ShadowCompile: Grounding Developer Build Latency in Solar Mechanics with Open AI

Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass Submission 🌿

In software engineering, compilation latency is an ergonomic trap.

Whether running a full C++/Rust rebuild, executing comprehensive end-to-end integration suites, or training machine learning epochs, developers frequently encounter 5 to 20-minute waiting windows. Conventional developer behavior during these intervals is counterproductive: we switch tabs to social media or monitor terminal logs, compounding eye fatigue and cognitive exhaustion.

The prompt for Week 1 demands that we "make the screen the shortest part of the experience."

ShadowCompile solves this by converting developer build delays into deterministic, real-world solar calibration breaks.

ShadowCompile is a developer wellness and break synchronizer built on astronomical physics and open-weight AI.

Closed proprietary APIs (like commercial OpenAI or Anthropic endpoints) are structurally unsuited for this application:

Calculating exact solar angles requires geographic coordinates. Transmitting a developer’s real-time latitude and longitude alongside timestamped development schedules to centralized cloud servers presents an unacceptable privacy footprint. Using open-weight models ensures user context stays strictly on the edge.

Developer tools must be frictionless. Gating a local compilation hook behind pay-per-token API tiers or rate limits prevents broad adoption. Open weights eliminate runtime billing constraints entirely.

Local inference operates independently of external network connectivity. A developer working in transit, in an offline lab environment, or in low-connectivity areas retains a fully functional tool without network degradation.

The system avoids heavy external dependencies, relying on a lightweight, modular Python architecture:

Solar declination and hour angles are computed directly via spherical astronomical approximations:

$$\delta = 23.45^\circ \cdot \sin\left(\frac{360}{365} \cdot (d - 81)\right)$$

$$\sin(\theta) = \sin(\phi)\sin(\delta) + \cos(\phi)\cos(\delta)\cos(H)$$

This calculates the exact solar elevation ($\theta$) and surface shadow length ratio ($1 / \tan \theta$) without requiring third-party celestial datasets.

The application interfaces with Google's Gemma-2-2B-it open-weight model to translate raw numeric telemetry (elevation angles and shadow multipliers) into concise, grounded outdoor sensory tasks: