascii=False)}\r\n" \
f"--{boundary}\r\n" \ {json.dumps(build_knowledge_request(tone), ensure_ascii=False)}
f"Content-Disposition: form-data; name=\"imageBase64\"\r\n\r\n" \ --{boundary}
f"{image_bin}\r\n" \ Content-Disposition: form-data; name="imageBase64"
f"--{boundary}--\r\n"
{image_bin}
--{boundary}--
"""
return data, boundary return data, boundary
def build_knowledge_request(tone: str) -> dict: def build_knowledge_request(tone: str) -> dict:
@ -102,14 +105,17 @@ def build_knowledge_request(tone: str) -> dict:
dict: The knowledge request payload. dict: The knowledge request payload.
""" """
return { return {
'invokedSkills': ["ImageById"], "imageInfo": {},
'subscriptionId': "Bing.Chat.Multimodal", "knowledgeRequest": {
'invokedSkillsRequestData': { 'invokedSkills': ["ImageById"],
'enableFaceBlur': True 'subscriptionId': "Bing.Chat.Multimodal",
}, 'invokedSkillsRequestData': {
'convoData': { 'enableFaceBlur': True
'convoid': "", },
'convotone': tone 'convoData': {
'convoid': "",
'convotone': tone
}
} }
} }

View File

@ -115,11 +115,11 @@
<textarea id="message-input" placeholder="Ask a question" cols="30" rows="10" <textarea id="message-input" placeholder="Ask a question" cols="30" rows="10"
style="white-space: pre-wrap;resize: none;"></textarea> style="white-space: pre-wrap;resize: none;"></textarea>
<label for="image" title="Works only with Bing and OpenaiChat"> <label for="image" title="Works only with Bing and OpenaiChat">
<input type="file" id="image" name="image" accept="image/png, image/gif, image/jpeg" required/> <input type="file" id="image" name="image" accept="image/png, image/gif, image/jpeg, image/svg+xml" required/>
<i class="fa-regular fa-image"></i> <i class="fa-regular fa-image"></i>
</label> </label>
<label for="file"> <label for="file">
<input type="file" id="file" name="file" accept="text/plain, text/html, text/xml, application/json, text/javascript, .sh, .py, .php, .css, .yaml, .sql, .svg, .log, .csv, .twig, .md" required/> <input type="file" id="file" name="file" accept="text/plain, text/html, text/xml, application/json, text/javascript, .sh, .py, .php, .css, .yaml, .sql, .log, .csv, .twig, .md" required/>
<i class="fa-solid fa-paperclip"></i> <i class="fa-solid fa-paperclip"></i>
</label> </label>
<div id="send-button"> <div id="send-button">

View File

@ -660,7 +660,13 @@ observer.observe(message_input, { attributes: true });
} }
document.getElementById("version_text").innerHTML = text document.getElementById("version_text").innerHTML = text
})() })()
imageInput.addEventListener('click', async (event) => {
imageInput.value = '';
});
fileInput.addEventListener('click', async (event) => {
fileInput.value = '';
delete fileInput.dataset.text;
});
fileInput.addEventListener('change', async (event) => { fileInput.addEventListener('change', async (event) => {
if (fileInput.files.length) { if (fileInput.files.length) {
type = fileInput.files[0].type; type = fileInput.files[0].type;

View File

@ -137,7 +137,7 @@ class Backend_Api:
if 'image' in request.files: if 'image' in request.files:
file = request.files['image'] file = request.files['image']
if file.filename != '' and is_allowed_extension(file.filename): if file.filename != '' and is_allowed_extension(file.filename):
kwargs['image'] = to_image(file.stream) kwargs['image'] = to_image(file.stream, file.filename.endswith('.svg'))
if 'json' in request.form: if 'json' in request.form:
json_data = json.loads(request.form['json']) json_data = json.loads(request.form['json'])
else: else:

