Have you ever worked on a chat application where you need to display messages in real-time or on applications that need to reflect changes based on backend updates? One common method used previously is polling the backend for changes at specific intervals. However, this approach is not ideal as it slows down the client-side application and increases traffic on the backend. One solution to this problem is SignalR.

In this blog, we’ll explore how to build a real-time web application using three cutting-edge technologies: SignalR, WebAssembly, and ASP.NET Core API. SignalR simplifies the process of adding real-time web functionality, WebAssembly brings near-native performance to web apps, and ASP.NET Core API provides a robust and scalable backend. By combining these technologies, you can create web applications that are both highly responsive and performant.

What is SignalR?

SignalR is a library for ASP.NET that enables real-time web functionality. It allows server-side code to push content to connected clients instantly as it becomes available. SignalR supports multiple communication protocols, including WebSockets, and automatically falls back to other compatible protocols for older browsers. This makes it an excellent choice for applications that require high-frequency updates, like chat applications, gaming, or live data feeds.

Setting Up the ASP.NET Core API

Create a New ASP.NET Core Project

First, create a new ASP.NET Core Web API project. Open your terminal or command prompt and run the following command:

dotnet new webapi -n RealTimeApp

This command creates a new ASP.NET Core Web API project named RealTimeApp.

Install SignalR

Next, navigate to the project directory and install the SignalR package:

cd RealTimeApp
dotnet add package Microsoft.AspNetCore.SignalR

Configure SignalR in program.cs

Open the program.cs file and configure SignalR. Your program.cs file should look like this:

var builder = WebApplication.CreateBuilder(args);

builder.Services.AddEndpointsApiExplorer();
builder.Services.AddSwaggerGen();

builder.Services.AddCors(options =>
{
options.AddPolicy(“CorsPolicy”, builder =>
{
builder.WithOrigins(“http://localhost:5051”)
.AllowAnyHeader()
.AllowAnyMethod()
.AllowCredentials();
});
});

builder.Services.AddControllers();

// Adding SignalR
builder.Services.AddSignalR();

var app = builder.Build();

if (app.Environment.IsDevelopment())
{
app.UseSwagger();
app.UseSwaggerUI();
}

app.UseHttpsRedirection();
app.UseCors(“CorsPolicy”);
app.UseRouting();

// Adding the endpoint for the ChatHub
app.UseEndpoints(endpoints =>
{
endpoints.MapControllers();
endpoints.MapHub<ChatHub>(“/chatHub”);
});

app.Run();

Create the SignalR Hub

Create a new class called ChatHub.cs in the project and define the SendMessage method:

using Microsoft.AspNetCore.SignalR;

public class ChatHub : Hub
{
public async Task SendMessage(string user, string message)
{
await Clients.All.SendAsync(“ReceiveMessage”, user, message);
}
}

Integrating WebAssembly

Create a New Blazor WebAssembly Project

First, create a new Blazor WebAssembly project. Open your terminal or command prompt and run the following command:

dotnet new blazorwasm -n RealTimeApp.Client

This command creates a new Blazor WebAssembly project named RealTimeApp.Client.

Add SignalR Client Library

Navigate to the project directory and add the SignalR client library:

cd RealTimeApp.Client
dotnet add package Microsoft.AspNetCore.SignalR.Client

Create the Blazor Component

Create a new Razor component for the chat interface. You can name it Chat.razor.

Add the following content to Chat.razor:

“/chat”
@using Microsoft.AspNetCore.SignalR.Client

@inject NavigationManager NavigationManager

<h3>Chat</h3>

<input @bind=”userInput” placeholder=”Your name” class=”form-control” />
<br>
<input @bind=”messageInput” placeholder=”Your message” @onkeypress=”@(e => { if (e.Key == “Enter”) Send().GetAwaiter().GetResult(); })” class=”form-control” />
<br>

<div style=”text-align: right”>
<button @onclick=”Send” class=”btn btn-primary”>Send</button>
</div>

