Charts Bigger, Duplicate Headings

app/DocumentGeneration/chart_generator.py

@@ -89,7 +89,7 @@ def create_enhanced_chart_style(chart_type, data, title, business_idea):
     colors = ['#6A2EAB', '#5A2E9B', '#C0C0C0', '#FF8C00', '#FFD700', '#696969', '#FF4500']
     
     if chart_type == "DC":  # Doughnut Chart
-        fig, ax = plt.subplots(figsize=(8, 6))
+        fig, ax = plt.subplots(figsize=(14, 10))
         
         # Extract labels and values from data
         labels = [item[0] for item in data]
@@ -98,15 +98,16 @@ def create_enhanced_chart_style(chart_type, data, title, business_idea):
         # Create doughnut chart with enhanced styling
         wedges, texts, autotexts = ax.pie(values, labels=labels, colors=colors[:len(values)],
                                          autopct='%1.1f%%', startangle=90,
-                                         wedgeprops=dict(width=0.6, edgecolor='white', linewidth=2))
+                                         wedgeprops=dict(width=0.6, edgecolor='white', linewidth=2),
+                                         textprops={'fontsize': 18})
         
         # Enhanced title styling
-        ax.set_title(title, fontsize=16, fontweight='bold', pad=20, color='#374151')
+        ax.set_title(title, fontsize=24, fontweight='bold', pad=20, color='#374151')
         
         # Add legend with enhanced styling
         legend = ax.legend(wedges, labels, title="Categories", 
                           loc="center left", bbox_to_anchor=(1, 0, 0.5, 1),
-                          fontsize=10, title_fontsize=12)
+                          fontsize=16, title_fontsize=18)
         legend.get_title().set_color('#FF6B6B')
         
         # Add business name watermark
@@ -117,7 +118,7 @@ def create_enhanced_chart_style(chart_type, data, title, business_idea):
         return fig
         
     elif chart_type == "BG":  # Bar Graph
-        fig, ax = plt.subplots(figsize=(10, 6))
+        fig, ax = plt.subplots(figsize=(16, 10))
         
         # Extract labels and values
         labels = [item[0] for item in data]
@@ -131,15 +132,16 @@ def create_enhanced_chart_style(chart_type, data, title, business_idea):
         for bar, value in zip(bars, values):
             height = bar.get_height()
             ax.text(bar.get_x() + bar.get_width()/2., height + max(values)*0.01,
-                   f'{value}', ha='center', va='bottom', fontweight='bold')
+                   f'{value}', ha='center', va='bottom', fontweight='bold', fontsize=18)
         
         # Enhanced styling
-        ax.set_title(title, fontsize=16, fontweight='bold', pad=20, color='#374151')
-        ax.set_xlabel('Categories', fontsize=12, color='#6B7280')
-        ax.set_ylabel('Values', fontsize=12, color='#6B7280')
+        ax.set_title(title, fontsize=24, fontweight='bold', pad=20, color='#374151')
+        ax.set_xlabel('Categories', fontsize=18, color='#6B7280')
+        ax.set_ylabel('Values', fontsize=18, color='#6B7280')
         
         # Rotate x-axis labels for better readability
-        plt.setp(ax.get_xticklabels(), rotation=45, ha='right')
+        plt.setp(ax.get_xticklabels(), rotation=45, ha='right', fontsize=16)
+        plt.setp(ax.get_yticklabels(), fontsize=16)
         
         # Add grid for better readability
         ax.yaxis.grid(True, alpha=0.3)
@@ -153,7 +155,7 @@ def create_enhanced_chart_style(chart_type, data, title, business_idea):
         return fig
         
     elif chart_type == "LG":  # Line Graph
-        fig, ax = plt.subplots(figsize=(10, 6))
+        fig, ax = plt.subplots(figsize=(16, 10))
         
         # Extract data
         x_values = [item[0] for item in data]
@@ -167,14 +169,18 @@ def create_enhanced_chart_style(chart_type, data, title, business_idea):
         ax.fill_between(x_values, y_values, alpha=0.3, color='#4ECDC4')
         
