feat: add PDF export for analysis reports

Add a new /export/{company_name}/pdf endpoint that generates a formatted
PDF report using reportlab, including a summary table and all analysis
results. Add the corresponding frontend Export PDF button alongside the
existing Export CSV button on the Analysis page.

Closes leeworks-agents/SPARC#85

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

SPARC/api.py

@@ -41,6 +41,7 @@ class CompanyAnalysisResponse(BaseModel):
     patent_count: int
     success: bool
     error: str | None = None
+    model: str | None = None
     timestamp: datetime
 
 
@@ -54,6 +55,15 @@ class BatchAnalysisResponse(BaseModel):
     timestamp: datetime
 
 
+class CompanyAnalysisRequest(BaseModel):
+    """Request model for single company analysis with optional model selection."""
+
+    model: str | None = Field(
+        default=None,
+        description="LLM model to use (e.g. 'anthropic/claude-3.5-sonnet', 'openai/gpt-4o'). Defaults to server config.",
+    )
+
+
 class BatchAnalysisRequest(BaseModel):
     """Request model for batch company analysis."""
 
@@ -63,6 +73,10 @@ class BatchAnalysisRequest(BaseModel):
     max_workers: int = Field(
         default=3, ge=1, le=5, description="Max concurrent analyses"
     )
+    model: str | None = Field(
+        default=None,
+        description="LLM model to use for all analyses in this batch. Defaults to server config.",
+    )
 
 
 class JobStatus(BaseModel):
@@ -140,6 +154,7 @@ def _convert_result(result: CompanyAnalysisResult) -> CompanyAnalysisResponse:
         patent_count=result.patent_count,
         success=result.success,
         error=result.error,
+        model=result.model,
         timestamp=result.timestamp,
     )
 
@@ -453,6 +468,32 @@ async def get_analytics(
     )
 
 
+# ============== Model Selection Endpoints ==============
+
+# Supported models via OpenRouter
+SUPPORTED_MODELS = [
+    {"id": "anthropic/claude-3.5-sonnet", "name": "Claude 3.5 Sonnet", "provider": "Anthropic"},
+    {"id": "openai/gpt-4o", "name": "GPT-4o", "provider": "OpenAI"},
+    {"id": "openai/gpt-4o-mini", "name": "GPT-4o Mini", "provider": "OpenAI"},
+    {"id": "google/gemini-pro-1.5", "name": "Gemini Pro 1.5", "provider": "Google"},
+    {"id": "meta-llama/llama-3.1-70b-instruct", "name": "Llama 3.1 70B", "provider": "Meta"},
+]
+
+
+@app.get("/models", tags=["System"])
+async def list_models():
+    """List supported LLM models for analysis.
+
+    Returns the available models that can be passed as the `model` field
+    in analysis requests. The default model is determined by the `MODEL`
+    environment variable on the server.
+    """
+    return {
+        "models": SUPPORTED_MODELS,
+        "default": config.model,
+    }
+
+
 @app.get("/analytics/trends", tags=["Analytics"])
 async def get_analytics_trends(
     days: int = Query(default=90, ge=7, le=365),
@@ -580,6 +621,164 @@ async def export_company_csv(
     )
 