<ul>
@foreach (var message in messages)
{
<li>@message</li>
}
</ul>

@code {
private HubConnection hubConnection;
private string userInput;
private string messageInput;
private List<string> messages = new List<string>();

protected override async Task OnInitializedAsync()
{
hubConnection = new HubConnectionBuilder()
.WithUrl(“http://localhost:5258/chathub”)
.Build();

hubConnection.On<string, string>(“ReceiveMessage”, (user, message) =>
{
var encodedMsg = $”{user}: {message}”;
messages.Add(encodedMsg);
InvokeAsync(StateHasChanged);
});

await hubConnection.StartAsync();
}

private async Task Send()
{
await hubConnection.SendAsync(“SendMessage”, userInput, messageInput);
messageInput = string.Empty;
}
}

Add Navigation to Chat.razor

Ensure your Chat.razor component can be accessed via navigation. Update your NavMenu.razor file to include a link to the chat component:

<ul class=”nav flex-column”>
<li class=”nav-item px-3″>
<NavLink class=”nav-link” href=””>
<span class=”oi oi-home” aria-hidden=”true”></span> Home
</NavLink>
</li>
<li class=”nav-item px-3″>
<NavLink class=”nav-link” href=”chat”>
<span class=”oi oi-chat” aria-hidden=”true”></span> Chat
</NavLink>
</li>
</ul>

Run the Application

Run the ASP.NET Core API:

dotnet run –project RealTimeApp

Run the Blazor WebAssembly app:

dotnet run –project RealTimeApp.Client

Navigate to /chat in your Blazor WebAssembly app to see the chat interface. You should be able to send messages, which will be broadcasted to all connected clients in real-time using SignalR.

Conclusion

In this blog, we’ve explored how to build a real-time web application using SignalR, WebAssembly, and ASP.NET Core API. By leveraging these technologies, we created a responsive chat application capable of instant updates without the need for constant polling. SignalR facilitated real-time communication, WebAssembly provided near-native performance, and ASP.NET Core API served as a robust backend.

We started by setting up an ASP.NET Core API and integrating SignalR to handle real-time messages. Then, we created a Blazor WebAssembly client that interacts with the SignalR hub to send and receive messages. This combination of technologies not only enhances performance but also simplifies the development of real-time web applications.

I hope this tutorial has provided you with a solid foundation for building your own real-time applications. To see the complete source code and explore further, visit the GitHub repository: GitHub Repository: RealTimeApp

Feel free to clone the repository, experiment with the code, and extend the application to suit your needs.

To explore more on this topic, you can check out the following resources:

Tutorial: Get started with ASP.NET Core SignalR
Introduction to SignalR

Happy coding!

​Microsoft Tech Community – Latest Blogs –Read More 

جـلب الـحبيب في السحر العجيب” 57838.5770 66 009-

​جـلب الـحبيب في السحر العجيب” 57838.5770 66 009-  Read More

​

جـلب الـحبيب بسرعة البرق 5940271.42 00.966

​جـلب الـحبيب بسرعة البرق 5940271.42 00.966  Read More

​

Hi, Microsoft Tech Community and @Ricky Simpson from Microsoft,

I cannot download Update 3 for Microsoft Advanced Threat Analytics 1.9.

Whenever I tried to access the download update from this article, it seemed the ID number 56725 was missing, and an error code of 404 was returned.

Tried URL:

https://www.microsoft.com/download/details.aspx?id=56725

Hope you can fix this problem as soon as possible, because Microsoft ATA still plays an important role in most of the enterprise network, including my company’s network.

Best regards for all people in the community

​Hi, Microsoft Tech Community and @Ricky Simpson from Microsoft, I cannot download Update 3 for Microsoft Advanced Threat Analytics 1.9. Whenever I tried to access the download update from this article, it seemed the ID number 56725 was missing, and an error code of 404 was returned.   Tried URL:https://www.microsoft.com/download/details.aspx?id=56725 Hope you can fix this problem as soon as possible, because Microsoft ATA still plays an important role in most of the enterprise network, including my company’s network. Best regards for all people in the community  Read More

​

Hi everyone,

I’m planning to sell my Windows 11 device and want to securely wipe the hard drive to ensure none of my personal data is recoverable. I’ve heard that just deleting files or even resetting the PC may not be enough to completely erase everything. Could anyone recommend a reliable method or tool to wipe a hard drive completely on Windows 11 before I sell it?

