Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/acdeveloper-sci/fortriage/llms.txt

Use this file to discover all available pages before exploring further.

Decades-old Fortran systems still power critical scientific, engineering, and industrial workloads — but modernizing them is rarely straightforward. Static analysis tools can produce complexity reports, dependency graphs, dead-code findings, clone reports, and coupling metrics, generating hundreds or thousands of individual data points across a single codebase. ForTriage is a Streamlit-based interactive dashboard designed for software engineers, architects, and modernization teams who need to cut through that noise and immediately understand where to focus their effort first, why it matters, and how any given component affects the rest of the system.

​
The problem ForTriage solves

Large Fortran codebases do not suffer from a lack of analysis data — they suffer from a lack of prioritized, actionable insight. Teams receive collections of technical reports: 
complexity.csv
dead_code.csv
clone_report.md
coupling_report.csv
dependency reports
hundreds of metrics
The information exists, but answering the questions that drive decisions remains difficult:
  • Which modules carry the biggest modernization risk?
  • Where should refactoring begin?
  • Which components are migration blockers?
  • How do you explain technical findings to non-engineering stakeholders?
ForTriage bridges the gap between raw analysis data and concrete engineering decisions.

​
How ForTriage works

ForTriage is the visualization and decision layer built on top of Forti4D, an open-source Fortran static analysis engine. The two projects have clearly separated responsibilities: 
Legacy Fortran Source Code
          ↓
  Forti4D — Static Analysis
  (parsing, metrics, risk scoring, reports)
          ↓
  ForTriage — Dashboard
  (risk prioritization, interactive exploration, executive views)
          ↓
  AI Insights — Conversational AI
  (context-aware explanations over validated results)
Forti4D handles Fortran parsing, static analysis, metrics calculation, risk scoring, and technical report generation. ForTriage consumes those reports and transforms them into a prioritized, interactive dashboard — it never modifies source code or the analysis engine itself. The result is a clear risk tier view across your entire codebase, with five severity levels — CRITICAL, HIGH, MEDIUM, LOW, and DEAD_CODE — so teams can immediately identify the units that demand attention.

​
Dashboard views

ForTriage organizes its analysis into five focused views, each designed to answer a distinct set of questions about your codebase.

Overview

The top-level risk dashboard. See tier summary cards, a distribution chart across all risk levels, and a sortable, filterable table of every analyzed unit ranked by prioritization score.

Drill-down

Inspect individual units in depth. Surfaces complexity metrics, prioritization score breakdown, clone detection results, and reachability data for any unit in the corpus.

Executive Summary

A narrative view of the overall system health. Produces human-readable explanations of technical findings suitable for software architects, technical leads, and modernization stakeholders.

Report Explorer

Browse the full set of curated, corpus-level reports produced by the analysis pipeline — including migration strategy, clone pairs, and common-block coupling — in a tabular format.

AI Insights

A conversational AI layer over your validated analysis results. Ask questions like “Why is this module CRITICAL?” or “What should we refactor first?” and receive context-aware answers grounded in your actual data.

​
Scope and data boundaries

ForTriage is ephemeral and session-isolated. It never persists any data between runs and never modifies your Fortran source code. Every analysis session is fully independent — closing the app or starting a new analysis resets all state completely. The Overview, Drill-down, Executive Summary, and Report Explorer views operate entirely offline; only AI Insights communicates with an external service (Google’s Gemini API).

​
Try ForTriage now

A live demo is available at fortriage.streamlit.app — no installation required. Click Load sample fixtures to explore a real analysis of the built-in sample corpus (26 units, 8 files) and try every dashboard view, including AI Insights, directly in your browser.

Build docs developers (and LLMs) love

Get started for freeTalk to us