encoding_model: cl100k_base
skip_workflows: []
llm:
  api_key: ${GRAPHRAG_API_KEY}
  type: openai_chat  
  model: gpt-4o-mini
  model_supports_json: true
  max_tokens: 4000
  temperature: 0

embeddings:
  async_mode: threaded
  batch_size: 16
  vector_store:
    type: lancedb
    db_uri: 'output/lancedb'
    container_name: default
    overwrite: true
  llm:
    api_key: ${GRAPHRAG_API_KEY}
    type: openai_embedding
    model: text-embedding-3-small

chunks:
  size: 500
  overlap: 50
  group_by_columns: [id]

input:
  type: file
  file_type: text
  base_dir: "input"
  file_pattern: ".*\\.txt$"
  recursive: true
  source_tracking: true  
  processing_order:      
    - path: "first_paragraphs"
      priority: 1
      purpose: "graph_building"
    - path: "full_documents"
      priority: 2
      purpose: "retrieval"

entity_extraction:
  prompt: "prompts/entity_extraction.txt"
  entity_types:
    - "Baggage Type"
    - "Dimension"
    - "Linear Dimension"
    - "Weight"
    - "Material Type"
    - "Wheel Configuration"
    - "Measurement Unit"
    - "Size Category"
    - "Weight Category"
    - "Airline"
    - "Alliance"
    - "Airport"
    - "Route Type"
    - "Travel Class"
    - "Cabin Section"
    - "Aircraft Type"
    - "Restriction"
    - "Exemption"
    - "Policy"
    - "Fee Structure"
    - "Currency"
    - "Allowance"
    - "Special Item"
    - "Prohibited Item"
    - "Restricted Item"
    - "Dangerous Good"
    - "Fragile Item"
    - "Valuable Item"
    - "Required Document"
    - "Label Type"
    - "Tag Category"
    - "Service Type"
    - "Handler Role"
    - "Service Location"
    - "Time Period"
    - "Passenger Type"
    - "Membership Level"
    - "Group Category"
  max_gleanings: 2
  source_filter: "first_paragraphs"

claim_extraction:
  enabled: true
  claim_types:
    - "Basic Size Restriction"
    - "Oversize Condition"
    - "Weight Limit Standard"
    - "Overweight Condition"
    - "Combined Dimension Limit"
    - "Cabin Storage Requirement"
    - "Standard Fee"
    - "Excess Fee"
    - "Oversize Fee"
    - "Overweight Fee"
    - "Special Handling Fee"
    - "Season Surcharge"
    - "Route-Specific Fee"
    - "Multi-Piece Pricing"
    - "Fee Waiver Condition"
    - "Basic Allowance"
    - "Class-Based Allowance"
    - "Status-Based Allowance"
    - "Route-Based Allowance"
    - "Special Group Allowance"
    - "Seasonal Allowance"
    - "Equipment Allowance"
    - "Prohibited Item Policy"
    - "Restricted Item Condition"
    - "Dangerous Goods Policy"
    - "Special Item Restriction"
    - "Packaging Requirement"
    - "Declaration Requirement"
    - "Check-in Deadline"
    - "Special Handling Procedure"
    - "Priority Handling Rule"
    - "Transfer Handling Policy"
    - "Delivery Service Policy"
    - "Storage Policy"
    - "Liability Limit"
    - "Insurance Requirement"
    - "Claim Procedure"
    - "Compensation Policy"
    - "Time Limit Policy"
    - "Weather Restriction"
    - "Seasonal Restriction"
    - "Aircraft Limitation"
    - "Route Restriction"
    - "Connection Impact"
    - "Tag Requirement"
    - "Label Requirement"
    - "Documentation Requirement"
    - "Declaration Policy"
    - "Handling Standard"
    - "Service Level Agreement"
    - "Priority Service Standard"
    - "Delivery Time Standard"
    - "Medical Exception"
    - "Military Exception"
    - "Diplomatic Exception"
    - "Event Exception"
    - "Emergency Exception"
  prompt: "prompts/claim_extraction.txt"
  description: "Extract baggage measurements, weight limits, and restrictions from airline documentation."
  max_gleanings: 2
  source_filter: "first_paragraphs"

local_search:
  text_unit_prop: 0.7
  community_prop: 0.3
  top_k_mapped_entities: 15
  top_k_relationships: 15
  max_tokens: 4000
  source_priority:
    graph_search: "first_paragraphs"
    answer_retrieval: "full_documents"

global_search:
  max_tokens: 4000
  data_max_tokens: 4000
  map_max_tokens: 1000
  reduce_max_tokens: 2000
  allow_general_knowledge: false
  min_score_threshold: 0.1
  concurrency: 10

embed_graph:
  enabled: true
  num_walks: 100
  walk_length: 10
  window_size: 5
  iterations: 10

umap:
  enabled: true
  n_neighbors: 15
  min_dist: 0.1
  n_components: 2

storage:
  type: file
  base_dir: "output"

cache:
  type: file
  base_dir: "cache"

reporting:
  type: file
  base_dir: "reports"
  include_source_tracking: true