STATUS.yaml

# STATUS.yaml — What is real vs aspirational
# stats_refreshed: 2026-06-10 (UTC) — run: python scripts/refresh_repo_stats.py

real_and_working:
  compiler:
    file: compiler_v2.py
    lines: 7052
    status: production
    notes: "Parses .ainl → IR. Multiple emitters. Fully tested."

  runtime:
    file: runtime/engine.py
    lines: 3106
    status: production
    notes: "Executes IR graphs. Async support. Observability."

  cli:
    file: cli/main.py
    lines: 3766
    status: production
    commands:
      - "ainl run"
      - "ainl validate [--strict] [--json-output]"
      - "ainl emit --target <target>"
      - "ainl serve [--port PORT]"
      - "ainl compile"
      - "ainl check"
      - "ainl visualize"
      - "ainl inspect"
      - "ainl init"
      - "ainl doctor"

  adapters:
    status: production
    python_modules_under_adapters: 70
    notes: "Count = recursive *.py under adapters/ (excludes __pycache__). Older docs said '54 modules' — this is the live count."
    key_adapters:
      - "solana"
      - "postgres"
      - "dynamodb"
      - "redis"
      - "supabase"
      - "mysql"
      - "airtable"
      - "llm/openrouter"
      - "llm/ollama"
      - "llm/anthropic"
      - "llm/cohere"
      - "http, memory, cache, queue"

  tests:
    definitions:
      py_files_under_tests: "Every *.py under tests/ (includes conftest.py, helpers, harnesses)."
      test_named_py_files: "Files named test_*.py under tests/ (pytest test modules)."
      lines_python_under_tests: "Sum of lines of all *.py under tests/."
      pytest_collect_only: "pytest --collect-only -q (selected vs total reflects default deselection)."
    py_files_under_tests: 264
    test_named_py_files: 234
    lines_python_under_tests: 31962
    lines_python_under_tests_rounded: "~32k"
    pytest_collect_only_selected: 1636
    pytest_collect_only_total: 1675
    pytest_deselected: 39
    status: "Pass/fail counts vary by optional deps; run CI or pytest locally. Collection counts above are machine-generated."

  examples:
    count: 110
    format: ".ainl under examples/ (real, compilable syntax)"

  http_server:
    command: "ainl serve --port 8080"
    endpoints:
      - "POST /validate"
      - "POST /compile"
      - "POST /run"
      - "GET /health"
    status: working
    notes: "Zero-dependency HTTP server using stdlib. Foundation for cloud."

  compact_syntax_preprocessor:
    file: ainl_preprocess.py
    lines: 501
    status: production
    tests: 54
    notes: >
      Human-friendly compact syntax (Python-like) that transpiles to standard
      opcodes before the compiler sees it. Zero compiler changes. Both strict
      and non-strict modes supported. 66% token reduction vs raw opcodes.
    examples: "examples/compact/"


telemetry_and_audit_surfaces:
  map_doc: docs/operations/AUDIT_AND_TELEMETRY_MAP.md
  runner_structured_audit:
    status: shipped
    notes: "HTTP runner service (scripts/runtime_runner_service.py); ainl.runner JSON — docs/operations/AUDIT_LOGGING.md"
  cli_trajectory_jsonl:
    status: shipped
    notes: "RuntimeEngine per-step JSONL via ainl run --trace-jsonl or --log-trajectory / AINL_LOG_TRAJECTORY — docs/trajectory.md"
  audit_trail_adapter:
    status: shipped
    notes: "Opt-in adapter; AINL_AUDIT_SINK / --audit-sink — docs/tutorials/production_with_estimates_and_audit.md"
  runtime_observability_jsonl:
    status: shipped
    notes: "Optional metrics JSONL (AINL_OBSERVABILITY, AINL_OBSERVABILITY_JSONL) — runtime/observability.py"
aspirational_not_built:
  graph_block_syntax:
    description: "graph X { node Y: LLM(...) { ... } } verbose block syntax in tutorials"
    status: "design preview only — does NOT compile"
    location: "docs/learning/intermediate/"
    notes: >
      Some tutorials show a verbose graph{} block syntax alongside the working
      compact syntax. The graph{} blocks are aspirational and marked with
      DESIGN PREVIEW disclaimers. Use compact syntax instead.

  cloud_marketplace_api:
    description: "Token-gated marketplace for buying/selling .ainl graphs"
    status: "design only — zero implementation code"
    location: "ainativelangcloud repo"

  hosted_runtime_saas:
    description: "Managed execution environment for enterprises"
    status: "design only"

  validation_saas_dashboard:
    description: "Cloud-hosted validation with compliance reports"
    status: "design only — ainl serve exists locally but not deployed"

  token_gating:
    description: "1M $AINL wallet balance check for marketplace access"
    status: "not built — Solana adapter exists but gating logic doesn't"

  marketplace_graphs:
    description: ".ainl.json files in ainativelangcloud/marketplace/graphs/"
    status: "hand-written mockups, not compiler output"

  enterprise_support_sla:
    description: "Tiered support contracts"
    status: "spec written, not operational"

  community_champions_program:
    description: "Token-incentivized contributor rewards"
    status: "spec written, not launched"

  ainl_native_cloud:
    description: "Running cloud services as AINL graphs instead of Lambda/Python"
    status: "plan only — no AINL backend graphs written"

  formal_verification_z3:
    description: "Z3-based safety property verification for IR adapter calls"
    status: "PoC — opt-in via --verify-adapters; requires z3-solver"
    notes: "Proves adapter allowlist compliance, effect tiers, acyclicity on sample IR. Not claiming full operational semantics."
    location: "tooling/verification/z3_adapter_safety.py"

  wasm_subgraph_emit:
    description: "Compile deterministic IR subgraphs to WASM modules"
    status: "research spike planned (P5) — no implementation exists"
    notes: "Analysis tool (tooling/wasm_subgraph_analysis.py) may identify candidates. No ainl emit --target wasm until end-to-end test runs."

  core_native_acceleration:
    description: "Native accelerators for hot core.* ops"
    status: "research spike planned (P5) — no implementation exists"
    notes: "Benchmark-first approach. Separate from ainl_native_cloud."