adk-doc-audit

Documentation audit system for ADK-Rust that validates documentation against actual crate implementations.

Overview

The adk-doc-audit crate provides a comprehensive system for ensuring that ADK-Rust's documentation stays accurate and up-to-date with the actual crate implementations. It performs static analysis of documentation files, validates code examples, and cross-references API signatures with actual implementations.

Features

• API Reference Validation: Ensures all API references in documentation match actual implementations
• Code Example Compilation: Validates that code examples compile with current crate versions
• Version Consistency Checking: Verifies version references are current and consistent
• Cross-Reference Validation: Validates internal links and references
• Missing Feature Detection: Identifies features that exist in code but are missing from documentation
• Automated Suggestions: Provides fix suggestions for identified issues
• Comprehensive Reporting: Generates detailed audit reports in multiple formats (Console, JSON, Markdown)
• CI/CD Integration: Supports integration with build pipelines with appropriate exit codes
• Incremental Auditing: Process only changed files for efficient CI/CD workflows

Installation

Add to your Cargo.toml:

[dependencies]
adk-doc-audit = { git = "https://github.com/zavora-ai/adk-rust" }

Or use the main ADK-Rust crate with the doc-audit feature:

[dependencies]
adk-rust = { git = "https://github.com/zavora-ai/adk-rust", features = ["doc-audit"] }

Quick Start

Basic Usage

use adk_doc_audit::{AuditConfig, AuditOrchestrator, OutputFormat};

#[tokio::main]
async fn main() -> Result<(), Box<dyn std::error::Error>> {
    let config = AuditConfig::builder()
        .workspace_path(".")
        .docs_path("docs/official_docs")
        .output_format(OutputFormat::Console)
        .build();
    
    let orchestrator = AuditOrchestrator::new(config)?;
    let report = orchestrator.run_full_audit().await?;
    
    println!("Audit Results:");
    println!("  Total files: {}", report.summary.total_files);
    println!("  Files with issues: {}", report.summary.files_with_issues);
    println!("  Total issues: {}", report.summary.total_issues);
    println!("  Critical issues: {}", report.summary.critical_issues);
    println!("  Coverage: {:.1}%", report.summary.coverage_percentage);
    
    // Exit with error code if critical issues found
    if report.summary.critical_issues > 0 {
        std::process::exit(1);
    }
    
    Ok(())
}

Single File Validation

use adk_doc_audit::{AuditConfig, AuditOrchestrator};
use std::path::Path;

#[tokio::main]
async fn main() -> Result<(), Box<dyn std::error::Error>> {
    let config = AuditConfig::builder()
        .workspace_path(".")
        .docs_path("docs")
        .build();
    
    let orchestrator = AuditOrchestrator::new(config)?;
    let result = orchestrator.validate_file(Path::new("docs/getting-started.md")).await?;
    
    for issue in &result.issues {
        println!("{:?}: {}", issue.severity, issue.message);
        if let Some(suggestion) = &issue.suggestion {
            println!("  Suggestion: {}", suggestion);
        }
    }
    
    Ok(())
}

Incremental Audit

use adk_doc_audit::{AuditConfig, AuditOrchestrator};
use std::path::PathBuf;

#[tokio::main]
async fn main() -> Result<(), Box<dyn std::error::Error>> {
    let config = AuditConfig::builder()
        .workspace_path(".")
        .docs_path("docs")
        .build();
    
    let orchestrator = AuditOrchestrator::new(config)?;
    
    // Only audit changed files (useful for CI/CD)
    let changed_files = vec![
        PathBuf::from("docs/api-reference.md"),
        PathBuf::from("docs/examples.md"),
    ];
    
    let report = orchestrator.run_incremental_audit(&changed_files).await?;
    
    println!("Incremental audit complete: {} issues in {} files", 
             report.summary.total_issues, changed_files.len());
    
    Ok(())
}

CLI Usage

The crate provides a command-line interface for easy integration with development workflows:

Full Audit

Run a complete audit of all documentation:

# Basic audit with console output
adk-doc-audit audit --workspace . --docs docs/official_docs

# Generate JSON report for programmatic consumption
adk-doc-audit audit --workspace . --docs docs/official_docs --format json > audit-report.json

# Generate Markdown report for documentation
adk-doc-audit audit --workspace . --docs docs/official_docs --format markdown > audit-report.md

