goalroute

Deterministic DAG routing based on business goals, with qualitative-to-quantitative KPI conversion (economic and impact) captured from multiple PDF documents.

What it does

1. Ingest business intent (text or plain-text extracted from PDFs).
2. Convert qualitative language into quantitative KPIs using named, reproducible evaluation scales (economics + impact).
3. Route through a deterministic Directed Acyclic Graph (DAG) of decisions/ actions, driven by the business goals and their quantified KPIs.

Everything is deterministic: the same inputs always yield the same route.

Install

pip install goalroute          # core (stdlib only)
pip install goalroute[pdf]     # adds PDF extraction via pypdf

Quick start

from goalroute import GoalRouteEngine, Goal, DocumentSource

goals = [
    Goal(name="reduce_cost", weight=0.6),
    Goal(name="customer_delight", weight=0.4),
]

engine = GoalRouteEngine(goals)

# Capture intent from a PDF (needs pypdf), or pass text directly.
pdf_text = DocumentSource.from_pdf("proposal.pdf")   # str or None if no pypdf

docs = [pdf_text, "Reduce operational cost while improving satisfaction."]

report = engine.run(docs)
print(report.route)      # list of decision nodes
print(report.scores)     # quantified KPI scores per goal

Core concepts

• DocumentSource — reads text from PDFs and/or raw text.
• Goal — a weighted business goal.
• KpiConverter — maps qualitative phrases to quantitative KPIs (economic value and impact) using built-in deterministic scales.
• DecisionDag — a DAG of decision nodes with deterministic path selection.
• GoalRouteEngine — orchestrates capture → convert → route.
• RouteReport — the deterministic result.

Determinism

The routing is fully deterministic given the same inputs:

• KPI scales are closed-form lookup tables (no randomness).
• Ties are broken by node order in the DAG, never by RNG.
• No network calls or mutable global state.

Development

pip install -e ".[dev]"
pytest

License

MIT