Expand description
batuta has moved to aprender-orchestrate.
This crate re-exports aprender-orchestrate for backward compatibility.
New code should depend on aprender-orchestrate directly.
Modules§
- agent
- Autonomous Agent Runtime (perceive-reason-act loop).
- analyzer
- ansi_
colors - ANSI Color Support
- audit
- Pipeline Audit Trail Integration
- backend
- bug_
hunter - Proactive Bug Hunting Module
- comply
- Stack Compliance Engine - Cross-project consistency enforcement
- config
- content
- Content Creation Tooling Module
- data
- Data Platforms Integration
- experiment
- Experiment Tracking Integration Module
- falsification
- Popperian Falsification Checklist Implementation
- hf
- HuggingFace Hub Integration
- mcp
- MCP (Model Context Protocol) Server Module
- numpy_
converter - NumPy to Trueno conversion module (BATUTA-008)
- oracle
- Oracle Mode - Intelligent query interface for the Sovereign AI Stack
- parf
- PARF (Pattern and Reference Finder) module (BATUTA-012)
- pipeline
- 5-phase transpilation pipeline with Jidoka stop-on-error validation.
- pipeline_
analysis - Pipeline Library Analysis
- playbook
- Batuta Playbook — deterministic pipeline orchestration
- plugin
- pytorch_
converter - PyTorch to Realizar conversion module (BATUTA-010)
- recipes
- Orchestration Recipes Module
- report
- serve
- Model Serving Ecosystem
- sklearn_
converter - sklearn to Aprender conversion module (BATUTA-009)
- stack
- PAIML Stack Dependency Orchestration
- timing
- Centralized timing utilities.
- tools
- tui
- TUI (Terminal User Interface) Module
- types
- viz
- Data Visualization Frameworks Integration
Structs§
- Analysis
Stage - Analysis stage - detects languages and dependencies
- Anomaly
- Detected anomaly in the stack
- Aprender
Algorithm - Aprender equivalent algorithm
- Artifact
Signature - Cryptographic signature for artifacts
- Backend
Selector - Cost model for backend selection Based on spec section 2.2 lines 191-204
- Build
Stage - Build stage - compiles to final binary
- Capability
- A capability provided by a component
- Chain
Verification - Result of hash chain verification.
- Chat
Message - A chat message with role and content
- Chat
Template Engine - Chat template engine for formatting prompts
- Check
Item - A single checklist item result.
- Checklist
Result - Complete checklist evaluation result.
- CiCd
Benchmark Recipe - CI/CD integration recipe for cost-performance benchmarking
- Circuit
Breaker Config - Cost circuit breaker configuration
- Citation
Metadata - Citation metadata for BibTeX/CFF generation
- Comply
Config - Main compliance configuration
- Comply
Report - Stack compliance report
- Component
Metrics - Quality and performance metrics for a component
- Component
Node - A component in the stack knowledge graph
- Component
Quality - Quality assessment for a single component
- Component
Recommendation - Component recommendation with confidence
- Compute
Recommendation - Compute recommendation
- Context
Manager - Context manager for handling token limits
- Context
Snapshot - Snapshot of pipeline context for audit purposes.
- Context
Window - Known model context window sizes
- Cost
Circuit Breaker - Cost circuit breaker
- Cost
Metrics - Cost metrics for experiments
- Cost
Performance Benchmark - Cost-performance benchmark with Pareto frontier analysis
- Cost
Performance Benchmark Recipe - Cost-performance benchmarking recipe
- Cost
Performance Point - A point in the cost-performance space
- Dead
Code - Dead code analysis result
- Distribution
Recommendation - Distribution recommendation
- Energy
Metrics - Energy consumption metrics
- Error
Forecaster - Simple exponential smoothing for time series forecasting
- Evidence
- Evidence collected for a check.
- Experiment
Run - Experiment run with full tracking metadata
- Experiment
Tracking Config - Experiment tracking recipe configuration
- Experiment
Tracking Recipe - Experiment tracking recipe
- Failover
Config - Failover configuration
- Failover
Manager - Stateful failover manager
- File
Dependency - Dependency relationship between files
- Forecast
Metrics - Forecast error metrics
- Graph
Metrics - Computed graph-level metrics
- Hardware
Spec - Hardware specification for queries
- Hash
Chain Entry - Hash chain entry for tamper-evident audit trail.
- Health
Summary - Summary of stack health for dashboard
- Hero
Image Result - Hero image detection result
- InMemory
Experiment Storage - In-memory experiment storage for testing
- Integration
Pattern - Integration pattern between components
- Isolation
Forest - Isolation Forest for anomaly detection Implements a simplified version of the algorithm from Liu et al. (2008)
- Knowledge
Graph - Knowledge graph containing all stack components and their relationships
- Migration
Report - Migration report data
- NumPy
Converter - NumPy to Trueno converter
- Offline
Registry Config - Offline model registry configuration
- Optimization
Stage - Optimization stage - applies performance optimizations using MoE routing
- Oracle
Query - Oracle query structure
- Oracle
Response - Complete Oracle response
- Orcid
- ORCID identifier (Open Researcher and Contributor ID)
- Parf
Analyzer - PARF analyzer for cross-codebase analysis
- Pipeline
Audit Collector - Collector for pipeline audit trails.
- Pipeline
Path - Decision path representing a pipeline stage execution.
- Pipeline
Trace - A single audit trace entry for pipeline execution.