I’d prefer a straightforward process that guarantees data can’t be recovered. Any step-by-step instructions or software suggestions would be greatly appreciated!

Thanks for your help!

​Hi everyone, I’m planning to sell my Windows 11 device and want to securely wipe the hard drive to ensure none of my personal data is recoverable. I’ve heard that just deleting files or even resetting the PC may not be enough to completely erase everything. Could anyone recommend a reliable method or tool to wipe a hard drive completely on Windows 11 before I sell it? I’d prefer a straightforward process that guarantees data can’t be recovered. Any step-by-step instructions or software suggestions would be greatly appreciated! Thanks for your help!  Read More

​

Hi, I’m trying to create a tracking spreadsheet indicative by color based on value. Using IF function, will work if I do it individually by cell but I’m wondering if there’s a better way. I’m working with < less than and > greater than.   Example: Reference Number in A3, if number in comparison cells (all in the same column) are greater, I need the cell green, if less I need the cell red. I tried conditional formatting but not having much luck.

​Hi, I’m trying to create a tracking spreadsheet indicative by color based on value. Using IF function, will work if I do it individually by cell but I’m wondering if there’s a better way. I’m working with < less than and > greater than.   Example: Reference Number in A3, if number in comparison cells (all in the same column) are greater, I need the cell green, if less I need the cell red. I tried conditional formatting but not having much luck.   Read More

​

Dear All,

I am planning to learn Azure Data engineering, I wanted to get hands-on in cloud environment. 

My free trial got expired, is there anyway we can get hands-on without cost involved.

I can go for pay as you go option, but I am not sure how the costing will happen. I am afraid as they may charge unexpectedly if I subscribed for some service unknowingly. I encountered this in my trial period as they charged 78$ (from free credit) overnight for some storage services I created and left it.

Any help is appreciated, thanks in advance.

​Dear All, I am planning to learn Azure Data engineering, I wanted to get hands-on in cloud environment.  My free trial got expired, is there anyway we can get hands-on without cost involved. I can go for pay as you go option, but I am not sure how the costing will happen. I am afraid as they may charge unexpectedly if I subscribed for some service unknowingly. I encountered this in my trial period as they charged 78$ (from free credit) overnight for some storage services I created and left it. Any help is appreciated, thanks in advance.  Read More

​

Viva Engage community insights doesn’t include the data from the first half of the year when exporting data for 365 days. I am trying to see our community insights data from January 2024 up to the current date, but the csv file only shows July to the current date, leaving the first half of the year.

​Viva Engage community insights doesn’t include the data from the first half of the year when exporting data for 365 days. I am trying to see our community insights data from January 2024 up to the current date, but the csv file only shows July to the current date, leaving the first half of the year.  Read More

​

You’re welcome to follow my GitHub repo and give it a star：https://github.com/xinyuwei-david/david-share.git，lots of useful code is here！

ONNX (Open Neural Network Exchange) is an open format used to represent machine learning and deep learning models. It was introduced by Microsoft and Facebook in 2017, aiming to facilitate model interoperability between different deep learning frameworks. With ONNX, you can seamlessly convert models between different deep learning frameworks such as PyTorch and TensorFlow.

Currently, ONNX fine-tuning can be done using Olive, but it does not yet support LoRA. If you want to perform LoRA fine-tuning with PyTorch and use ORT for inference, how can this be achieved?

First, fine-tune the model using LoRA. Do not use QLoRA, as it may result in significant precision loss during subsequent merging.
Merge the Adapter with the PyTorch base model.
Convert the merged safetensors to ONNX.
Generate the genai_config.json file using ONNX Runtime GenAI Model Builder.
Perform inference using onnxruntime-genai.

Merge adapter：

Consolidated results:

Export to ONNX

Export result：

Generate genai_config.json.

when doing the conversion, you need to use FP32.