 
+@app.get("/export/{company_name}/pdf", tags=["Export"])
+async def export_company_pdf(
+    company_name: str,
+    _: UserResponse = Depends(get_current_user),
+):
+    """Export analysis results for a company as a formatted PDF report.
+
+    Returns all stored analysis records for the given company, including
+    analysis type, model used, response text, and timestamp, formatted
+    as a downloadable PDF document.
+
+    Args:
+        company_name: Company name to export results for
+
+    Returns:
+        PDF file download
+    """
+    import io
+    import textwrap
+
+    from reportlab.lib import colors
+    from reportlab.lib.pagesizes import letter
+    from reportlab.lib.styles import ParagraphStyle, getSampleStyleSheet
+    from reportlab.lib.units import inch
+    from reportlab.platypus import (
+        Paragraph,
+        SimpleDocTemplate,
+        Spacer,
+        Table,
+        TableStyle,
+    )
+
+    db = get_db_client()
+    with db.get_conn() as conn:
+        with conn.cursor() as cur:
+            cur.execute(
+                """
+                SELECT company_name, analysis_type, model, response, timestamp
+                FROM llm_messages
+                WHERE LOWER(company_name) = LOWER(%s) AND is_cached = FALSE
+                ORDER BY timestamp DESC
+                """,
+                (company_name,),
+            )
+            rows = cur.fetchall()
+
+    if not rows:
+        raise HTTPException(status_code=404, detail=f"No analysis results found for '{company_name}'")
+
+    buffer = io.BytesIO()
+    doc = SimpleDocTemplate(
+        buffer,
+        pagesize=letter,
+        rightMargin=0.75 * inch,
+        leftMargin=0.75 * inch,
+        topMargin=0.75 * inch,
+        bottomMargin=0.75 * inch,
+    )
+
+    styles = getSampleStyleSheet()
+    title_style = ParagraphStyle(
+        "CustomTitle",
+        parent=styles["Title"],
+        fontSize=20,
+        spaceAfter=6,
+    )
+    subtitle_style = ParagraphStyle(
+        "Subtitle",
+        parent=styles["Normal"],
+        fontSize=11,
+        textColor=colors.grey,
+        spaceAfter=20,
+    )
+    heading_style = ParagraphStyle(
+        "SectionHeading",
+        parent=styles["Heading2"],
+        fontSize=13,
+        spaceBefore=16,
+        spaceAfter=8,
+        textColor=colors.HexColor("#1a1a2e"),
+    )
+    body_style = ParagraphStyle(
+        "BodyText",
+        parent=styles["Normal"],
+        fontSize=9,
+        leading=13,
+        spaceAfter=10,
+    )
+
+    elements = []
+
+    # Title and date
+    display_name = rows[0][0]  # Use the casing from the database
+    analysis_date = datetime.now().strftime("%Y-%m-%d")
+    elements.append(Paragraph(f"SPARC Analysis Report: {display_name}", title_style))
+    elements.append(Paragraph(f"Generated on {analysis_date}", subtitle_style))
+
+    # Summary table
+    summary_data = [
+        ["Total Analyses", str(len(rows))],
+        ["Analysis Types", ", ".join(sorted(set(r[1] for r in rows)))],
+        ["Models Used", ", ".join(sorted(set(r[2] for r in rows)))],
+    ]
+    summary_table = Table(summary_data, colWidths=[2 * inch, 4.5 * inch])
+    summary_table.setStyle(
+        TableStyle(
+            [
+                ("BACKGROUND", (0, 0), (0, -1), colors.HexColor("#f0f0f5")),
+                ("FONTNAME", (0, 0), (0, -1), "Helvetica-Bold"),
+                ("FONTSIZE", (0, 0), (-1, -1), 9),
+                ("PADDING", (0, 0), (-1, -1), 6),
+                ("GRID", (0, 0), (-1, -1), 0.5, colors.HexColor("#cccccc")),
+                ("VALIGN", (0, 0), (-1, -1), "TOP"),
+            ]
+        )
+    )
+    elements.append(summary_table)
+    elements.append(Spacer(1, 16))
+
+    # Individual analysis sections
+    for i, row in enumerate(rows, 1):
+        _, analysis_type, model, response, timestamp = row
+        ts_str = timestamp.strftime("%Y-%m-%d %H:%M:%S") if hasattr(timestamp, "strftime") else str(timestamp)
+
+        elements.append(
+            Paragraph(f"Analysis {i}: {analysis_type} (via {model})", heading_style)
+        )
+        elements.append(
+            Paragraph(f"<i>Performed: {ts_str}</i>", body_style)
+        )
+
+        # Wrap long response text into paragraphs, escaping XML special chars
+        safe_response = (
+            response.replace("&", "&amp;")
+            .replace("<", "&lt;")
+            .replace(">", "&gt;")
+        )
+        # Split into manageable paragraphs to avoid overflow
+        for line in safe_response.split("\n"):
+            if line.strip():
+                elements.append(Paragraph(line, body_style))
+            else:
+                elements.append(Spacer(1, 4))
+
+        elements.append(Spacer(1, 10))
+
+    doc.build(elements)
+    buffer.seek(0)
+
+    safe_name = company_name.replace(" ", "_").lower()
+    filename = f"{safe_name}-analysis-{analysis_date}.pdf"
+    return StreamingResponse(
+        iter([buffer.getvalue()]),
+        media_type="application/pdf",
+        headers={"Content-Disposition": f'attachment; filename="{filename}"'},
+    )
+
+
 # ============== System Endpoints ==============
 
