Update app.py

app.py

@@ -25,105 +25,35 @@ def extract_text_from_pdf(pdf_path):
         return f"Error extracting text from PDF: {str(e)}"
 
 # Function to generate PRISMA flow chart HTML
-def generate_prisma_flow_chart():
-    # Placeholder values for the PRISMA flow chart (as per the provided image)
-    records_db = 0  # Records from databases
-    records_other = 0  # Records from other sources
-    duplicates_removed = 0  # After duplicates removed
-    records_screened = 0  # Records screened
-    records_excluded_screening = 0  # Records excluded during screening
-    reports_sought = 0  # Reports sought for retrieval
-    reports_not_retrieved = 0  # Reports not retrieved
-    reports_assessed = 0  # Reports assessed for eligibility
-    reports_excluded_eligibility = 0  # Reports excluded during eligibility
-    studies_included = 0  # Studies included in review
-
-    # HTML and CSS for PRISMA flow chart
+def generate_prisma_flow_chart(total_records_db, additional_records, duplicates_removed, screened_records, excluded_title_abstract, full_text_assessed, excluded_full_text, included_studies):
     prisma_html = f"""
-    <div class="prisma-flowchart" style="margin: 40px 0; font-family: 'Times New Roman', Times, serif; font-size: 14px;">
-        <h3 style="text-align: center; margin-bottom: 20px;">PRISMA Flow Diagram</h3>
-        <div style="display: flex; justify-content: space-between;">
-            <!-- Left Side: Database Search -->
-            <div style="width: 45%;">
-                <div style="background-color: #f0f0f0; padding: 10px; border: 1px solid #ccc; text-align: center;">
-                    Identification of new studies via databases and registers<br>
-                    Records identified from:<br>
-                    Databases (n = {records_db})<br>
-                    Registers (n = 0)
-                </div>
-                <div style="text-align: center; margin: 10px 0;">↓</div>
-                <div style="background-color: #f0f0f0; padding: 10px; border: 1px solid #ccc; text-align: center;">
-                    Records removed before screening:<br>
-                    Duplicate records (n = 0)<br>
-                    Records marked as ineligible by automation tools (n = 0)<br>
-                    Records removed for other reasons (n = 0)
-                </div>
-                <div style="text-align: center; margin: 10px 0;">↓</div>
-                <div style="background-color: #f0f0f0; padding: 10px; border: 1px solid #ccc; text-align: center;">
-                    Records screened (n = {records_screened})
-                </div>
-                <div style="display: flex; justify-content: space-between; align-items: center;">
-                    <div style="width: 50%; text-align: center; margin: 10px 0;">↓</div>
-                    <div style="width: 50%; background-color: #f0f0f0; padding: 10px; border: 1px solid #ccc; text-align: center;">
-                        Records excluded (n = {records_excluded_screening})
-                    </div>
-                </div>
-                <div style="background-color: #f0f0f0; padding: 10px; border: 1px solid #ccc; text-align: center;">
-                    Reports sought for retrieval (n = {reports_sought})
-                </div>
-                <div style="display: flex; justify-content: space-between; align-items: center;">
-                    <div style="width: 50%; text-align: center; margin: 10px 0;">↓</div>
-                    <div style="width: 50%; background-color: #f0f0f0; padding: 10px; border: 1px solid #ccc; text-align: center;">
-                        Reports not retrieved (n = {reports_not_retrieved})
-                    </div>
-                </div>
-                <div style="background-color: #f0f0f0; padding: 10px; border: 1px solid #ccc; text-align: center;">
-                    Reports assessed for eligibility (n = {reports_assessed})
-                </div>
-                <div style="display: flex; justify-content: space-between; align-items: center;">
-                    <div style="width: 50%; text-align: center; margin: 10px 0;">↓</div>
-                    <div style="width: 50%; background-color: #f0f0f0; padding: 10px; border: 1px solid #ccc; text-align: center;">
-                        Reports excluded:<br>
-                        Reason 1 (n = NA)<br>
-                        Reason 2 (n = NA)
-                    </div>
-                </div>
-                <div style="background-color: #f0f0f0; padding: 10px; border: 1px solid #ccc; text-align: center;">
-                    New studies included in review (n = {studies_included})<br>
-                    Reports of new studies (n = 0)
-                </div>
+    <div style="text-align: center; margin: 20px auto; max-width: 800px;">
+        <h2>PRISMA Flow Diagram</h2>
+        <div style="border: 1px solid #000; padding: 10px; background-color: #f9f9f9;">
+            <div style="margin: 10px 0;">
+                <strong>Identification of studies via databases and registers</strong>
+                <div>Records identified from: Databases (n = {total_records_db})</div>
+                <div>Records removed before screening: Duplicates marked as ineligible by automation tools (n = {duplicates_removed})</div>
             </div>
-
-            <!-- Right Side: Other Methods -->
-            <div style="width: 45%;">
-                <div style="background-color: #f0f0f0; padding: 10px; border: 1px solid #ccc; text-align: center;">
-                    Identification of new studies via other methods<br>
-                    Records identified from:<br>
-                    Websites (n = 0)<br>
-                    Organisations (n = 0)<br>
-                    Citation searching (n = 0)
-                </div>
-                <div style="text-align: center; margin: 10px 0;">↓</div>
-                <div style="background-color: #f0f0f0; padding: 10px; border: 1px solid #ccc; text-align: center;">
-                    Reports sought for retrieval (n = {reports_sought})
-                </div>
-                <div style="display: flex; justify-content: space-between; align-items: center;">
-                    <div style="width: 50%; text-align: center; margin: 10px 0;">↓</div>
-                    <div style="width: 50%; background-color: #f0f0f0; padding: 10px; border: 1px solid #ccc; text-align: center;">
-                        Reports not retrieved (n = {reports_not_retrieved})
-                    </div>
-                </div>
-                <div style="background-color: #f0f0f0; padding: 10px; border: 1px solid #ccc; text-align: center;">
-                    Reports assessed for eligibility (n = {reports_assessed})
-                </div>
-                <div style="display: flex; justify-content: space-between; align-items: center;">
-                    <div style="width: 50%; text-align: center; margin: 10px 0;">↓</div>
-                    <div style="width: 50%; background-color: #f0f0f0; padding: 10px; border: 1px solid #ccc; text-align: center;">
-                        Reports excluded:<br>
-                        Reason 1 (n = NA)<br>
-                        Reason 2 (n = NA)
-                    </div>
-                </div>
+            <div style="margin: 10px 0;">
+                <strong>Screening</strong>
+                <div>Records screened (n = {screened_records})</div>
+                <div>Records excluded (n = {excluded_title_abstract})</div>
+            </div>
+            <div style="margin: 10px 0;">
+                <strong>Eligibility</strong>
+                <div>Reports sought for retrieval (n = {full_text_assessed})</div>
+                <div>Reports not retrieved (n = 0)</div>
+                <div>Reports assessed for eligibility (n = {full_text_assessed})</div>
+                <div>Reports excluded (n = {excluded_full_text})</div>
+            </div>
+            <div style="margin: 10px 0;">
+                <strong>Included</strong>
+                <div>New studies included in review (n = {included_studies})</div>
+            </div>
+            <div style="margin: 10px 0;">
+                <strong>Identification of studies via other methods</strong>
+                <div>Records identified from: Websites (n = 0), Organizations (n = 0), Citation searching (n = {additional_records})</div>
             </div>
         </div>
     </div>
@@ -145,7 +75,7 @@ def generate_systematic_review(pdf_files, review_question, include_tables=True):
         openai.api_key = api_key
         