Incremental Audit

Audit only specific files (useful for CI/CD when you know which files changed):

adk-doc-audit incremental --workspace . --docs docs/official_docs --changed file1.md file2.md

Single File Validation

Validate a specific documentation file:

adk-doc-audit validate docs/official_docs/getting-started.md

Configuration File

Create a .adk-doc-audit.toml configuration file:

workspace_path = "."
docs_path = "docs/official_docs"
output_format = "console"

# Files to exclude from auditing
excluded_files = [
    "docs/internal/*",
    "docs/drafts/*"
]

# Crates to exclude from analysis
excluded_crates = [
    "example-crate",
    "test-utils"
]

# Severity threshold for failing builds
severity_threshold = "warning"
fail_on_critical = true

# Timeout for code example compilation
example_timeout = "30s"

Then run without arguments:

adk-doc-audit audit

Configuration

AuditConfig Builder

use adk_doc_audit::{AuditConfig, OutputFormat, IssueSeverity};
use std::time::Duration;

let config = AuditConfig::builder()
    .workspace_path(".")
    .docs_path("docs/official_docs")
    .output_format(OutputFormat::Json)
    .excluded_files(vec!["docs/internal/*".to_string()])
    .excluded_crates(vec!["test-utils".to_string()])
    .severity_threshold(IssueSeverity::Warning)
    .fail_on_critical(true)
    .example_timeout(Duration::from_secs(30))
    .build();

Configuration Options

Issue Types and Severity

Issue Categories

• ApiMismatch: API references that don't match actual implementations
• VersionInconsistency: Version numbers that don't match workspace versions
• CompilationError: Code examples that fail to compile
• BrokenLink: Internal links that point to non-existent files
• MissingDocumentation: Public APIs not documented
• DeprecatedApi: References to deprecated APIs

Severity Levels

• Critical: Issues that make documentation incorrect or unusable
• Warning: Issues that may confuse users but don't break functionality
• Info: Minor issues or suggestions for improvement

CI/CD Integration

GitHub Actions

name: Documentation Audit
on: [push, pull_request]

jobs:
  audit:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v3
      - uses: actions-rs/toolchain@v1
        with:
          toolchain: stable
      - name: Run documentation audit
        run: |
          cargo install --git https://github.com/zavora-ai/adk-rust adk-doc-audit
          adk-doc-audit audit --workspace . --docs docs --format json

Exit Codes

• 0: Audit passed (no critical issues)
• 1: Audit failed (critical issues found or configuration error)

Incremental CI/CD

For faster CI/CD, audit only changed files:

# Get changed files from git
CHANGED_FILES=$(git diff --name-only HEAD~1 HEAD -- '*.md' | tr '\n' ' ')

if [ -n "$CHANGED_FILES" ]; then
    adk-doc-audit incremental --workspace . --docs docs --changed $CHANGED_FILES
else
    echo "No documentation files changed"
fi

Integration with ADK-Rust Workflows

Spec-Driven Development

The audit system integrates with ADK-Rust's spec-driven development workflow:

1. Requirements Phase: Validate that all requirements are documented
2. Design Phase: Ensure design documents reference valid APIs
3. Implementation Phase: Verify code examples compile with new implementations
4. Testing Phase: Validate that property tests are documented

Hook Integration

Set up automatic audits when crates are updated:

use adk_doc_audit::{AuditOrchestrator, AuditConfig};

// Register hook for crate updates
let config = AuditConfig::builder()
    .workspace_path(".")
    .docs_path("docs")
    .build();

let orchestrator = AuditOrchestrator::new(config)?;

// This would be called by the hook system when a crate is updated
orchestrator.register_hook("crate_updated", |crate_name| async move {
    println!("Running audit after {} was updated", crate_name);
    // Run incremental audit on affected documentation
});

Report Formats

Console Output

Human-readable output with colored indicators:

=== Documentation Audit Report ===
📊 Summary:
  Total files: 15
  Files with issues: 3
  Total issues: 7
  Critical: 2, Warning: 3, Info: 2
  Coverage: 87.5%

❌ docs/api-reference.md:42
   [Critical] API reference 'GeminiModel::new_with_config' not found in crate
   💡 Suggestion: Use 'GeminiModel::with_config' instead

⚠️  docs/examples.md:15
   [Warning] Version reference '0.1.8' doesn't match workspace version '0.2.0'
   💡 Suggestion: Update to version '0.2.0'

JSON Output

Machine-readable format for integration:

{
  "summary": {
    "total_files": 15,
    "files_with_issues": 3,
    "total_issues": 7,
    "critical_issues": 2,
    "warning_issues": 3,
    "info_issues": 2,
    "coverage_percentage": 87.5
  },
  "issues": [
    {
      "file_path": "docs/api-reference.md",
      "line_number": 42,
      "severity": "Critical",
      "category": "ApiMismatch",
      "message": "API reference 'GeminiModel::new_with_config' not found in crate",
      "suggestion": "Use 'GeminiModel::with_config' instead"
    }
  ],
  "timestamp": "2024-01-15T10:30:00Z"
}

Markdown Output

Documentation-friendly format:

# Documentation Audit Report

Generated: 2024-01-15 10:30:00 UTC

## Summary

- **Total files**: 15
- **Files with issues**: 3  
- **Total issues**: 7
- **Critical issues**: 2
- **Coverage**: 87.5%

## Issues by File

### docs/api-reference.md

#### ❌ Line 42 - Critical
**API reference 'GeminiModel::new_with_config' not found in crate**
- Category: ApiMismatch
- Suggestion: Use 'GeminiModel::with_config' instead

Advanced Usage

Custom Validators

Extend the audit system with custom validation logic:

use adk_doc_audit::{Validator, ValidationResult, CodeExample};
use async_trait::async_trait;

struct CustomValidator;

#[async_trait]
impl Validator for CustomValidator {
    async fn validate_example(&self, example: &CodeExample) -> ValidationResult {
        // Custom validation logic
        ValidationResult {
            success: true,
            errors: vec![],
            warnings: vec![],
            suggestions: vec!["Consider adding error handling".to_string()],
        }
    }
}

// Register custom validator
let mut orchestrator = AuditOrchestrator::new(config)?;
orchestrator.add_validator(Box::new(CustomValidator));

Programmatic Report Processing

Process audit reports programmatically:

use adk_doc_audit::{AuditReport, IssueSeverity, IssueCategory};

fn process_report(report: &AuditReport) {
    // Filter critical API mismatches
    let critical_api_issues: Vec<_> = report.issues
        .iter()
        .filter(|issue| {
            matches!(issue.severity, IssueSeverity::Critical) &&
            matches!(issue.category, IssueCategory::ApiMismatch)
        })
        .collect();
    
    if !critical_api_issues.is_empty() {
        println!("Found {} critical API issues that need immediate attention:", 
                 critical_api_issues.len());
        
        for issue in critical_api_issues {
            println!("  - {}: {}", issue.file_path.display(), issue.message);
        }
    }
    
    // Generate metrics for tracking
    let metrics = serde_json::json!({
        "total_issues": report.summary.total_issues,
        "critical_issues": report.summary.critical_issues,
        "coverage": report.summary.coverage_percentage,
        "timestamp": report.timestamp
    });
    
    // Send to monitoring system, save to database, etc.
}

Troubleshooting

Common Issues

1. "Workspace not found": Ensure the workspace path points to a directory with Cargo.toml
2. "Documentation directory not found": Check that the docs path exists and contains markdown files
3. "Compilation timeout": Increase example_timeout for complex examples
4. "Permission denied": Ensure the process has read access to workspace and docs directories

Debug Mode

Enable debug logging for troubleshooting:

use tracing_subscriber;

// Initialize debug logging
tracing_subscriber::fmt()
    .with_max_level(tracing::Level::DEBUG)
    .init();

// Run audit with debug output
let report = orchestrator.run_full_audit().await?;

Performance Tuning

For large documentation sets:

let config = AuditConfig::builder()
    .workspace_path(".")
    .docs_path("docs")
    .example_timeout(Duration::from_secs(60))  // Increase timeout
    .excluded_files(vec![
        "docs/generated/*",  // Exclude auto-generated docs
        "docs/archive/*",    // Exclude archived content
    ])
    .build();

Contributing

Contributions are welcome! Please see the ADK-Rust contributing guide for details.

Adding New Validators

To add new validation capabilities:

1. Implement the Validator trait
2. Add configuration options if needed
3. Write property tests for the validator
4. Update documentation and examples

Extending Report Formats

To add new report formats:

1. Implement the ReportFormatter trait
2. Add the format to the OutputFormat enum
3. Update CLI argument parsing
4. Add tests and documentation

License

Licensed under the Apache License, Version 2.0. See LICENSE for details.