 

SPARC/llm.py

@@ -40,12 +40,13 @@ class LLMAnalyzer:
         else:
             self.client = None
 
-    def analyze_patent_content(self, patent_content: str, company_name: str) -> str:
+    def analyze_patent_content(self, patent_content: str, company_name: str, model: str | None = None) -> str:
         """Analyze patent content to estimate company innovation and performance.
 
         Args:
           patent_content: Minimized patent text (abstract, claims, summary)
           company_name: Name of the company for context
+          model: Optional model override (e.g. "openai/gpt-4o"). Defaults to config.
 
         Returns:
           Analysis text describing innovation quality and potential impact
@@ -63,6 +64,8 @@ Patent Content:
 
 Provide a concise analysis (2-3 paragraphs) focusing on what this patent reveals about the company's technical direction and competitive advantage."""
 
+        effective_model = model or self.model
+
         if self.test_mode:
             logger.debug("TEST MODE - Prompt that would be sent to LLM:\n%s", prompt)
             return "[TEST MODE - No API call made]"
@@ -81,7 +84,7 @@ Provide a concise analysis (2-3 paragraphs) focusing on what this patent reveals
                     response=cached["response"],
                     company_name=company_name,
                     analysis_type="single_patent",
-                    model=self.model,
+                    model=effective_model,
                     metadata={
                         "patent_content_length": len(patent_content),
                         "cache_hit": True,
@@ -94,7 +97,7 @@ Provide a concise analysis (2-3 paragraphs) focusing on what this patent reveals
         # Call API if no cache hit and client is available
         if self.client:
             response = self.client.chat.completions.create(
-                model=self.model,
+                model=effective_model,
                 max_tokens=1024,
                 messages=[{"role": "user", "content": prompt}],
             )
@@ -106,7 +109,7 @@ Provide a concise analysis (2-3 paragraphs) focusing on what this patent reveals
                 response=response_text,
                 company_name=company_name,
                 analysis_type="single_patent",
-                model=self.model,
+                model=effective_model,
                 metadata={"patent_content_length": len(patent_content)},
                 token_usage={
                     "prompt_tokens": response.usage.prompt_tokens,
@@ -124,13 +127,13 @@ Provide a concise analysis (2-3 paragraphs) focusing on what this patent reveals
             response=placeholder,
             company_name=company_name,
             analysis_type="single_patent",
-            model=self.model,
+            model=effective_model,
             metadata={"patent_content_length": len(patent_content), "pending": True}
         )
         return placeholder
 
     def analyze_patent_portfolio(
-        self, patents_data: list[Dict[str, str]], company_name: str
+        self, patents_data: list[Dict[str, str]], company_name: str, model: str | None = None
     ) -> str:
         """Analyze multiple patents to estimate overall company performance.
 
@@ -165,13 +168,16 @@ Patent Portfolio:
 
 Provide a comprehensive analysis (4-5 paragraphs) with a final verdict on the company's innovation strength and performance outlook."""
 
