Feat/automation loop node

[ppatel9703]

Description

Adds the backend contract and execution lifecycle for bounded Automation loops. A loop owns a nested acyclic body graph, while Java controls iteration, limits, cancellation, routing, and run-history persistence. Existing node executors continue to run the work inside each iteration.

This PR also adds the data-extraction node used by file-processing workflows and hardens runtime scope-path handling for loop inputs.

Changes Made

• Added control.loop with three bounded modes:
  • For each — iterate over a list, optionally in batches.
  • Repeat — run a fixed number of times.
  • While — repeat while a validated condition remains true.
• Added loop-body validation for:
  • one executable entry path;
  • unique node IDs and output variables;
  • bounded iterations, batch sizes, and body size;
  • no nested loops or agent input waits in phase 1.
• Added per-iteration scope values, including the current item/batch, index, iteration number, and prior body outputs.
• Persisted loop and inner-node results so the UI can show each iteration and step independently in run history.
• Added data.extract for JSON/XML values with configurable paths and missing/null fallbacks.
• Preserved complete storage file paths for downstream loop processing.
• Hardened runtime placeholder parsing, including root bracket paths such as $["labName"] and keys containing dots.
• Corrected generated Automation source so the runtime module is imported exactly once after the package refactor.
• Added focused Java and Python regression coverage for catalog definitions, validation, rendering, execution, result persistence, and path resolution.

Deferred browser/Playwright and general data-transformation nodes are intentionally not included in the final scope.

How to Test

1. Rebuild and restart the backend.
2. Create an Automation with a list-producing node followed by a loop.
3. Configure the loop as For each, Repeat, or While, then add one or more ordinary nodes to its body.
4. Reference the loop item/index from a body node and run the Automation.
5. Confirm the loop is bounded and that each iteration records its inner-node results.
6. For extraction coverage, process JSON with a path such as labName or $["labName"] and verify the same value is returned.
7. Verify invalid loop bodies, excessive bounds, nested loops, and agent waits inside loops are rejected.

Verification

• Focused Python runtime suite: 7 tests passed.
• Live local file-processing Automation: passed, including 8 loop iterations and nested results.
• Root-bracket extraction ($["labName"]): passed after a backend rebuild.
• Java regression tests were added; Maven was not run locally.

Notes

• Phase 1 executes loop iterations sequentially and keeps the parent workflow graph acyclic.
• Nested loops and agent input waits inside loop bodies remain intentionally unsupported.
• The work stays within the Automation runtime, definition, execution, and scheduling integration areas.

[snyk-io]

✅ Snyk checks have passed. No issues have been found so far.

Status	Scan Engine	Critical	High	Medium	Low	Total (0)
✅	Open Source Security	0	0	0	0	0 issues

💻 Catch issues earlier using the plugins for VS Code, JetBrains IDEs, Visual Studio, and Eclipse.