Codex Prompt & Context Architecture Workbench

1. System Directives & Signature Contract PYTHON

System Role & Framing Guardrails

You are an expert Python data engineer writing deterministic typing-annotated data pipelines using Polars and Pydantic v2.

Target Language

Model Context Window

Function / Interface Signature Contract

def transform_timeseries(raw_stream: Iterable[Dict[str, Any]], window_size: int = 15) -> pl.DataFrame:

2. Few-Shot Exemplars & Defenses 2 Exemplars

Defensive Constraint Toggles

✓ Explicit Type Annotations

✓ Zero In-Memory Buffer Overflows

✓ Disallow Undocumented Imports

+ Exhaustive Null/NaN Filtering

+ Include NumPy/Google Docstrings

Few-Shot In-Context Demonstrations

# Example 1: Input aggregation
# transform_timeseries([{"ts": 1690000000, "val": 42.1}], window_size=5) -> DataFrame with rolling stats

# Example 2: Schema guarantee
# Returns Polars LazyFrame schema with Float64 columns and strict timezone metadata

3. Live Context Budget & AST Analysis AST Syntax Valid

Prompt Tokens

428

Constraint Score

96 / 100

Context Headroom

3,668

■ System: 48t ■ Signature: 32t ■ Few-Shot: 94t ■ Guards: 46t

AST Boundary & Structural Node Verification

Type annotations explicitly pinned PASSED

Polars lazy-frame boundary defined PASSED

Zero runtime import hallucinations PASSED