NormalObjects - Strict Complaint Processor

This lab implements a strict LangGraph-style complaint workflow for the Downside Up Complaint Bureau.

Files

• normalobjects_langgraph.py - self-contained complaint processor, workflow engine fallback, and demo run.
• lab_summary.md - short LangGraph vs LangChain comparison for this use case.
• README.md - setup and execution notes.

How to run

Use the local Anaconda interpreter on this machine:

C:\Users\NunoAlmeida\anaconda3\python.exe normalobjects_langgraph.py

If langgraph is installed in your environment, the script will use it. If not, it falls back to a local compatible runner so the lab still executes.

What the script does

• Categorizes each complaint into exactly one of five categories.
• Enforces validation before investigation.
• Produces category-specific evidence, resolutions, and closure logs.
• Demonstrates both valid and invalid complaints.
• Prints a workflow path for every complaint.
• Leaves rejected and escalated complaints in a deferred state instead of pretending they were fully resolved.