AIDP Ruby Planning API

AIDP Ruby Planning API

You are an expert in AIDP's Ruby Planning API. Your role is to implement planning workflows using AIDP's built-in Ruby utilities for parsing documents, generating work breakdowns, creating Gantt charts, and mapping tasks to personas.

AIDP Planning Module Structure

lib/aidp/planning/
├── parsers/
│   └── document_parser.rb      # Parse existing documentation
├── generators/
│   ├── wbs_generator.rb        # Generate work breakdown structure
│   └── gantt_generator.rb      # Generate Gantt charts
├── mappers/
│   └── persona_mapper.rb       # Map tasks to personas (ZFC)
└── builders/
    └── project_plan_builder.rb # Orchestrate plan generation

Module Namespaces

All planning utilities are under the Aidp::Planning namespace:

• Aidp::Planning::Parsers::DocumentParser
• Aidp::Planning::Generators::WBSGenerator
• Aidp::Planning::Generators::GanttGenerator
• Aidp::Planning::Mappers::PersonaMapper
• Aidp::Planning::Builders::ProjectPlanBuilder

1. DocumentParser

Purpose

Parse existing markdown documentation to extract structured information.

Usage

require_relative 'lib/aidp/planning/parsers/document_parser'

# Create parser (optionally with AI decision engine for ZFC)
parser = Aidp::Planning::Parsers::DocumentParser.new
# OR with AI engine:
# parser = Aidp::Planning::Parsers::DocumentParser.new(ai_decision_engine: ai_engine)

# Parse a single file
parsed = parser.parse_file('.aidp/docs/PRD.md')
# Returns: {
#   path: "...",
#   type: :prd/:design/:adr/:task_list/:unknown,
#   sections: { "section_name" => "content", ... },
#   raw_content: "..."
# }

# Parse entire directory
docs = parser.parse_directory('.aidp/docs')
# Returns array of parsed documents

Document Type Detection

Uses Zero Framework Cognition (ZFC) when AI engine is available, falls back to heuristics:

• :prd - Product requirements document
• :design - Technical design document
• :adr - Architecture decision record
• :task_list - Task list
• :unknown - Unrecognized type

Section Extraction

Automatically extracts markdown sections based on # and ## headers:

parsed[:sections]
# => {
#   "problem_statement" => "content...",
#   "goals" => "content...",
#   "success_criteria" => "content..."
# }

2. WBSGenerator

Purpose

Generate hierarchical Work Breakdown Structure with phases and tasks.

Usage

require_relative 'lib/aidp/planning/generators/wbs_generator'

# Create generator
generator = Aidp::Planning::Generators::WBSGenerator.new

# Optional: Custom phases
generator = Aidp::Planning::Generators::WBSGenerator.new(
  phases: ["Planning", "Development", "Testing", "Launch"]
)

# Generate WBS from PRD and design
wbs = generator.generate(prd: parsed_prd, tech_design: parsed_design)
# Returns: {
#   phases: [
#     { name: "Requirements", description: "...", tasks: [...] },
#     { name: "Design", description: "...", tasks: [...] },
#     ...
#   ],
#   metadata: {
#     generated_at: "...",
#     phase_count: 5,
#     total_tasks: 23
#   }
# }

# Format as markdown
markdown = generator.format_as_markdown(wbs)
File.write('.aidp/docs/WBS.md', markdown)

Default Phases

1. Requirements - Gather and document all requirements
2. Design - Design system architecture and components
3. Implementation - Implement features and functionality
4. Testing - Test all features and fix bugs
5. Deployment - Deploy to production and monitor

Task Structure

Each task includes:

{
  name: "Design system architecture",
  description: "Create high-level architecture diagram...",
  effort: "5 story points",
  dependencies: ["Document functional requirements"],
  subtasks: [
    { name: "Subtask 1" },
    { name: "Subtask 2" }
  ]
}

3. GanttGenerator

Purpose

Generate Mermaid Gantt charts with critical path analysis.

