How inabubblewithyou/azure-documenter-tool works
Project Type and Purpose
• This is a CLI-based Python utility (“Azure Documenter Tool”) whose purpose is to inventory, analyze, and document Azure environments.
High-Level Architectural Components
a. CLI Orchestrator
• azure_documenter/main.py: entry point, parses CLI options (mode, subscriptions, flags), loads configuration, drives the workflow.
b. Configuration Layer
• azure_documenter/config.py: holds settings (authentication, output paths, LLM provider config).
c. Data-Access/Fetch Layer
• azure_documenter/fetchers/: modules that call Azure APIs (via DefaultAzureCredential) across categories (compute, network, storage, security, etc.).
• Each fetcher returns JSON-serializable data.
d. Analysis & Reporting Layer
• azure_documenter/generators/: orchestrates data consumption and builds outputs.
• analysis_helpers.py: data transformations, summaries, delta computations.
• delta_report_writer.py: builds “Compare” mode outputs.
• design_document_writer.py: high-level architecture and recommendations.
• markdown_writer.py & html_exporter.py: convert analysis into Markdown/HTML.
• llm_writer.py: optional AI-enhanced summaries (requires LLM provider).
e. Output Store
f. External Dependencies
• Azure Resource Manager APIs
• Optional LLM endpoints (Azure OpenAI / OpenAI)
• Python runtime & third-party libraries from requirements.txt
Relationships & Data Flow
1. User invokes main.py → CLI Parser
2. Config loaded → Authentication context created
3. Fetchers sequentially invoke Azure SDK → raw resource data collected
4. Raw JSON data saved to outputs/data/
5. Generators consume JSON:
• analysis_helpers perform aggregations or delta comparisons
• design_document_writer builds high-level docs
• delta_report_writer builds diff reports
• markdown_writer + html_exporter produce navigable reports in outputs/reports/
• llm_writer optionally inserts AI-generated insights
6. Logs written throughout to outputs/logs/
Architectural Patterns & Principles
• Modular Separation of Concerns:
• Fetch-only modules (data layer)
• Generator modules (presentation layer)
• Single orchestrator (main)
• Config isolated from code
• Pipeline Workflow: CLI → Fetch → Save → Generate → Export
• Extensibility Points:
• New fetchers for additional Azure services
• Plug in new writers (e.g., PDF exporter)
• Swap LLM provider via config
• Idempotent Outputs:
• Versioned JSON data files support Compare & Regenerate modes