         # Enhanced styling
-        ax.set_title(title, fontsize=16, fontweight='bold', pad=20, color='#374151')
-        ax.set_xlabel('Time Period', fontsize=12, color='#6B7280')
-        ax.set_ylabel('Values', fontsize=12, color='#6B7280')
+        ax.set_title(title, fontsize=24, fontweight='bold', pad=20, color='#374151')
+        ax.set_xlabel('Time Period', fontsize=18, color='#6B7280')
+        ax.set_ylabel('Values', fontsize=18, color='#6B7280')
         
         # Add grid
         ax.grid(True, alpha=0.3)
         ax.set_axisbelow(True)
         
+        # Increase tick label font sizes
+        plt.setp(ax.get_xticklabels(), fontsize=16)
+        plt.setp(ax.get_yticklabels(), fontsize=16)
+        
         # Add business name watermark
         fig.text(0.5, 0.02, f"Generated for: {business_idea}", 
                 ha='center', va='bottom', fontsize=8, color='#9CA3AF', style='italic')
@@ -183,7 +189,7 @@ def create_enhanced_chart_style(chart_type, data, title, business_idea):
         return fig
         
     elif chart_type == "PG":  # Pie Graph
-        fig, ax = plt.subplots(figsize=(8, 6))
+        fig, ax = plt.subplots(figsize=(14, 10))
         
         # Extract labels and values
         labels = [item[0] for item in data]
@@ -192,10 +198,11 @@ def create_enhanced_chart_style(chart_type, data, title, business_idea):
         # Create pie chart with enhanced styling
         wedges, texts, autotexts = ax.pie(values, labels=labels, colors=colors[:len(values)],
                                          autopct='%1.1f%%', startangle=90,
-                                         wedgeprops=dict(edgecolor='white', linewidth=2))
+                                         wedgeprops=dict(edgecolor='white', linewidth=2),
+                                         textprops={'fontsize': 18})
         
         # Enhanced title styling
-        ax.set_title(title, fontsize=16, fontweight='bold', pad=20, color='#374151')
+        ax.set_title(title, fontsize=24, fontweight='bold', pad=20, color='#374151')
         
         # Add business name watermark
         fig.text(0.5, 0.02, f"Generated for: {business_idea}", 
@@ -221,7 +228,7 @@ def create_chart_from_data(chart_info: Dict[str, Any], business_idea: str) -> by
         if fig is None:
             # Fallback to original method if enhanced styling fails
             if chart_type == "DC":  # Doughnut Chart
-                fig, ax = plt.subplots(figsize=(6, 4.5))  # Smaller size for better page fit
+                fig, ax = plt.subplots(figsize=(14, 10))  # Larger size for better visibility
                 
                 # Extract labels and values from data
                 labels = [item[0] for item in data]
@@ -233,17 +240,18 @@ def create_chart_from_data(chart_info: Dict[str, Any], business_idea: str) -> by
                 # Create doughnut chart
                 wedges, texts, autotexts = ax.pie(values, labels=labels, colors=colors[:len(values)],
                                                  autopct='%1.1f%%', startangle=90,
-                                                 wedgeprops=dict(width=0.6, edgecolor='white', linewidth=2))
+                                                 wedgeprops=dict(width=0.6, edgecolor='white', linewidth=2),
+                                                 textprops={'fontsize': 18})
                 
                 # Enhanced title
-                ax.set_title(chart_title, fontsize=14, fontweight='bold', pad=15, color='#374151')
+                ax.set_title(chart_title, fontsize=22, fontweight='bold', pad=15, color='#374151')
                 
                 # Add legend
-                ax.legend(wedges, labels, title="Categories", loc="center left", bbox_to_anchor=(1, 0, 0.5, 1))
+                ax.legend(wedges, labels, title="Categories", loc="center left", bbox_to_anchor=(1, 0, 0.5, 1), fontsize=16)
                 
                 plt.tight_layout()
             elif chart_type == "BG":  # Bar Graph
-                fig, ax = plt.subplots(figsize=(8, 5))
+                fig, ax = plt.subplots(figsize=(16, 10))
                 
