Improve AI alternative model guidance

[codoom1]

What changed

• gives the chatbot the complete recommendation context: research question, goal, outcome, predictor types, sample size, missingness, distribution, expected relationship, correlation, and verified alternatives
• teaches the assistant to answer replacement-model questions with 3–6 concise alternatives
• requires each alternative to explain when it is preferable and its main assumption or tradeoff
• prioritizes engine-verified alternatives and labels any broader suggestion as conditional
• adds a signed-in “Compare alternatives with AI” action to recommendation results
• preserves bullet and paragraph formatting in chatbot responses

Why

The assistant previously saw only the recommended model name and research question. That was not enough context to explain credible replacements for a particular study design, and the generic 3–6 sentence response policy did not fit comparison questions.

User impact

From a recommendation result, a signed-in user can ask what should replace the leading model with one click. The response will resemble a statistical consultation: a bounded list of alternatives, the condition favoring each method, and the design or diagnostic evidence needed to choose.

The deterministic engine remains authoritative. AI advice does not alter the saved recommendation and cannot claim that assumptions or diagnostics were tested.

Validation

• python -m pytest -q tests/ — 130 passed
• focused integration and route tests — 45 passed
• node --check public/static/js/chatbot.js — passed
• fatal-error flake8 checks — passed
• git diff --check — passed

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
statistical-model-suggester	Ready	Preview, Comment	Jul 25, 2026 12:15am

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.