feat: add model card packaging (#180)

* feat: add model card packaging

* fix: align model card metadata with plexe naming

* chore: bump version to 1.3.3

Author: marcellodebernardi

plexe/CODE_INDEX.md

@@ -1,6 +1,6 @@
 # Code Index: plexe
 
-> Generated on 2026-02-27 14:59:37
+> Generated on 2026-02-27 15:11:29
 
 Code structure and public interface documentation for the **plexe** package.
 
@@ -477,6 +477,13 @@ Standard XGBoost predictor - NO Plexe dependencies.
 - `__init__(self, model_dir: str)`
 - `predict(self, x: pd.DataFrame) -> pd.DataFrame` - Make predictions on input DataFrame.
 
+---
+## `templates/packaging/model_card_template.py`
+Model card template generator.
+
+**Functions:**
+- `generate_model_card(context, final_metrics: dict, evaluation_report: Any | None) -> str` - Generate a Markdown model card for the final package.
+
 ---
 ## `templates/training/train_catboost.py`
 Hardcoded robust CatBoost training loop.
@@ -700,4 +707,4 @@ Main workflow orchestrator.
 - `evaluate_final(spark: SparkSession, context: BuildContext, solution: Solution, config: Config, on_checkpoint_saved: Callable[[str, Path, Path], None] | None) -> dict` - Phase 5: Final evaluation on test set sample.
 - `package_final_model(spark: SparkSession, context: BuildContext, solution: Solution, final_metrics: dict, on_checkpoint_saved: Callable[[str, Path, Path], None] | None) -> Path` - Package all final deliverables into a unified directory.
 
----
+---