                 # Extract labels and values
                 labels = [item[0] for item in data]
@@ -260,14 +268,15 @@ def create_chart_from_data(chart_info: Dict[str, Any], business_idea: str) -> by
                 for bar, value in zip(bars, values):
                     height = bar.get_height()
                     ax.text(bar.get_x() + bar.get_width()/2., height + max(values)*0.01,
-                           f'{value}', ha='center', va='bottom', fontweight='bold')
+                           f'{value}', ha='center', va='bottom', fontweight='bold', fontsize=18)
                 
-                ax.set_title(chart_title, fontsize=14, fontweight='bold', pad=15, color='#374151')
-                ax.set_xlabel('Categories', fontsize=12, color='#6B7280')
-                ax.set_ylabel('Values', fontsize=12, color='#6B7280')
+                ax.set_title(chart_title, fontsize=22, fontweight='bold', pad=15, color='#374151')
+                ax.set_xlabel('Categories', fontsize=18, color='#6B7280')
+                ax.set_ylabel('Values', fontsize=18, color='#6B7280')
                 
                 # Rotate x-axis labels
-                plt.setp(ax.get_xticklabels(), rotation=45, ha='right')
+                plt.setp(ax.get_xticklabels(), rotation=45, ha='right', fontsize=16)
+                plt.setp(ax.get_yticklabels(), fontsize=16)
                 
                 # Add grid
                 ax.yaxis.grid(True, alpha=0.3)
@@ -275,7 +284,7 @@ def create_chart_from_data(chart_info: Dict[str, Any], business_idea: str) -> by
                 
                 plt.tight_layout()
             elif chart_type == "LG":  # Line Graph
-                fig, ax = plt.subplots(figsize=(8, 5))
+                fig, ax = plt.subplots(figsize=(16, 10))
                 
                 # Extract data
                 x_values = [item[0] for item in data]
@@ -288,17 +297,21 @@ def create_chart_from_data(chart_info: Dict[str, Any], business_idea: str) -> by
                 # Fill area under line
                 ax.fill_between(x_values, y_values, alpha=0.3, color='#4ECDC4')
                 
-                ax.set_title(chart_title, fontsize=14, fontweight='bold', pad=15, color='#374151')
-                ax.set_xlabel('Time Period', fontsize=12, color='#6B7280')
-                ax.set_ylabel('Values', fontsize=12, color='#6B7280')
+                ax.set_title(chart_title, fontsize=22, fontweight='bold', pad=15, color='#374151')
+                ax.set_xlabel('Time Period', fontsize=18, color='#6B7280')
+                ax.set_ylabel('Values', fontsize=18, color='#6B7280')
                 
                 # Add grid
                 ax.grid(True, alpha=0.3)
                 ax.set_axisbelow(True)
                 
+                # Increase tick label font sizes
+                plt.setp(ax.get_xticklabels(), fontsize=16)
+                plt.setp(ax.get_yticklabels(), fontsize=16)
+                
                 plt.tight_layout()
             elif chart_type == "PG":  # Pie Graph
-                fig, ax = plt.subplots(figsize=(6, 4.5))
+                fig, ax = plt.subplots(figsize=(14, 10))
                 
                 # Extract labels and values
                 labels = [item[0] for item in data]
@@ -310,9 +323,10 @@ def create_chart_from_data(chart_info: Dict[str, Any], business_idea: str) -> by
                 # Create pie chart
                 wedges, texts, autotexts = ax.pie(values, labels=labels, colors=colors[:len(values)],
                                                  autopct='%1.1f%%', startangle=90,
-                                                 wedgeprops=dict(edgecolor='white', linewidth=2))
+                                                 wedgeprops=dict(edgecolor='white', linewidth=2),
+                                                 textprops={'fontsize': 18})
                 
-                ax.set_title(chart_title, fontsize=14, fontweight='bold', pad=15, color='#374151')
+                ax.set_title(chart_title, fontsize=22, fontweight='bold', pad=15, color='#374151')
                 
                 plt.tight_layout()
             else:

app/DocumentGeneration/document_helpers.py

@@ -113,17 +113,46 @@ def remove_duplicate_headings(content: str) -> str:
             continue
         