Files needed during ONNX inference：

Detailed Code

LoRA SFT

model_name = “microsoft/Phi-3.5-Mini-instruct”

tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True, add_eos_token=True, use_fast=True)
tokenizer.pad_token = tokenizer.unk_token
tokenizer.pad_token_id = tokenizer.convert_tokens_to_ids(tokenizer.pad_token)
tokenizer.padding_side = ‘left’

ds = load_dataset(“timdettmers/openassistant-guanaco”)

model = AutoModelForCausalLM.from_pretrained(
model_name, torch_dtype=compute_dtype, trust_remote_code=True, device_map={“”: 0}, attn_implementation=attn_implementation
)

model.gradient_checkpointing_enable(gradient_checkpointing_kwargs={‘use_reentrant’:True})

peft_config = LoraConfig(
lora_alpha=16,
lora_dropout=0.05,
r=16,
bias=”none”,
task_type=”CAUSAL_LM”,
target_modules= [‘k_proj’, ‘q_proj’, ‘v_proj’, ‘o_proj’, “gate_proj”, “down_proj”, “up_proj”]
)

training_arguments = SFTConfig(
output_dir=”./Phi-3.5/Phi-3.5-Mini_LoRA”,
eval_strategy=”steps”,
do_eval=True,
optim=”adamw_torch”,
per_device_train_batch_size=4,
gradient_accumulation_steps=8,
per_device_eval_batch_size=4,
log_level=”debug”,
save_strategy=”epoch”,
logging_steps=25,
learning_rate=1e-4,
fp16 = not torch.cuda.is_bf16_supported(),
bf16 = torch.cuda.is_bf16_supported(),
eval_steps=25,
num_train_epochs=1,
warmup_ratio=0.1,
lr_scheduler_type=”linear”,
dataset_text_field=”text”,
max_seq_length=512
)

trainer = SFTTrainer(
model=model,
train_dataset=ds[‘train’],
eval_dataset=ds[‘test’],
peft_config=peft_config,
tokenizer=tokenizer,
args=training_arguments,
)

trainer.train()

Merge the adapter

model_name = “microsoft/Phi-3.5-Mini-instruct”
adapter_path = “/root/Phi-3.5-Mini_LoRA/checkpoint-411/”

# 加载 tokenizer
tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True, use_fast=True)

# 加载模型
model = AutoModelForCausalLM.from_pretrained(model_name, trust_remote_code=True)

# 加载适配器
model = PeftModel.from_pretrained(model, adapter_path)

# 设置模型为评估模式
model.eval()

# 定义推理函数
def generate_text(prompt, max_length=500):
inputs = tokenizer(prompt, return_tensors=”pt”)
attention_mask = inputs[‘attention_mask’]
with torch.no_grad():
outputs = model.generate(
inputs.input_ids,
attention_mask=attention_mask,
max_length=max_length,
num_return_sequences=1,
do_sample=True,
top_k=50,
top_p=0.95
)
return tokenizer.decode(outputs[0], skip_special_tokens=True)

# 示例推理
prompt = (“1+1=?”)
generated_text = generate_text(prompt)
print(generated_text)

ONNX Export

model_checkpoint = “/root/Phi-3.5-Mini-LoRA-Merge”
save_directory = “/root/onnx1/”

# Load a model from transformers and export it to ONNX
ort_model = ORTModelForCausalLM.from_pretrained(model_checkpoint, export=True)
tokenizer = AutoTokenizer.from_pretrained(model_checkpoint)

# Save the onnx model and tokenizer
ort_model.save_pretrained(save_directory)
tokenizer.save_pretrained(save_directory)

Generate genai_config.json

(phi3.5) root@h100vm:~/onnxruntime-genai/src/python/py/models# python3 builder.py -m microsoft/Phi-3.5-mini-instruct -o /root/onnx4 -p fp16 -e cuda -c /root/onnx1 –extra_options config_only=true

