Cleaning repo.

LICENSE

@@ -186,7 +186,7 @@
       same "printed page" as the copyright notice for easier
       identification within third-party archives.
 
-   Copyright [yyyy] [name of copyright owner]
+   Copyright 2025 Jeremy Vachier
 
    Licensed under the Apache License, Version 2.0 (the "License");
    you may not use this file except in compliance with the License.

Makefile

@@ -25,53 +25,53 @@ help:
 
 # Dependency management
 install:
-	@echo "📦 Installing dependencies with uv..."
+	@echo "Installing dependencies with uv..."
 	uv sync --all-extras
 
 # Code quality with Ruff
 format:
-	@echo "🎨 Formatting code with ruff..."
+	@echo "Formatting code with ruff..."
 	uv run ruff format src/ dash_app/ tests/ scripts/
 
 lint:
-	@echo "🔍 Linting code with ruff..."
+	@echo "Linting code with ruff..."
 	uv run ruff check . --fix
 	uv run ruff format --check .
 
 # Type checking
 typecheck:
-	@echo "🔎 Type checking with mypy..."
+	@echo "Type checking with mypy..."
 	uv run mypy src/ --ignore-missing-imports
 
 # Security checking
 security:
-	@echo "🔒 Security checking with bandit..."
+	@echo "Security checking with bandit..."
 	uv run bandit -r src/ -f json
 
 # Run all quality checks
 check-all: lint typecheck security
-	@echo "✅ All code quality checks completed!"
+	@echo "All code quality checks completed!"
 
 # Testing
 test:
-	@echo "🧪 Running tests..."
+	@echo "Running tests..."
 	uv run pytest tests/ -v
 
 # Pipeline execution
 run:
-	@echo "🚀 Running modular pipeline..."
+	@echo "Running modular pipeline..."
 	uv run python src/main_modular.py
 
 # Model training
 train-models:
-	@echo "🤖 Training and saving ML models..."
+	@echo "Training and saving ML models..."
 	uv run python scripts/train_and_save_models.py
 
 # Dash application
 dash:
-	@echo "📊 Starting Dash application..."
+	@echo "Starting Dash application..."
 	uv run python dash_app/main.py --model-name ensemble
 
 stop-dash:
-	@echo "🛑 Stopping Dash application..."
+	@echo "Stopping Dash application..."
 	@lsof -ti:8050 | xargs kill -9 2>/dev/null || echo "No process found on port 8050"

README.md

@@ -63,18 +63,18 @@ uv run python src/main_modular.py   # Run pipeline
 
 ```
 src/
-├── main_modular.py                 # 🎯 Main production pipeline (MLOps-enhanced)
-├── modules/                        # 🧩 Core modules
-│   ├── config.py                   # ⚙️ Configuration & logging
-│   ├── data_loader.py              # 📊 Data loading & external merge
-│   ├── preprocessing.py            # 🔧 Feature engineering
-│   ├── data_augmentation.py        # 🎲 Advanced synthetic data
-│   ├── model_builders.py           # 🏭 Model stack construction
-│   ├── ensemble.py                 # 🎯 Ensemble & OOF predictions
-│   ├── optimization.py             # 🔍 Optuna utilities
-│   └── utils.py                    # 🛠️ Utility functions
-
-dash_app/                           # 🖥️ Interactive Dashboard
+├── main_modular.py                 # Main production pipeline (MLOps-enhanced)
+├── modules/                        # Core modules
+│   ├── config.py                   # Configuration & logging
+│   ├── data_loader.py              # Data loading & external merge
+│   ├── preprocessing.py            # Feature engineering
+│   ├── data_augmentation.py        # Advanced synthetic data
+│   ├── model_builders.py           # Model stack construction
+│   ├── ensemble.py                 # Ensemble & OOF predictions
+│   ├── optimization.py             # Optuna utilities
+│   └── utils.py                    # Utility functions
+
+dash_app/                           # Interactive Dashboard
 ├── dashboard/                            # Application source
 │   ├── app.py                      # Main Dash application
 │   ├── layout.py                   # UI layout components
@@ -84,21 +84,21 @@ dash_app/                           # 🖥️ Interactive Dashboard
 ├── Dockerfile                      # Container configuration
 └── docker-compose.yml             # Multi-service orchestration
 
-models/                             # 🤖 Trained Models
+models/                             # Trained Models
 ├── ensemble_model.pkl              # Production ensemble model
 ├── ensemble_metadata.json         # Model metadata and labels
 ├── stack_*_model.pkl              # Individual stack models
 └── stack_*_metadata.json          # Stack-specific metadata
 
-scripts/                            # 🛠️ Utility Scripts
+scripts/                            # Utility Scripts
 └── train_and_save_models.py        # Model training and persistence
 
-data/                               # 📊 Datasets
+data/                               # Datasets
 
-docs/                               # 📝 Documentation
+docs/                               # Documentation
 └── [Generated documentation]       # Technical guides
 
-best_params/                        # 💾 Optimized parameters
+best_params/                        # Optimized parameters
 └── stack_*_best_params.json        # Per-stack best parameters
 ```
 