         # Create the system message with systematic review guidelines
-        system_prompt = """You are an expert academic assistant. Create a systematic review using academic research paper formatting. The Systematic Review must be in great detail. Structure it using these steps:
+        system_prompt = """You are an expert academic assistant. Create a systematic review using academic research paper formatting. The Systematic Review must be in great details. Structure it using these steps:
         
         Step 1: Identify a Research Field
         The first step in writing a systematic review paper is to identify a research field. This involves selecting a specific area of study that you are interested in and want to explore further.
@@ -202,6 +132,18 @@ def generate_systematic_review(pdf_files, review_question, include_tables=True):
             pdf_texts.append(pdf_text)
             pdf_names.append(pdf_name)
         
+        # Prepare PRISMA flow chart data (example values, adjust based on actual process)
+        total_records_db = len(pdf_files)
+        additional_records = 0
+        duplicates_removed = 0
+        screened_records = total_records_db
+        excluded_title_abstract = 0
+        full_text_assessed = total_records_db
+        excluded_full_text = 0
+        included_studies = total_records_db
+        
+        prisma_chart = generate_prisma_flow_chart(total_records_db, additional_records, duplicates_removed, screened_records, excluded_title_abstract, full_text_assessed, excluded_full_text, included_studies)
+        
         # Prepare the user prompt with the review question and instructions
         table_instruction = ""
         if include_tables:
@@ -227,16 +169,7 @@ def generate_systematic_review(pdf_files, review_question, include_tables=True):
         # Get the AI response
         review_content = response["choices"][0]["message"]["content"]
         
-        # Generate PRISMA flow chart
-        prisma_flow_chart = generate_prisma_flow_chart()
-        
-        # Insert PRISMA flow chart into the review content (after the methodology section, assuming section 3 is methodology)
-        review_content = review_content.replace(
-            '<h2>4. Results</h2>',
-            '<h2>4. Results</h2>\n' + prisma_flow_chart
-        )
-        
-        # Apply professional academic paper styling
+        # Apply professional academic paper styling with PRISMA chart in the middle
         styled_html = f"""
         <!DOCTYPE html>
         <html lang="en">
@@ -373,6 +306,8 @@ def generate_systematic_review(pdf_files, review_question, include_tables=True):
         <body>
             <div class="paper-container">
                 {review_content}
+                <!-- Insert PRISMA Flow Chart in the middle -->
+                {prisma_chart}
             </div>
         </body>
         </html>
@@ -564,7 +499,6 @@ with gr.Blocks(css=custom_css) as demo:
         - Upload papers that are related to the same research topic or field
         - Be specific in your review question to get more focused results
         - The generated review will follow a systematic structure including research field identification, data extraction, analysis, and conclusions
-        - A PRISMA flow chart will be included in the review to visualize the study selection process
         - The more papers you upload, the more comprehensive the review will be
         - The review will be formatted as a professional academic paper with proper sections and citations
         """)