Valid precision + execution provider combinations are: FP32 CPU, FP32 CUDA, FP16 CUDA, FP16 DML, INT4 CPU, INT4 CUDA, INT4 DML
Extra options: {‘config_only’: ‘true’}
config.json: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 3.45k/3.45k [00:00<00:00, 33.5MB/s]
configuration_phi3.py: 100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 11.2k/11.2k [00:00<00:00, 79.0MB/s]
A new version of the following files was downloaded from https://huggingface.co/microsoft/Phi-3.5-mini-instruct:
– configuration_phi3.py
. Make sure to double-check they do not contain any added malicious code. To avoid downloading new versions of the code file, you can pin a revision.
GroupQueryAttention (GQA) is used in this model.
generation_config.json: 100%|████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 195/195 [00:00<00:00, 2.19MB/s]
Saving GenAI config in /root/onnx4
tokenizer_config.json: 100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 3.98k/3.98k [00:00<00:00, 42.8MB/s]
tokenizer.model: 100%|██████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 500k/500k [00:00<00:00, 105MB/s]
tokenizer.json: 100%|████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 1.84M/1.84M [00:00<00:00, 2.13MB/s]
added_tokens.json: 100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 306/306 [00:00<00:00, 3.40MB/s]
special_tokens_map.json: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 665/665 [00:00<00:00, 7.23MB/s]
Saving processing files in /root/onnx4 for GenAI

Copy generated files to ONNX model file:

(phi3.5) root@h100vm:~/onnx4# ls
added_tokens.json genai_config.json special_tokens_map.json tokenizer.json tokenizer.model tokenizer_config.json
(phi3.5) root@h100vm:~/onnx4# cp ./* /root/onnx1

Inference test with ONNX:

model = og.Model(‘/root/onnx1/1’)
tokenizer = og.Tokenizer(model)
tokenizer_stream = tokenizer.create_stream()

# Set the max length to something sensible by default,
# since otherwise it will be set to the entire context length
search_options = {}
search_options[‘max_length’] = 2048

chat_template = ‘<|user|>n{input} <|end|>n<|assistant|>’

text = input(“Input: “)
if not text:
print(“Error, input cannot be empty”)
exit

prompt = f'{chat_template.format(input=text)}’

input_tokens = tokenizer.encode(prompt)

params = og.GeneratorParams(model)
params.set_search_options(**search_options)
params.input_ids = input_tokens
generator = og.Generator(model, params)

print(“Output: “, end=”, flush=True)

try:
while not generator.is_done():
generator.compute_logits()
generator.generate_next_token()

new_token = generator.get_next_tokens()[0]
print(tokenizer_stream.decode(new_token), end=”, flush=True)
except KeyboardInterrupt:
print(” –control+c pressed, aborting generation–“)

print()
del generator

​Microsoft Tech Community – Latest Blogs –Read More 

On December 5th, 2024, we will host a public webinar on how to effectively integrate APIs with Microsoft Sentinel and the Unified Security Platform. This session will cover when to use APIs, how to set them up, and potential challenges. We will present live demos to guide you through the process. To ensure this webinar is as engaging and relevant as possible for you, we’d love your input to help us create its agenda!

Help us plan this webinar
Do you have any use cases you think we should feature? Or have you encountered any blockers that you’d like us to address? We’re eager to find out what content matches your needs the most! Please answer this survey to help us with your input. It will remain open until October 31st, 2024.

Take the survey here: https://forms.office.com/r/hrWtm34WFu

Join the webinar on December 5th!
In addition to helping us plan it, we hope to count on your participation. Register at Register for this webinar at https://aka.ms/MasteringAPISentinelUSOPWebinar. 

Thank you for your contributions!

Naomi Chistis and Jeremey Tan – Microsoft SIEM & XDR Team

​Hello on behalf of the Microsoft SIEM & XDR Engineering organization! 
On December 5th, 2024, we will host a public webinar on how to effectively integrate APIs with Microsoft Sentinel and the Unified Security Platform. This session will cover when to use APIs, how to set them up, and potential challenges. We will present live demos to guide you through the process. To ensure this webinar is as engaging and relevant as possible for you, we’d love your input to help us create its agenda!Help us plan this webinarDo you have any use cases you think we should feature? Or have you encountered any blockers that you’d like us to address? We’re eager to find out what content matches your needs the most! Please answer this survey to help us with your input. It will remain open until October 31st, 2024.Take the survey here: https://forms.office.com/r/hrWtm34WFuJoin the webinar on December 5th!In addition to helping us plan it, we hope to count on your participation. Register at Register for this webinar at https://aka.ms/MasteringAPISentinelUSOPWebinar. Thank you for your contributions!Naomi Chistis and Jeremey Tan – Microsoft SIEM & XDR Team  Read More

​