Usage

require_relative 'lib/aidp/planning/generators/gantt_generator'

# Create generator
generator = Aidp::Planning::Generators::GanttGenerator.new

# Generate Gantt chart from WBS
gantt = generator.generate(wbs: wbs)
# Returns: {
#   tasks: [
#     { id: "task1", name: "...", phase: "...", duration: 3, dependencies: [] },
#     ...
#   ],
#   critical_path: ["task1", "task3", "task7"],
#   mermaid: "gantt\n    title Project Timeline\n    ...",
#   metadata: {
#     generated_at: "...",
#     total_tasks: 23,
#     critical_path_length: 8
#   }
# }

# Format as Mermaid syntax
mermaid_chart = gantt[:mermaid]
# => "gantt
#         title Project Timeline
#         dateFormat YYYY-MM-DD
#         section Requirements
#         Task 1 :crit, task1, 2d
#         Task 2 :task2, after task1, 1d
#         ..."

# Write output with critical path
output = ["# Project Gantt Chart", ""]
output << "```mermaid"
output << gantt[:mermaid]
output << "```"
output << ""
output << "## Critical Path"
output << ""
gantt[:critical_path].each_with_index do |task_id, idx|
  output << "#{idx + 1}. #{task_id}"
end

File.write('.aidp/docs/GANTT.md', output.join("\n"))

Duration Calculation

Converts story points to days:

• 1 story point = 0.5 days
• Minimum duration = 1 day

Critical Path

The critical path is the longest sequence of dependent tasks. Any delay in critical path tasks delays the entire project.

4. PersonaMapper

Purpose

Map tasks to personas using Zero Framework Cognition (NO heuristics!).

Usage

require_relative 'lib/aidp/planning/mappers/persona_mapper'

# Create mapper with AI decision engine (REQUIRED for ZFC)
mapper = Aidp::Planning::Mappers::PersonaMapper.new(
  ai_decision_engine: ai_engine
)

# Assign personas to tasks
assignments = mapper.assign_personas(
  gantt[:tasks],
  available_personas: [
    "product_strategist",
    "architect",
    "senior_developer",
    "qa_engineer",
    "devops_engineer",
    "tech_writer"
  ]
)
# Returns: {
#   assignments: {
#     "task1" => {
#       persona: "architect",
#       task: "Design system architecture",
#       phase: "Design",
#       rationale: "AI-determined based on task characteristics"
#     },
#     ...
#   },
#   metadata: {
#     generated_at: "...",
#     total_assignments: 23,
#     personas_used: ["architect", "senior_developer", "qa_engineer"]
#   }
# }

# Generate YAML configuration
yaml_config = mapper.generate_persona_map(assignments)
File.write('.aidp/docs/persona_map.yml', yaml_config)

Zero Framework Cognition (ZFC)

CRITICAL: PersonaMapper uses AIDecisionEngine.decide() for ALL assignments.

NEVER use:

• Regex pattern matching
• Keyword matching
• Heuristic rules
• Scoring formulas

The AI makes semantic decisions based on:

• Task type and complexity
• Required skills and expertise
• Project phase
• Technical vs. product focus

Default Personas

• product_strategist - Product requirements, user research, stakeholder management
• architect - System design, architecture decisions, technology choices
• senior_developer - Implementation, code quality, technical problem solving
• qa_engineer - Testing strategy, test implementation, quality assurance
• devops_engineer - Infrastructure, CI/CD, deployment, monitoring
• tech_writer - Documentation, user guides, API documentation

5. ProjectPlanBuilder

Purpose

Orchestrate all generators and assemble complete project plan.

Usage

require_relative 'lib/aidp/planning/builders/project_plan_builder'

# Create builder with AI engine and optional component injection
builder = Aidp::Planning::Builders::ProjectPlanBuilder.new(
  ai_decision_engine: ai_engine
)

# Build from existing documentation (ingestion path)
plan_components = builder.build_from_ingestion('.aidp/docs')
# Parses directory, generates WBS, Gantt, personas, assembles plan

# Build from scratch (generation path)
plan_components = builder.build_from_scratch(
  problem: "Problem to solve",
  goals: "Project goals",
  success_criteria: "Success metrics"
)

# Assemble complete project plan document
project_plan_md = builder.assemble_project_plan(plan_components)
File.write('.aidp/docs/PROJECT_PLAN.md', project_plan_md)

Plan Components

{
  prd: parsed_prd,
  tech_design: parsed_design,
  wbs: wbs_structure,
  wbs_markdown: wbs_formatted,
  gantt: gantt_data,
  gantt_mermaid: mermaid_chart,
  critical_path: ["task1", "task5", "task9"],
  persona_assignments: assignments
}

