Cookbooks¶

1. Basic KPI deck from CSV¶

Use when you need a fast deck with a few headline metrics.

Config pattern¶

provider:
  type: "google_slides"
  config:
    template_id: "<template-id>"

presentation:
  name: "Weekly KPI Report"
  slides:
    - id: "<slide-id-1>"
      replacements:
        - type: "text"
          config:
            placeholder: "{{WEEK}}"
            replacement: "Week 07"
        - type: "text"
          config:
            placeholder: "{{TOTAL_REVENUE}}"
            data_source:
              type: "csv"
              name: "kpi"
              file_path: "./data/kpi.csv"
            value_fn: "get_first_value"
            value_fn_args:
              column: "revenue"

Command¶

slideflow validate config.yml --registry registry.py
slideflow build config.yml --registry registry.py

2. Reusable chart templates with built-ins¶

Use when your team needs consistent visuals across many decks.

Pattern¶

Define reusable YAML chart templates in templates/.
Use type: template charts with template_config values.
Keep styling in templates, business logic in config/data.

presentation:
  name: "Template Driven Deck"
  slides:
    - id: "<slide-id>"
      charts:
        - type: "template"
          config:
            template_name: "bars/bar_basic"
            data_source:
              type: "csv"
              name: "sales"
              file_path: "./data/sales.csv"
            template_config:
              title: "Revenue by Month"
              x_column: "month"
              y_column: "revenue"

3. Fully custom charts + bulk generation¶

Use when each audience segment needs a tailored deck.

Pattern¶

Create custom chart and formatter functions in registry.py
Use batch params CSV for segment variants
Run with controlled concurrency and rate limit

registry:
  - "./registry.py"

presentation:
  name: "Q{quarter} {region} Review"
  slides:
    - id: "<slide-id>"
      charts:
        - type: "custom"
          config:
            chart_fn: "build_segment_chart"
            data_source:
              type: "databricks"
              name: "segment_sales"
              query: "SELECT * FROM mart.sales WHERE region = '{region}'"
            chart_config:
              quarter: "{quarter}"

variants.csv:

quarter,region
Q1,NA
Q1,EMEA
Q1,APAC

Command:

slideflow build config.yml --params-path variants.csv --threads 2 --rps 0.8

Practical tips¶

Use --dry-run before high-volume runs.
Keep one shared base registry and add team-specific registries as needed.
Start with CSV-based fixtures for deterministic testing before moving to live connectors.