diff --git a/Smart_Summarize/AIAssitViewModel.cs b/Smart_Summarize/AIAssitViewModel.cs
index e137a13..f27ac79 100644
--- a/Smart_Summarize/AIAssitViewModel.cs
+++ b/Smart_Summarize/AIAssitViewModel.cs
@@ -94,26 +94,15 @@ public AIAssitViewModel(PdfViewerControl viewer)
///
public async Task GenerateMessages()
{
- Chats.Add(new TextMessage
- {
- Author = currentUser,
- DateTime = DateTime.Now,
- Text = "Summarizing the PDF document..."
- });
-
// Execute the following asynchronously
await ExtractDetailsFromPDF();
- string summaryText = await SummarizePDF();
-
- // Update chats on the UI thread
Chats.Add(new TextMessage
{
- Author = new Author { Name = "AIAssistant" },
+ Author = currentUser,
DateTime = DateTime.Now,
- Text = summaryText
+ Text = "Summarizing the PDF document..."
});
- await AddSuggestions(summaryText);
}
///
/// Generate suggestion from the answers
@@ -133,48 +122,20 @@ private async Task AddSuggestions(String text)
///
private async Task ExtractDetailsFromPDF()
{
- StringBuilder extractedText = new StringBuilder();
+ List extractedText = new List();
Syncfusion.Pdf.TextLines textLines = new Syncfusion.Pdf.TextLines();
//Extract the text from the PDF document
for (int pageIndex = 0; pageIndex < pdfViewer.PageCount; pageIndex++)
{
string text = $"... Page {pageIndex + 1} ...\n";
text += pdfViewer.ExtractText(pageIndex, out textLines);
- extractedText.AppendLine(text);
+ extractedText.Add(text);
}
- return ProcessExtractedText(extractedText.ToString());
- }
- ///
- /// Processes the extracted full text from a document by splitting it into pages.
- ///
- /// The complete extracted text from the document.
- /// A formatted string containing the processed text.
- private string ProcessExtractedText(string fullText)
- {
- string[] pages = fullText.Split(new string[] { "\f", "\n\nPage " }, StringSplitOptions.RemoveEmptyEntries);
- for (int i = 0; i < pages.Length; i++)
- {
- processedText.AppendLine($"... Page {i + 1} ...");
- string[] lines = pages[i].Split(new[] { '\n' }, StringSplitOptions.RemoveEmptyEntries);
- int maxLines = Math.Min(1000, lines.Length);
+ microsoftAIExtension.CreateEmbeddedPage(extractedText.ToArray());
- for (int j = 0; j < maxLines; j++)
- {
- processedText.AppendLine(lines[j]);
- }
- processedText.AppendLine();
- }
- return processedText.ToString();
- }
- ///
- /// Summarizes the extracted text from the PDF using Extension AI.
- ///
- private async Task SummarizePDF()
- {
- //Summarize the text using the Semantic Kernel AI
- string summary = await microsoftAIExtension.GetAnswerFromGPT(processedText.ToString());
- return summary;
+ return extractedText.ToString();
}
+
///
/// Handles the event when the chat collection changes.
///
@@ -187,23 +148,18 @@ private async void Chats_CollectionChanged(object sender, System.Collections.Spe
var item = e.NewItems[0] as ITextMessage;
if (item != null)
{
- if (item.Text != "Summarizing the PDF document...")
+ if (item.Author.Name == currentUser.Name)
{
- if (item.Author.Name == currentUser.Name)
+ string answer = await microsoftAIExtension.AnswerQuestion(item.Text);
+ Chats.Add(new TextMessage
{
- string answer = await microsoftAIExtension.GetAnswerFromGPT("You are a helpful assistant. Your task is to analyze the provided question and answer the question based on the pdf", item.Text);
- Chats.Add(new TextMessage
- {
- Author = new Author { Name = "AIAssistant" },
- DateTime = DateTime.Now,
- Text = answer
- });
- Suggestion.Clear();
- await AddSuggestions(answer);
- }
-
+ Author = new Author { Name = "AIAssistant" },
+ DateTime = DateTime.Now,
+ Text = answer
+ });
+ Suggestion.Clear();
+ await AddSuggestions(answer);
}
-
}
}
}
diff --git a/Smart_Summarize/MicrosoftAIExtension.cs b/Smart_Summarize/MicrosoftAIExtension.cs
index e5c58c2..18f5beb 100644
--- a/Smart_Summarize/MicrosoftAIExtension.cs
+++ b/Smart_Summarize/MicrosoftAIExtension.cs
@@ -1,5 +1,6 @@
-using Microsoft.Extensions.AI;
-using Azure.AI.OpenAI;
+using Azure.AI.OpenAI;
+using Microsoft.Extensions.AI;
+using SmartComponents.LocalEmbeddings;
using System.Text;
namespace Smart_Summarize
{
@@ -8,10 +9,11 @@ internal class MicrosoftAIExtension
const string endpoint = "YOUR-AI-ENDPOINT";
const string deploymentName = "YOUR-DEPLOYMENT-NAME";
- internal string key = string.Empty;
private IChatClient clientAI;
+ public Dictionary? PageEmbeddings { get; set; }
+
///
/// Initializes a new instance of the class.
///
@@ -64,5 +66,33 @@ public async Task GetAnswerFromGPT(string systemPrompt, string userText)
}
return string.Empty;
}
+
+ ///
+ /// Create the embedded page from the extracted chunks in the PDF
+ ///
+ /// Extracted text from pdfViewer
+ ///
+ public async Task CreateEmbeddedPage(string[] chunks)
+ {
+ var embedder = new LocalEmbedder();
+ PageEmbeddings = chunks.Select(x => KeyValuePair.Create(x, embedder.Embed(x))).ToDictionary(k => k.Key, v => v.Value);
+ }
+
+
+ public async Task AnswerQuestion(string question)
+ {
+ var embedder = new LocalEmbedder();
+ var questionEmbedding = embedder.Embed(question);
+ var results = LocalEmbedder.FindClosestWithScore(questionEmbedding, PageEmbeddings.Select(x => (x.Key, x.Value)), 5, 0.5f);
+ StringBuilder builder = new StringBuilder();
+ foreach (var result in results)
+ {
+ builder.AppendLine(result.Item);
+ }
+ string message = builder.ToString();
+ var answer = await GetAnswerFromGPT("You are a helpful assistant. Use the provided PDF document pages and pick a precise page to answer the user question. Provide the answer in plain text without any special formatting or Markdown syntax. Pages: " + message, question);
+
+ return answer;
+ }
}
-}
+}
\ No newline at end of file
diff --git a/Smart_Summarize/Smart_Summarize.csproj b/Smart_Summarize/Smart_Summarize.csproj
index 3aad4f5..204faa9 100644
--- a/Smart_Summarize/Smart_Summarize.csproj
+++ b/Smart_Summarize/Smart_Summarize.csproj
@@ -8,9 +8,6 @@
true
-
-
-
@@ -21,10 +18,9 @@
+
-
-
-
+