View File

@ -4,9 +4,9 @@ import base64
from .typing import ImageType, Union from .typing import ImageType, Union
from PIL import Image from PIL import Image
ALLOWED_EXTENSIONS = {'png', 'jpg', 'jpeg', 'gif', 'webp'} ALLOWED_EXTENSIONS = {'png', 'jpg', 'jpeg', 'gif', 'webp', 'svg'}
def to_image(image: ImageType) -> Image.Image: def to_image(image: ImageType, is_svg: bool = False) -> Image.Image:
""" """
Converts the input image to a PIL Image object. Converts the input image to a PIL Image object.
@ -16,6 +16,16 @@ def to_image(image: ImageType) -> Image.Image:
Returns: Returns:
Image.Image: The converted PIL Image object. Image.Image: The converted PIL Image object.
""" """
if is_svg:
try:
import cairosvg
except ImportError:
raise RuntimeError('Install "cairosvg" package for open svg images')
if not isinstance(image, bytes):
image = image.read()
buffer = BytesIO()
cairosvg.svg2png(image, write_to=buffer)
image = Image.open(buffer)
if isinstance(image, str): if isinstance(image, str):
is_data_uri_an_image(image) is_data_uri_an_image(image)
image = extract_data_uri(image) image = extract_data_uri(image)
@ -153,6 +163,8 @@ def to_base64(image: Image.Image, compression_rate: float) -> str:
str: The base64-encoded image. str: The base64-encoded image.
""" """
output_buffer = BytesIO() output_buffer = BytesIO()
if image.mode != "RGB":
image = image.convert('RGB')
image.save(output_buffer, format="JPEG", quality=int(compression_rate * 100)) image.save(output_buffer, format="JPEG", quality=int(compression_rate * 100))
return base64.b64encode(output_buffer.getvalue()).decode() return base64.b64encode(output_buffer.getvalue()).decode()

View File

@ -5,6 +5,7 @@ from .Provider import (
Chatgpt4Online, Chatgpt4Online,
ChatgptDemoAi, ChatgptDemoAi,
GeminiProChat, GeminiProChat,
PerplexityAi,
ChatgptNext, ChatgptNext,
HuggingChat, HuggingChat,
ChatgptDemo, ChatgptDemo,
@ -78,7 +79,7 @@ gpt_35_long = Model(
gpt_35_turbo = Model( gpt_35_turbo = Model(
name = 'gpt-3.5-turbo', name = 'gpt-3.5-turbo',
base_provider = 'openai', base_provider = 'openai',
best_provider=RetryProvider([ best_provider = RetryProvider([
GptGo, You, GptGo, You,
GptForLove, ChatBase, GptForLove, ChatBase,
Chatgpt4Online, Chatgpt4Online,
@ -114,20 +115,20 @@ llama2_13b = Model(
llama2_70b = Model( llama2_70b = Model(
name = "meta-llama/Llama-2-70b-chat-hf", name = "meta-llama/Llama-2-70b-chat-hf",
base_provider = "huggingface", base_provider = "huggingface",
best_provider = RetryProvider([Llama2, DeepInfra, HuggingChat]) best_provider = RetryProvider([Llama2, DeepInfra, HuggingChat, PerplexityAi])
) )
# Mistal # Mistal
mixtral_8x7b = Model( mixtral_8x7b = Model(
name = "mistralai/Mixtral-8x7B-Instruct-v0.1", name = "mistralai/Mixtral-8x7B-Instruct-v0.1",
base_provider = "huggingface", base_provider = "huggingface",
best_provider = RetryProvider([DeepInfra, HuggingChat]) best_provider = RetryProvider([DeepInfra, HuggingChat, PerplexityAi])
) )
mistral_7b = Model( mistral_7b = Model(
name = "mistralai/Mistral-7B-Instruct-v0.1", name = "mistralai/Mistral-7B-Instruct-v0.1",
base_provider = "huggingface", base_provider = "huggingface",
best_provider = RetryProvider([DeepInfra, HuggingChat]) best_provider = RetryProvider([DeepInfra, HuggingChat, PerplexityAi])
) )
# Dolphin # Dolphin
@ -311,7 +312,7 @@ llama70b_v2_chat = Model(
pi = Model( pi = Model(
name = 'pi', name = 'pi',
base_provider = 'inflection', base_provider = 'inflection',
best_provider=Pi best_provider = Pi
) )
class ModelUtils: class ModelUtils: