Python Package Documentation Tile

Problem/Feature Description

Your data science team has built an internal Python library called smart_data_loader (version 2.4.0) published on your private PyPI registry. The library helps data scientists load, validate, and transform datasets from various sources. The AI assistant needs documentation about this package to help team members use it correctly.

You need to create a Tessl tile that documents this package. The tile should be in workspace "ds-team" named "ds-team/pypi-smart-data-loader", version 2.4.0, and should be a private tile.

Here is the library's public API extracted from its source:

Module: smart_data_loader

• SmartLoader(config: LoaderConfig) — Main class. Config takes cache_dir: str, max_retries: int = 3, timeout: float = 30.0
• SmartLoader.load_csv(path: str, schema: Schema | None = None) -> DataFrame — Load CSV with optional schema validation
• SmartLoader.load_parquet(path: str, columns: list[str] | None = None) -> DataFrame — Load Parquet with optional column selection
• SmartLoader.load_json(path: str, flatten: bool = False) -> DataFrame — Load JSON, optionally flatten nested structures
• SmartLoader.validate(df: DataFrame, schema: Schema) -> ValidationResult — Validate a DataFrame against a schema
• SmartLoader.transform(df: DataFrame, pipeline: list[Transform]) -> DataFrame — Apply transformation pipeline

Module: smart_data_loader.schemas

• Schema(fields: dict[str, FieldType]) — Define expected structure
• FieldType — Enum: STRING, INTEGER, FLOAT, BOOLEAN, DATETIME, JSON
• ValidationResult — Dataclass with is_valid: bool, errors: list[ValidationError]

Module: smart_data_loader.transforms

• Rename(mapping: dict[str, str]) — Rename columns
• Cast(column: str, dtype: FieldType) — Cast column type
• Filter(condition: str) — SQL-like filter expression
• Deduplicate(subset: list[str] | None = None) — Remove duplicate rows

Output Specification

Produce a complete tile directory at ./smart-data-loader-tile/ containing:

1. tile.json — properly configured manifest for a Python package
2. Documentation files with comprehensive API coverage
3. A file called setup-notes.md in the working directory explaining:
   • The CLI command you would run to scaffold this tile
   • Any manifest configuration decisions you made and why

Use realistic, brief content — enough to demonstrate the structure but not exhaustive prose.