Add FastMCP tool for web scrapping.

Add logging info for tools calling
2026-07-29 16:28:15 +09:00
parent e5bb49a596
commit afb10aa52e
8 changed files with 606 additions and 73 deletions
@@ -0,0 +1,133 @@
+import markdown
+from weasyprint import HTML
+
+# ==========================================
+# REUSABLE HTML & CSS LAYOUT TEMPLATE
+# ==========================================
+PDF_REPORT_TEMPLATE = """<!DOCTYPE html>
+<html>
+<head>
+    <meta charset="utf-8">
+    <style>
+        @page {{
+            size: A4;
+            margin: 20mm;
+            @bottom-right {{
+                content: counter(page);
+                font-family: Arial, sans-serif;
+                font-size: 9pt;
+            }}
+        }}
+        
+        body {{
+            font-family: Arial, sans-serif;
+            color: #333333;
+            line-height: 1.6;
+            font-size: 11pt;
+        }}
+
+        h1 {{ font-size: 18pt; margin-top: 24pt; border-bottom: 1px solid #ddd; padding-bottom: 6px; }}
+        h2 {{ font-size: 14pt; margin-top: 18pt; color: #2c3e50; }}
+        h3 {{ font-size: 12pt; margin-top: 14pt; color: #7f8c8d; }}
+
+        /* --- Layout Tables --- */
+        table {{
+            width: 100%;
+            border-collapse: collapse;
+            margin: 15pt 0;
+        }}
+        th, td {{
+            border: 1px solid #dddddd;
+            padding: 8px;
+            text-align: left;
+        }}
+        th {{ background-color: #f8f9fa; font-weight: bold; }}
+        tr:nth-child(even) {{ background-color: #fdfdfd; }}
+
+        /* --- Links --- */
+        a {{ color: #3498db; text-decoration: none; }}
+        a:hover {{ text-decoration: underline; }}
+
+        /* ==========================================
+           DYNAMIC PAGED MEDIA TOC LAYOUT
+        ========================================== */
+        .toc-wrapper {{
+            margin: 20pt 0;
+            padding: 15pt;
+            background: #fdfdfd;
+            border: 1px solid #eaeaea;
+            border-radius: 5px;
+            page-break-after: always;
+        }}
+        
+        /* Clear default markdown padding/bullets on root list element */
+        .toc-wrapper > ul {{
+            list-style-type: none;
+            padding-left: 0;
+        }}
+        
+        /* Main Level 1 entries (bolded for contrast) */
+        .toc-wrapper > ul > li {{
+            font-weight: bold;
+            margin-top: 12px;
+            margin-bottom: 6px;
+        }}
+
+        /* Indent Subsections (Level 2 nested lists) */
+        .toc-wrapper li ul {{
+            list-style-type: none;
+            padding-left: 25px;
+            margin-top: 4px;
+            margin-bottom: 4px;
+        }}
+        
+        /* Ensure subsection text lines are regular weight */
+        .toc-wrapper li ul li {{
+            font-weight: normal;
+            margin-bottom: 6px;
+        }}
+
+        .toc-wrapper li {{
+            position: relative;
+            display: block;
+        }}
+        
+        /* Pulls the target page number of the markdown heading reference link */
+        .toc-wrapper a::after {{
+            content: target-counter(attr(href), page);
+            float: right;
+            color: #7f8c8d;
+            font-weight: normal;
+        }}
+    </style>
+</head>
+<body>
+    <div class="content">
+        {html_body}
+    </div>
+</body>
+</html>
+"""
+
+def export_to_pdf(llm_response: str, output_filename: str = "output.pdf") -> None:
+    """Converts structured LLM response with a fully clickable, indented manual TOC into a clean PDF."""
+    
+    # 1. Standard markdown conversion with 'toc' extension enabled.
+    # This automatically adds unique ID anchors to your <h1> and <h2> elements 
+    # so that clicking the TOC links will correctly jump down to that section.
+    html_body = markdown.markdown(
+        llm_response, 
+        extensions=['fenced_code', 'tables', 'toc']
+    )
+    
+    # 2. Inject a styling hook class wrapper around your Table of Contents list block
+    html_body = html_body.replace('<h3>Table of Contents</h3>', '<h3>Table of Contents</h3><div class="toc-wrapper">')
+    
+    # Close the div wrapper cleanly right before the next main topic heading begins
+    html_body = html_body.replace('<h1>1.', '</div><h1>1.')
+    
+    # 3. Format the final output document string structure
+    final_html_content = PDF_REPORT_TEMPLATE.format(html_body=html_body)
+    
+    # 4. Generate the clickable PDF file
+    HTML(string=final_html_content).write_pdf(output_filename)