@@ -231,19 +231,19 @@ The pipeline employs six specialized ensemble stacks, each optimized for differe
 The pipeline is designed to achieve high accuracy through ensemble learning and advanced optimization techniques. Performance will vary based on:
 
 ```
-📊 Dataset Statistics
+Dataset Statistics
 ├── Training Samples: ~18,000+ (with augmentation)
 ├── Test Samples: ~6,000+
 ├── Original Features: 8 personality dimensions
 ├── Engineered Features: 14+ (with preprocessing)
 ├── Augmented Samples: Variable (adaptive, typically 5-10%)
 └── Class Balance: Extrovert/Introvert classification
 
-🔧 Technical Specifications
+Technical Specifications
 ├── Memory Usage: <4GB peak (configurable)
 ├── CPU Utilization: 4 cores (configurable)
-├── Model Persistence: ✅ Best parameters saved
-└── Reproducibility: ✅ Fixed random seeds
+├── Model Persistence: Yes - Best parameters saved
+└── Reproducibility: Yes - Fixed random seeds
 ```
 
 ## Testing & Validation

docs/README.md

@@ -40,7 +40,7 @@ docker build -t personality-classifier .
 docker run -p 8080:8080 personality-classifier
 ```
 
-## 📚 Resources
+## Resources
 
 - Code: `src/main_modular.py`, `examples/`
 - Config templates: [Configuration Guide](configuration.md)

docs/data-augmentation.md

@@ -397,14 +397,14 @@ def calculate_adaptive_ratio(data_characteristics):
 
 ### When to Use Augmentation
 
-✅ **Recommended**:
+**Recommended**:
 
 - Small to medium datasets (<10K samples)
 - Class imbalanced problems
 - High-stakes applications requiring robustness
 - When overfitting is detected
 
-❌ **Not Recommended**:
+**Not Recommended**:
 
 - Very large datasets (>100K samples)
 - When computational resources are limited

scripts/train_and_save_models.py

@@ -174,14 +174,14 @@ def main():
     setup_logging()
     logger = get_logger(__name__)
 
-    logger.info("🚀 Starting model training and saving process...")
+    logger.info("Starting model training and saving process...")
 
     # Create models directory
     models_dir = Path("models")
     models_dir.mkdir(exist_ok=True)
 
     # Load and prepare data
-    logger.info("📊 Loading and preparing data...")
+    logger.info("Loading and preparing data...")
     df_tr, df_te, submission = load_data_with_external_merge()
 
     # Preprocess data (prep function expects target column in df_tr)
@@ -215,7 +215,7 @@ def main():
     except Exception as e:
         logger.error(f"Failed to train ensemble model: {e}")
 
-    logger.info("✅ Model training and saving complete!")
+    logger.info("Model training and saving complete!")
     logger.info(f"Models saved in: {models_dir.absolute()}")
 
 

src/main_modular.py

@@ -88,7 +88,7 @@ def load_and_prepare_data(
     testing_mode: bool = True, test_size: int = 1000
 ) -> TrainingData:
     """Load and prepare training data."""
-    logger.info("🎯 Six-Stack Personality Classification Pipeline (Modular)")
+    logger.info("Six-Stack Personality Classification Pipeline (Modular)")
     logger.info("=" * 60)
 
     # Load data using advanced merge strategy
@@ -97,11 +97,11 @@ def load_and_prepare_data(
     # FOR TESTING: Limit to specified samples for faster execution
     if testing_mode and len(df_tr) > test_size:
         logger.info(
-            f"🔬 TESTING MODE: Limiting dataset to {test_size} samples "
+            f"TESTING MODE: Limiting dataset to {test_size} samples "