         # Check if this is any type of heading
+        is_heading = False
+        heading_text = ""
+        
+        # Markdown headings (# ## ### etc.)
         if line.startswith('#'):
-            heading_text = re.sub(r'^#+\s*', '', line)
-            # heading_text = re.sub(r'[*_`~]', '', heading_text).strip()  # Commented out to preserve ** symbols
-            heading_text = heading_text.strip()
+            heading_text = re.sub(r'^#+\s*', '', line).strip()
+            is_heading = True
+        
+        # Bold headings (**text**)
+        elif line.startswith('**') and line.endswith('**') and len(line) > 4:
+            heading_text = line[2:-2].strip()  # Remove ** symbols
+            is_heading = True
+        
+        # Check for numbered headings (e.g., "8.1 Management Team")
+        elif re.match(r'^\d+\.?\d*\s+', line):
+            heading_text = re.sub(r'^\d+\.?\d*\s+', '', line).strip()
+            is_heading = True
+        
+        if is_heading and heading_text:
+            # Normalize the heading text for comparison
+            normalized_text = heading_text.lower().strip()
             
-            # Check for duplicates (case-insensitive)
-            if heading_text.lower() in seen_headings:
-                continue
+            # Also check for variations with "and" vs "&"
+            variations = [
+                normalized_text,
+                normalized_text.replace(' & ', ' and '),
+                normalized_text.replace(' and ', ' & '),
+                normalized_text.replace('&', 'and'),
+                normalized_text.replace('and', '&')
+            ]
             
-            seen_headings.add(heading_text.lower())
-            cleaned_lines.append(line)
+            # Check if any variation has been seen before
+            is_duplicate = any(var in seen_headings for var in variations)
+            
+            if not is_duplicate:
+                # Add all variations to seen set
+                for var in variations:
+                    seen_headings.add(var)
+                cleaned_lines.append(line)
+            # Skip duplicate headings
         else:
             cleaned_lines.append(line)
     

app/DocumentGeneration/pdf_generator.py

@@ -161,7 +161,7 @@ def calculate_real_page_numbers(request: ExportRequest, filtered_sections):
             
             for chart_info in section_charts:
                 try:
-                    chart_height_needed = 95  # Chart height (75) + title (20) + minimal spacing
+                    chart_height_needed = 140  # Chart height (120) + title (20) + minimal spacing
                     current_y = temp_pdf.get_y()
                     charts_per_page += 1
                     
@@ -191,7 +191,7 @@ def calculate_real_page_numbers(request: ExportRequest, filtered_sections):
                     temp_pdf.line(25, y_title + 14, 25 + line_width, y_title + 14)
                     
                     # Add chart image
-                    img_width = 100
+                    img_width = 160  # Much larger chart width
                     x = (210 - img_width) / 2
                     y_img = y_title + 20
                     
@@ -201,7 +201,7 @@ def calculate_real_page_numbers(request: ExportRequest, filtered_sections):
                         temp_img_path = tmp_img.name
                     
                     # Add chart
-                    chart_height = 75
+                    chart_height = 120  # Much larger chart height
                     temp_pdf.image(temp_img_path, x=x, y=y_img, w=img_width)
                     temp_pdf.ln(chart_height + 5)
                     
@@ -455,24 +455,60 @@ def create_pdf_document(request: ExportRequest):
             
