ShedBoxAI for Automated Reporting

data_sources:
  sales:
    type: csv
    path: "weekly_sales.csv"

  customers:
    type: csv
    path: "new_customers.csv"

  support:
    type: csv
    path: "support_metrics.csv"

ai_interface:
  model:
    type: rest
    url: "https://api.anthropic.com/v1/messages"
    method: POST
    headers:
      x-api-key: "${ANTHROPIC_API_KEY}"
      Content-Type: "application/json"
    options:
      model: "claude-sonnet-4-20250514"

  prompts:
    weekly_report:
      system: "You are a business analyst writing a weekly report."
      user_template: |
        Generate a weekly business report:

        Sales Data:
        {% for sale in sales %}
        - {{ sale.date }}: ${{ sale.amount }}, {{ sale.product }}
        {% endfor %}

        New Customers: {{ customers | length }}
        Support Tickets: {{ support | length }}

        Include: key highlights, trends, and recommendations.

output:
  type: file
  path: "weekly_report.md"
  format: json

data_sources:
  metrics:
    type: csv
    path: "daily_metrics.csv"

processing:
  contextual_filtering:
    metrics:
      - field: "error_rate"
        condition: ">0.05"
        new_name: "high_error_metrics"
      - field: "response_time"
        condition: ">2000"
        new_name: "slow_response_metrics"

ai_interface:
  model:
    type: rest
    url: "https://api.anthropic.com/v1/messages"
    method: POST
    headers:
      x-api-key: "${ANTHROPIC_API_KEY}"
      Content-Type: "application/json"
    options:
      model: "claude-sonnet-4-20250514"

  prompts:
    alert:
      system: "You are a system monitoring analyst."
      user_template: |
        Generate an alert summary for these issues:

        High Error Rates:
        {% for m in high_error_metrics %}
        - {{ m.service }}: {{ m.error_rate }} error rate
        {% endfor %}

        Slow Responses:
        {% for m in slow_response_metrics %}
        - {{ m.endpoint }}: {{ m.response_time }}ms
        {% endfor %}

        Prioritize by severity and suggest remediation.

output:
  type: file
  path: "alerts.md"
  format: json

data_sources:
  revenue:
    type: rest
    url: "https://api.stripe.com/v1/balance"
    headers:
      Authorization: "Bearer ${STRIPE_SECRET_KEY}"

  users:
    type: csv
    path: "user_metrics.csv"

processing:
  content_summarization:
    revenue:
      method: "statistical"
      fields: ["available.amount"]
      summarize: ["sum"]

  content_summarization:
    users:
      method: "statistical"
      fields: ["active"]
      summarize: ["count", "sum"]

output:
  type: file
  path: "executive_dashboard.json"
  format: json