             f"(original: {len(df_tr)})"
         )
         df_tr = df_tr.sample(n=test_size, random_state=RND).reset_index(drop=True)
-        logger.info(f"   📊 Using {len(df_tr)} samples for testing")
+        logger.info(f"   Using {len(df_tr)} samples for testing")
 
     # Preprocess data with advanced competitive approach (do this first)
     X_full, X_test, y_full, le = prep(df_tr, df_te)
@@ -235,7 +235,7 @@ def train_single_stack(config: StackConfig, data: TrainingData) -> optuna.Study:
 
 def train_all_stacks(data: TrainingData) -> dict[str, optuna.Study]:
     """Train all stacks in the ensemble."""
-    logger.info("\n🔍 Training 6 specialized stacks...")
+    logger.info("\nTraining 6 specialized stacks...")
 
     stack_configs = get_stack_configurations()
     studies = {}
@@ -250,7 +250,7 @@ def create_model_builders(
     studies: dict[str, optuna.Study], data: TrainingData
 ) -> dict[str, Callable[[], Any]]:
     """Create model builder functions for each stack."""
-    logger.info("\n📊 Creating model builders for ensemble...")
+    logger.info("\nCreating model builders for ensemble...")
 
     builders = {
         "A": lambda: build_stack(studies["A"].best_trial, seed=RND, wide_hp=False),
@@ -274,7 +274,7 @@ def generate_oof_predictions(
     builders: dict[str, Callable[[], Any]], data: TrainingData
 ) -> dict[str, pd.Series]:
     """Generate out-of-fold predictions for all stacks."""
-    logger.info("\n🔮 Generating out-of-fold predictions...")
+    logger.info("\nGenerating out-of-fold predictions...")
 
     oof_predictions = {}
 
@@ -325,7 +325,7 @@ def optimize_ensemble_blending(
     oof_predictions: dict[str, pd.Series], y_full: pd.Series
 ) -> tuple[dict[str, float], float]:
     """Optimize ensemble blending weights."""
-    logger.info("\n⚖️ Optimizing ensemble blending...")
+    logger.info("\nOptimizing ensemble blending...")
 
     study_blend = optuna.create_study(direction="maximize")
     blend_objective = create_blend_objective(oof_predictions, y_full)
@@ -345,7 +345,7 @@ def optimize_ensemble_blending(
         "F": best_weights_list[5],
     }
 
-    logger.info("\n🏆 Best ensemble weights:")
+    logger.info("\nBest ensemble weights:")
     for stack_name, weight in best_weights.items():
         logger.info(f"   Stack {stack_name}: {weight:.3f}")
     logger.info(f"Best CV score: {study_blend.best_value:.6f}")
@@ -377,7 +377,7 @@ def refit_and_predict(
     models["F"].fit(data.X_full, y_full_noisy)
 
     # Generate final predictions
-    logger.info("\n🎯 Generating final predictions...")
+    logger.info("\nGenerating final predictions...")
     probabilities = {}
     for stack_name in ["A", "B", "C", "D", "E", "F"]:
         probabilities[stack_name] = models[stack_name].predict_proba(data.X_test)[:, 1]
@@ -408,11 +408,11 @@ def apply_pseudo_labelling(
 ) -> TrainingData:
     """Apply pseudo labelling using ensemble predictions."""
     if not ENABLE_PSEUDO_LABELLING:
-        logger.info("🔮 Pseudo labelling disabled")
+        logger.info("Pseudo labelling disabled")
         return data
 
     logger.info(
-        f"\n🔮 Applying pseudo labelling (threshold={PSEUDO_CONFIDENCE_THRESHOLD}, max_ratio={PSEUDO_MAX_RATIO})..."
+        f"\nApplying pseudo labelling (threshold={PSEUDO_CONFIDENCE_THRESHOLD}, max_ratio={PSEUDO_MAX_RATIO})..."
     )
 