+        effective_model = model or self.model
+
         if self.test_mode:
             logger.debug("TEST MODE - Portfolio prompt:\n%s", prompt)
             return "[TEST MODE]"
 
         metadata = {
             "patent_count": len(patents_data),
-            "patent_ids": [p['patent_id'] for p in patents_data]
+            "patent_ids": [p['patent_id'] for p in patents_data],
+            "model": effective_model,
         }
 
         # Check cache first
@@ -188,7 +194,7 @@ Provide a comprehensive analysis (4-5 paragraphs) with a final verdict on the co
                     response=cached["response"],
                     company_name=company_name,
                     analysis_type="portfolio",
-                    model=self.model,
+                    model=effective_model,
                     metadata={
                         **metadata,
                         "cache_hit": True,
@@ -202,7 +208,7 @@ Provide a comprehensive analysis (4-5 paragraphs) with a final verdict on the co
         if self.client:
             try:
                 response = self.client.chat.completions.create(
-                    model=self.model,
+                    model=effective_model,
                     max_tokens=2048,
                     messages=[{"role": "user", "content": prompt}],
                 )
@@ -215,7 +221,7 @@ Provide a comprehensive analysis (4-5 paragraphs) with a final verdict on the co
                     response=response_text,
                     company_name=company_name,
                     analysis_type="portfolio",
-                    model=self.model,
+                    model=effective_model,
                     metadata=metadata,
                     token_usage={
                         "prompt_tokens": response.usage.prompt_tokens,
@@ -235,7 +241,7 @@ Provide a comprehensive analysis (4-5 paragraphs) with a final verdict on the co
             response=placeholder,
             company_name=company_name,
             analysis_type="portfolio",
-            model=self.model,
+            model=effective_model,
             metadata={**metadata, "pending": True}
         )
         return placeholder

SPARC/types.py

@@ -24,6 +24,7 @@ class CompanyAnalysisResult:
     patent_count: int
     success: bool
     error: str | None = None
+    model: str | None = None
     timestamp: datetime = field(default_factory=datetime.now)
 
 

frontend/src/api/client.ts

@@ -141,6 +141,21 @@ export const exportApi = {
     link.remove();
     window.URL.revokeObjectURL(url);
   },
+  exportPdf: async (companyName: string): Promise<void> => {
+    const response = await api.get(`/export/${encodeURIComponent(companyName)}/pdf`, {
+      responseType: 'blob',
+    });
+    const safeName = companyName.toLowerCase().replace(/\s+/g, '_');
+    const date = new Date().toISOString().split('T')[0];
+    const url = window.URL.createObjectURL(new Blob([response.data], { type: 'application/pdf' }));
+    const link = document.createElement('a');
+    link.href = url;
+    link.setAttribute('download', `${safeName}-analysis-${date}.pdf`);
+    document.body.appendChild(link);
+    link.click();
+    link.remove();
+    window.URL.revokeObjectURL(url);
+  },
 };
 
 // Analytics API

frontend/src/pages/Analysis.tsx

@@ -110,6 +110,7 @@ export function Analysis() {
                 <h3 className="text-lg font-semibold text-text-primary">
                   AI Analysis Results
                 </h3>
+                <div className="flex items-center gap-2">
                   <button
                     onClick={() => exportApi.exportCsv(result.company_name)}
                     className="flex items-center gap-2 text-sm bg-primary/20 hover:bg-primary/30 text-primary font-medium px-3 py-1.5 rounded-lg transition-colors"
@@ -117,6 +118,14 @@ export function Analysis() {
                     <Download size={14} />
                     Export CSV
                   </button>
+                  <button
+                    onClick={() => exportApi.exportPdf(result.company_name)}
+                    className="flex items-center gap-2 text-sm bg-primary/20 hover:bg-primary/30 text-primary font-medium px-3 py-1.5 rounded-lg transition-colors"
+                  >
+                    <FileText size={14} />
+                    Export PDF
+                  </button>
+                </div>
               </div>
               <div className="prose prose-invert max-w-none">
                 <div className="text-text-primary whitespace-pre-wrap leading-relaxed">

requirements.txt

@@ -17,3 +17,4 @@ PyJWT
 slowapi
 apscheduler
 boto3
+reportlab