# MCP Data Visualization Server Configuration (Flattened for Pydantic)
# Server Configuration
server__name: "data-viz-server"
server__version: "1.0.0"
server__description: "Natural Language Data Visualization MCP Server"
server__transport: "stdio"
server__log_level: "INFO"
# Database Configuration
database__type: "duckdb"
database__connection__path: "./data/mcp.duckdb"
database__connection__timeout: 30
database__connection__read_only: false
database__settings__memory_limit: "1GB"
database__settings__threads: 4
database__settings__enable_extensions: true
# LLM Configuration - Disabled (Use Claude Desktop instead)
llm__provider: "disabled"
llm__ollama__base_url: "http://localhost:11434"
llm__ollama__model: "qwen2:0.5b"
llm__ollama__timeout: 30
llm__ollama__max_tokens: 2048
llm__ollama__temperature: 0.1
llm__prompts__chart_type_detection_template: |
Analyze this visualization request and determine the best chart type.
Request: "{request}"
Available columns: {columns}
Chart types: bar, line, scatter, pie, histogram, box, heatmap, area
Respond with JSON: {{"chart_type": "type", "confidence": 0.8, "reasoning": "explanation"}}
llm__prompts__column_suggestion_template: |
For a {chart_type} chart with columns {columns}, suggest the most appropriate column mappings.
Respond with JSON: {{"suggestions": {{"x_axis": "col1", "y_axis": "col2", "explanation": "why"}}}}
llm__prompts__insights_description_template: |
Generate a clear, concise, and insightful description for the chart, highlighting key findings and observations. Use accessible language.
llm__prompts__chart_explanation_template: |
Explain what this chart is showing and how to interpret it, using clear, non-technical language.
llm__prompts__data_quality_check_template: |
Analyze this dataset for potential data quality issues, including missing values, outliers, inconsistencies, or incorrect data types. Suggest potential data cleaning and preparation if needed.
llm__prompts__followup_questions_template: |
Based on this {chart_type} chart, its context, and generated insights, propose 3-5 relevant and insightful follow-up questions a user might ask. Format them as a simple numbered list.
# Visualization Configuration
visualization__default_theme: "plotly_white"
visualization__output_format: "html"
visualization__figure_size__width: 800
visualization__figure_size__height: 600
visualization__color_schemes__categorical: ["#1f77b4", "#ff7f0e", "#2ca02c", "#d62728", "#9467bd"]
visualization__color_schemes__sequential: "Blues"
visualization__color_schemes__diverging: "RdBu"
visualization__chart_defaults__bar__orientation: "v"
visualization__chart_defaults__bar__text_auto: true
visualization__chart_defaults__bar__show_legend: true
visualization__chart_defaults__line__mode: "lines+markers"
visualization__chart_defaults__line__line_width: 2
visualization__chart_defaults__line__marker_size: 6
visualization__chart_defaults__scatter__mode: "markers"
visualization__chart_defaults__scatter__marker_size: 8
visualization__chart_defaults__scatter__opacity: 0.7
visualization__chart_defaults__pie__hole: 0
visualization__chart_defaults__pie__text_info: "label+percent"
visualization__chart_defaults__histogram__bins: 30
visualization__chart_defaults__histogram__opacity: 0.7
visualization__chart_defaults__histogram__show_distribution: true
# Insights Configuration
insights__enabled_types:
- "max"
- "min"
- "mean"
- "median"
- "distinct_count"
- "total_count"
- "correlation"
- "trend"
insights__correlation_threshold: 0.5
insights__trend_detection__min_points: 5
insights__trend_detection__significance_level: 0.05
insights__formatting__decimal_places: 2
insights__formatting__percentage_format: ".1%"
insights__formatting__large_number_format: ".2e"
# Data Processing Configuration
data__max_rows_preview: 1000
data__max_file_size_mb: 100
data__supported_formats: ["csv", "parquet", "json", "xlsx"]
data__preprocessing__auto_detect_types: true
data__preprocessing__handle_missing_values: true
data__preprocessing__date_formats: ["%Y-%m-%d", "%d/%m/%Y", "%m/%d/%Y", "%Y-%m-%d %H:%M:%S"]
data__sampling__large_dataset_threshold: 100000
data__sampling__sample_size: 10000
data__sampling__sampling_method: "random"
# Security Configuration
security__max_query_length: 10000
security__allowed_sql_keywords:
- "SELECT"
- "FROM"
- "WHERE"
- "GROUP BY"
- "ORDER BY"
- "LIMIT"
- "JOIN"
- "INNER JOIN"
- "LEFT JOIN"
- "RIGHT JOIN"
security__blocked_sql_keywords:
- "DROP"
- "DELETE"
- "INSERT"
- "UPDATE"
- "CREATE"
- "ALTER"
- "TRUNCATE"
security__file_access__allowed_paths: ["./data/", "./samples/"]
security__file_access__blocked_extensions: [".exe", ".bat", ".sh", ".py"]
# Development Configuration
development__debug_mode: true
development__auto_reload: true
development__sample_data__generate_on_startup: true
development__sample_data__datasets:
- name: "sales"
file: "./data/samples/sales_data.csv"
- name: "customers"
file: "./data/samples/customers.csv"
- name: "products"
file: "./data/samples/products.csv"
development__testing__use_test_database: true
development__testing__test_data_size: 100