     # First train models to get test predictions for pseudo labelling
@@ -464,9 +464,7 @@ def apply_pseudo_labelling(
 
     # Create new TrainingData with pseudo labels added
     if pseudo_stats["n_pseudo_added"] > 0:
-        logger.info(
-            f"✅ Pseudo labelling added {pseudo_stats['n_pseudo_added']} samples"
-        )
+        logger.info(f"Pseudo labelling added {pseudo_stats['n_pseudo_added']} samples")
 
         # Create new TrainingData object with enhanced training set
         enhanced_data = TrainingData(
@@ -478,14 +476,14 @@ def apply_pseudo_labelling(
         )
         return enhanced_data
     else:
-        logger.info("⚠️ No pseudo labels added, using original data")
+        logger.info("No pseudo labels added, using original data")
         return data
 
 
 def main():
     """Main execution function for the Six-Stack Personality Classification Pipeline."""
 
-    logger.info("🚀 Starting Six-Stack Personality Classification Pipeline")
+    logger.info("Starting Six-Stack Personality Classification Pipeline")
 
     try:
         # Load and prepare data
@@ -494,7 +492,7 @@ def main():
         )
 
         logger.info(
-            f"📊 Loaded data: {len(data.X_full)} training samples, {len(data.X_test)} test samples"
+            f"Loaded data: {len(data.X_full)} training samples, {len(data.X_test)} test samples"
         )
 
         # Train all stacks
@@ -503,7 +501,7 @@ def main():
         # Log stack optimization results
         for stack_name, study in studies.items():
             logger.info(
-                f"📈 Stack {stack_name}: Best score = {study.best_value:.6f} ({len(study.trials)} trials)"
+                f"Stack {stack_name}: Best score = {study.best_value:.6f} ({len(study.trials)} trials)"
             )
 
         # Create model builders
@@ -517,8 +515,8 @@ def main():
             oof_predictions, data.y_full
         )
 
-        logger.info(f"🎯 Best ensemble CV score: {best_cv_score:.6f}")
-        logger.info(f"⚖️ Ensemble weights: {best_weights}")
+        logger.info(f"Best ensemble CV score: {best_cv_score:.6f}")
+        logger.info(f"Ensemble weights: {best_weights}")
 
         # Apply pseudo labelling using ensemble predictions
         enhanced_data = apply_pseudo_labelling(builders, best_weights, data)
@@ -534,12 +532,12 @@ def main():
         )
 
         # Print final results
-        logger.info(f"\n✅ Predictions saved to '{output_file}'")
-        logger.info(f"📊 Final submission shape: {submission_df.shape}")
-        logger.info("🎉 Six-stack ensemble pipeline completed successfully!")
+        logger.info(f"\nPredictions saved to '{output_file}'")
+        logger.info(f"Final submission shape: {submission_df.shape}")
+        logger.info("Six-stack ensemble pipeline completed successfully!")
 
         # Print summary
-        logger.info("\n📋 Summary:")
+        logger.info("\nSummary:")
         logger.info(f"   - Training samples: {len(enhanced_data.X_full):,}")
         logger.info(f"   - Test samples: {len(enhanced_data.X_test):,}")
         logger.info(f"   - Features: {enhanced_data.X_full.shape[1]}")
@@ -551,7 +549,7 @@ def main():
         logger.info("   - Modular architecture")
 
     except Exception as e:
-        logger.error(f"❌ Pipeline failed: {e}")
+        logger.error(f"Pipeline failed: {e}")
         raise