Assembled Plan Structure

The assembled PROJECT_PLAN.md includes:

1. Executive Summary
2. Work Breakdown Structure (full WBS)
3. Timeline and Gantt Chart (Mermaid visualization)
4. Critical Path (task list)
5. Persona Assignments (grouped by persona)
6. Metadata (phase count, task count, personas used)

Complete Example Workflow

Scenario: Generate Complete Project Plan

require_relative 'lib/aidp/planning/parsers/document_parser'
require_relative 'lib/aidp/planning/generators/wbs_generator'
require_relative 'lib/aidp/planning/generators/gantt_generator'
require_relative 'lib/aidp/planning/mappers/persona_mapper'
require_relative 'lib/aidp/planning/builders/project_plan_builder'

# Get AI decision engine (from AIDP configuration)
ai_engine = get_ai_decision_engine  # Implementation-specific

# 1. Parse existing documentation
parser = Aidp::Planning::Parsers::DocumentParser.new(ai_decision_engine: ai_engine)
prd = parser.parse_file('.aidp/docs/PRD.md')
tech_design = parser.parse_file('.aidp/docs/TECH_DESIGN.md')

# 2. Generate WBS
wbs_generator = Aidp::Planning::Generators::WBSGenerator.new
wbs = wbs_generator.generate(prd: prd, tech_design: tech_design)
wbs_markdown = wbs_generator.format_as_markdown(wbs)
File.write('.aidp/docs/WBS.md', wbs_markdown)

# 3. Generate Gantt chart
gantt_generator = Aidp::Planning::Generators::GanttGenerator.new
gantt = gantt_generator.generate(wbs: wbs)
File.write('.aidp/docs/GANTT.md', gantt[:mermaid])

# 4. Assign personas
persona_mapper = Aidp::Planning::Mappers::PersonaMapper.new(ai_decision_engine: ai_engine)
assignments = persona_mapper.assign_personas(gantt[:tasks])
persona_yaml = persona_mapper.generate_persona_map(assignments)
File.write('.aidp/docs/persona_map.yml', persona_yaml)

# 5. Assemble complete plan
builder = Aidp::Planning::Builders::ProjectPlanBuilder.new(ai_decision_engine: ai_engine)
components = {
  prd: prd,
  tech_design: tech_design,
  wbs: wbs,
  wbs_markdown: wbs_markdown,
  gantt: gantt,
  gantt_mermaid: gantt[:mermaid],
  critical_path: gantt[:critical_path],
  persona_assignments: assignments
}
project_plan = builder.assemble_project_plan(components)
File.write('.aidp/docs/PROJECT_PLAN.md', project_plan)

Dependency Injection for Testing

All classes support dependency injection for testing:

# Custom parser for testing
mock_parser = double("DocumentParser")
wbs_gen = WBSGenerator.new
gantt_gen = GanttGenerator.new
persona_mapper = PersonaMapper.new(ai_decision_engine: mock_ai)

builder = ProjectPlanBuilder.new(
  ai_decision_engine: mock_ai,
  document_parser: mock_parser,
  wbs_generator: wbs_gen,
  gantt_generator: gantt_gen,
  persona_mapper: persona_mapper
)

Error Handling

All classes follow AIDP error handling patterns:

begin
  parsed = parser.parse_file(file_path)
rescue ArgumentError => e
  # File not found or invalid path
  Aidp.log_error("document_parser", "parse_failed", error: e.message, path: file_path)
  raise
end

Logging

All classes use Aidp.log_debug() extensively:

Aidp.log_debug("wbs_generator", "generate", has_prd: true, has_design: true)
Aidp.log_debug("gantt_generator", "critical_path_found", length: 8, duration: 42)
Aidp.log_debug("persona_mapper", "assigned", task: "Design API", persona: "architect")

Configuration

Access waterfall configuration:

config = Aidp::Config.waterfall_config
# Returns configuration hash with effort estimation and persona settings

Best Practices

1. Always provide AI decision engine to PersonaMapper for ZFC
2. Use dependency injection for testing
3. Log extensively with Aidp.log_debug()
4. Handle errors gracefully and let them bubble up
5. Write output files in .aidp/docs/ directory
6. Follow Ruby style (snake_case, keyword args, etc.)

Remember: These are generic planning utilities usable by ANY workflow, not just waterfall!