Skip to content

Agent

generate_report(request, orchestrator=Depends(get_orchestrator)) async

Generates the Executive Report as required by the GenAI Challenge.

This endpoint acts as the main entry point for the AI reporting workflow. It constructs a strict system prompt and delegates execution to the SRAGAgentOrchestrator.

Workflow:

  1. Prompt Construction: Injects a structured prompt ensuring the Agent calculates 4 key metrics (Case Increase, Mortality, ICU, Vaccination) and generates 2 mandatory charts (Trend, History).
  2. Context Injection: If focus_area is provided in the request, it is appended to the prompt to guide the news search.
  3. Execution: Runs the ReAct agent loop.
  4. Artifact Handling: Extracts generated plot filenames from the agent's output and uploads artifacts (text + plots) to storage (MinIO/S3).

Parameters:

Name Type Description Default
request ReportRequest

Input payload containing focus areas.

 required
orchestrator SRAGAgentOrchestrator

The dependency-injected agent orchestrator.

 Depends(get_orchestrator)

Returns:

Name Type Description
AgentResponse

Object containing the generated markdown, list of plot paths,

and execution metrics.

Raises:

Type Description
HTTPException

If the agent execution fails or an internal error occurs (500).
Source code in api/src/routers/agent.py 
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
@router.post("/report", response_model=AgentResponse)
async def generate_report(
    request: ReportRequest,
    orchestrator: SRAGAgentOrchestrator = Depends(get_orchestrator),
):
    """
    Generates the Executive Report as required by the GenAI Challenge.

    This endpoint acts as the main entry point for the AI reporting workflow.
    It constructs a strict system prompt and delegates execution to the `SRAGAgentOrchestrator`.

    **Workflow:**

    1.  **Prompt Construction**: Injects a structured prompt ensuring the Agent calculates
        4 key metrics (Case Increase, Mortality, ICU, Vaccination) and generates
        2 mandatory charts (Trend, History).
    2.  **Context Injection**: If `focus_area` is provided in the request, it is appended
        to the prompt to guide the news search.
    3.  **Execution**: Runs the ReAct agent loop.
    4.  **Artifact Handling**: Extracts generated plot filenames from the agent's output
        and uploads artifacts (text + plots) to storage (MinIO/S3).

    Args:
        request (ReportRequest): Input payload containing focus areas.
        orchestrator (SRAGAgentOrchestrator): The dependency-injected agent orchestrator.

    Returns:
        AgentResponse: Object containing the generated markdown, list of plot paths,
        and execution metrics.

    Raises:
        HTTPException: If the agent execution fails or an internal error occurs (500).
    """
    start_time = time.time()

    tags = {
        "genai.task": "executive_report",
        "genai.focus_area": request.focus_area if request.focus_area else "general",
    }
    set_trace_tags(tags)

    report_prompt = (
        "Generate a comprehensive **Executive Report** on the current **SRAG situation**.\n"
        "STRICTLY FOLLOW this structure and formatting rules:\n\n"
        "1. **KEY METRICS**: Calculate and present:\n"
        "   - Case Increase Rate (Growth %)\n"
        "   - Mortality Rate (CFR %)\n"
        "   - ICU Occupation Rate (%)\n"
        "   - Vaccination Rate for COVID-19 (%)\n"
        "   - Vaccination Rate for Influenza (Flu) (%)\n\n"
        "2. **VISUAL ANALYSIS** (Mandatory):\n"
        "   - You MUST generate two charts: 'trend_30d' and 'history_12m'.\n"
        "   - **IMPORTANT:** When embedding charts, use the EXACT filename returned by the tool (including .png):\n"
        "     `![Desc](/api/v1/plots/<exact_filename_from_tool_output>)`\n"
        "   - Do NOT use the local 'data/plots/' path in the Markdown link.\n\n"
        "3. **CONTEXTUAL ANALYSIS**:\n"
        "   Make web search to find relevant news (e.g., outbreaks, variants, public health events)\n"
        "   to provide context and support for the presented metrics, helping explain observed trends\n"
        "   and anomalies in the data.\n\n"
        "4. **CONCLUSION**: Brief executive summary. "
        "5. **OUTPUT FORMAT**:\n"
        "   - The report includes only the analytical sections defined above.\n"
        "   - The content is written in a neutral, report-style format.\n"
        "   - The document finishes at the conclusion section."
    )

    if request.focus_area:
        report_prompt += (
            f"\n\nAdditional Focus: Please specifically analyze {request.focus_area}."
        )

    try:
        result = await orchestrator.run(report_prompt)

        generated_plots = extract_plots_from_result(result)
        response_text = result.output

        upload_run_artifacts(response_text, generated_plots)

        return AgentResponse(
            response=response_text,
            plots=generated_plots,
            execution_time=time.time() - start_time,
        )

    except Exception as e:
        logger.error(f"Report generation failed: {e}", exc_info=True)
        raise HTTPException(status_code=500, detail=str(e))