- Plugin
Metadata - Metadata describing a transpiler plugin
- Plugin
Registry - Plugin registry for managing transpiler plugins
- Plugin
Stage - Wrapper to integrate a plugin as a pipeline stage
- PreRegistration
- Pre-registration for reproducible research
- Project
Analysis - Complete project analysis results
- PyTorch
Converter - PyTorch to Realizar converter
- Quality
Checker - Quality matrix checker for PAIML stack components
- Quality
Issue - A quality issue found during analysis
- Quality
Summary - Summary statistics for stack quality
- Query
Constraints - Query constraints
- Query
Engine - Query engine that processes queries and generates responses
- Query
Preferences - Query preferences
- Realizar
Operation - Realizar equivalent operation
- Recipe
Result - Recipe execution result
- Recommender
- Oracle recommendation engine
- Research
Artifact - Research artifact with full academic metadata
- Research
Artifact Recipe - Academic research artifact recipe
- Research
Contributor - Research contributor with ORCID and CRediT roles
- Router
Config - Spillover routing configuration
- Score
- Score with context
- Serve
Backend Selector - Backend selection configuration
- Sklearn
Converter - sklearn to Aprender converter
- Sovereign
Artifact - Artifact in a sovereign distribution
- Sovereign
Deployment Config - Sovereign deployment recipe configuration
- Sovereign
Deployment Recipe - Sovereign deployment recipe
- Sovereign
Distribution - Sovereign distribution manifest for air-gapped deployments
- Spillover
Router - Spillover router for hybrid cloud routing
- Stack
Comply Engine - Stack Compliance Engine
- Stack
Component - A component in the Sovereign AI Stack
- Stack
Diagnostics - Main diagnostics engine for stack analysis
- Stack
Quality Report - Complete quality report for the stack
- Stage
Timer - Timer for measuring stage execution duration.
- Streaming
Context - Cached streaming context for failover recovery
- Symbol
Reference - A reference to a symbol in the codebase
- Token
Estimator - Simple token estimator (approximation without full tokenizer)
- Transpilation
Pipeline - Main transpilation pipeline
- Transpilation
Stage - Transpilation stage - converts source to Rust
- Validation
Stage - Validation stage - verifies semantic equivalence
- Workflow
State - Complete workflow state tracking
Enums§
- Andon
Status - Andon board status for the entire stack
- Anomaly
Category - Categories of anomalies
- Apple
Chip - Apple Silicon chip variants
- Artifact
Type - Types of distributable artifacts
- Backend
- Compute backend options
- Capability
Category - Category of capability
- Check
Status - Result status for a checklist item.
- Citation
Type - Citation type
- Code
Pattern - Code pattern detected in the codebase
- Comply
Report Format - Report output format
- Compute
Device - Compute device abstraction for heterogeneous hardware
- Compute
Intensity - Compute intensity levels
- CpuArchitecture
- CPU architecture variants
- Credit
Role - CRediT (Contributor Roles Taxonomy) roles
- Data
Size - Data size specification
- Dependency
Kind - Type of dependency between files
- Evidence
Type - Type of evidence collected.
- Experiment
Error - Errors that can occur during experiment tracking operations
- GpuVendor
- GPU vendor identification
- Health
Status - Health status for components (Andon-style visual control)
- Image
Format - Supported image formats for hero images
- Language
- Programming language detected in the project
- Model
Paradigm - Model training paradigm classification
- NumPyOp
- NumPy operation types
- OpComplexity
- Operation complexity for MoE (Mixture-of-Experts) routing
- Optimization
Target - Optimization target for queries
- Oracle
Backend - Compute backend selection
- Oracle
OpComplexity - Operation complexity classification
- Phase
Status - Status of a workflow phase
- Platform
Efficiency - Platform efficiency classification
- Privacy
Tier - Privacy tier for backend selection
- Problem
Domain - Problem domain for classification
- PyTorch
Operation - PyTorch operation types (inference-focused)
- Quality
Grade - Quality grade levels based on score percentages
- Quality
Stack Layer - Layer in the PAIML stack architecture
- Reject
Reason - Reason for rejection
- Report
Format - Role
- Chat message role
- Routing
Decision - Routing decision result
- RunStatus
- Run status
- Serving
Backend - Supported serving backends
- Severity
- Severity level for checklist items.
- Signature
Algorithm - Supported signature algorithms
- Sklearn
Algorithm - sklearn algorithm types
- Stack
Layer - Layer in the Sovereign AI Stack hierarchy
- Symbol
Kind - Symbol kind for reference finding
- Template
Format - Chat template format
- TpsGrade
- TPS-aligned assessment grade.
- TpuVersion
- TPU version variants
- Truncation
Strategy - Truncation strategy when context is exceeded
- Validation
Strategy - Validation strategy for pipeline stages
- Workflow
Phase - Workflow phase in the 5-phase Batuta pipeline
Traits§
- Experiment
Storage - Experiment storage backend trait
- Pipeline
Stage - Trait for pipeline stages
- Transpiler
Plugin - Core trait for transpiler plugins
Functions§
- evaluate_
critical_ only - Run only the critical architectural invariants.
- evaluate_
project - Run the complete falsification checklist against a project.
- new_
audit_ collector - Create a new pipeline audit collector with a generated run ID.
- record_
failure - Record a failed stage execution.
- record_
success - Record a successful stage execution.
- render_
dashboard - Render diagnostics as ASCII dashboard