             # Remove the section title from content to prevent duplication (since we already added it above)
             section_title_clean = section.title.strip()
-            print(f"DEBUG: Section title: '{section_title_clean}'")
-            print(f"DEBUG: Original content before filtering: {content_text[:300]}...")
             content_lines = content_text.split('\n')
             filtered_lines = []
             for line in content_lines:
                 line_clean = line.strip()
                 # Skip lines that match the section title (with or without markdown)
                 # Check for various markdown formats and variations
-                if (line_clean == section_title_clean or 
-                    line_clean == f"**{section_title_clean}**" or
-                    line_clean == f"# {section_title_clean}" or
-                    line_clean == f"## {section_title_clean}" or
-                    line_clean == f"### {section_title_clean}" or
-                    line_clean == f"#### {section_title_clean}" or
-                    # Also check if the line starts with the section title (for numbered headings)
-                    line_clean.startswith(f"{section_title_clean}") and len(line_clean) <= len(section_title_clean) + 10):
-                    continue
-                filtered_lines.append(line)
+                should_skip = False
+                
+                # Exact matches
+                if line_clean == section_title_clean:
+                    should_skip = True
+                
+                # Markdown heading variations
+                elif (line_clean == f"**{section_title_clean}**" or
+                      line_clean == f"# {section_title_clean}" or
+                      line_clean == f"## {section_title_clean}" or
+                      line_clean == f"### {section_title_clean}" or
+                      line_clean == f"#### {section_title_clean}" or
+                      line_clean == f"##### {section_title_clean}" or
+                      line_clean == f"###### {section_title_clean}"):
+                    should_skip = True
+                
+                # Numbered headings (e.g., "8.1 Management Team")
+                elif (line_clean.startswith(f"{section_title_clean}") and 
+                      len(line_clean) <= len(section_title_clean) + 15):
+                    should_skip = True
+                
+                # Check for headings that contain the section title with minimal additional text
+                elif (section_title_clean.lower() in line_clean.lower() and 
+                      len(line_clean) <= len(section_title_clean) + 20 and
+                      (line_clean.startswith('#') or line_clean.startswith('**'))):
+                    should_skip = True
+                
+                # Check for variations with "and" vs "&" (e.g., "Management Team" vs "Management and Team")
+                title_variations = [
+                    section_title_clean.replace(' & ', ' and '),
+                    section_title_clean.replace(' and ', ' & '),
+                    section_title_clean.replace('&', 'and'),
+                    section_title_clean.replace('and', '&')
+                ]
+                
+                for variation in title_variations:
+                    if (line_clean == variation or
+                        line_clean == f"**{variation}**" or
+                        line_clean == f"# {variation}" or
+                        line_clean == f"## {variation}" or
+                        line_clean == f"### {variation}" or
+                        line_clean == f"#### {variation}"):
+                        should_skip = True
+                        break
+                
+                if not should_skip:
+                    filtered_lines.append(line)
+            
             content_text = '\n'.join(filtered_lines)
             
             # Skip enhance_subheadings_with_bold for grouped rendering to preserve ** markers
@@ -549,7 +585,7 @@ def create_pdf_document(request: ExportRequest):
                 for chart_info in section_charts:
                     try:
                         # Check if we need a new page for this chart
-                        chart_height_needed = 95  # Chart height (75) + title (20) + minimal spacing
+                        chart_height_needed = 140  # Chart height (120) + title (20) + minimal spacing
                         
                         # Smart page break logic - allow multiple charts per page
                         current_y = pdf.get_y()
@@ -571,8 +607,8 @@ def create_pdf_document(request: ExportRequest):
                         # Subheader (title) and subtle rule
                         y_title = add_subheader(pdf, chart_info["title"], 16, y_title)
 
-                        # Centered image with minimal border - optimized for more charts per page
-                        img_width = 100  # Optimized for smaller charts (6x4.5 ratio)
+                        # Centered image with larger dimensions for better visibility
+                        img_width = 160  # Much larger chart width for better visibility
                         x = (210 - img_width) / 2
                         y_img = y_title
 
@@ -582,14 +618,14 @@ def create_pdf_document(request: ExportRequest):
                             temp_img_path = tmp_img.name
 
                         # Chart dimensions (no border needed)
-                        chart_height = 75  # Optimized height for 6:4.5 aspect ratio (100 * 4.5/6)
+                        chart_height = 120  # Much larger height for better visibility
                         chart_width = img_width
 
                         # Image
                         pdf.image(temp_img_path, x=x, y=y_img, w=chart_width)
 
                         # Move below image and clean up (spacing based on actual chart height)
-                        pdf.ln(chart_height + 5)  # Chart height (75) + minimal spacing between charts
+                        pdf.ln(chart_height + 5)  # Chart height (120) + minimal spacing between charts
                         try:
                             os.remove(temp_img_path)
                         except Exception: