Add LangGraph ROI preset and integration example for Satya

Satya_AI already provides an ROI dashboard and audit logging for agent tasks, but users of popular agent frameworks like **LangGraph** currently have to design their own mapping from framework runs to Satya events. A small, opinionated "LangGraph ROI preset" + example project would make it much easier to plug Satya into existing LangGraph workflows.

Task
----

Create an integration example that shows how to:

1. Map **LangGraph node runs** to Satya tasks:
- node name → task name
- run status → task status
- timing / tokens → ROI fields (duration, cost if available)

2. Send logs to Satya:
- log intermediate events, errors, and final outputs as Satya audit events
- include minimal metadata (graph name, run id)

3. Visualize ROI in Satya:
- document how to filter by graph/agent name
- show which metrics become interesting (success rate, latency, etc.)

Implementation Steps
--------------------

- [ ] Create `examples/integrations/langgraph_satya/` directory
- [ ] Add a minimal LangGraph example that uses Satya's Python SDK
- [ ] Document the mapping decisions in a local `README.md`
- [ ] Add a docs section (or page) called **"LangGraph ROI preset"** and link it from the main `README.md`

Resources
---------

- • Satya SDK client and audit logging APIs in this repo
- • LangGraph docs / example patterns (external)

Why This Matters
----------------

A clear, copy-pasteable integration makes it easy for LangGraph users to adopt Satya for governance and ROI tracking, and positions Satya as a default observability layer for their agents.