Serverlex ⚡

The Production-Safe Machine Learning Preprocessing, Diagnostics, & Serving Platform

🚀 Coming Soon to PyPI! Serverlex is currently undergoing final release preparations and will soon be installable with a simple pip install serverlex. Star this repository to get notified when the initial PyPI package is published!

Serverlex is an end-to-end Python library that takes raw, dirty data and transforms it into model-ready features, baseline benchmarks, tuned algorithms, and live production endpoints—all without data leakage or boilerplate code.

📦 Installation

Coming Soon via PyPI:

pip install serverlex

Current Installation (From GitHub / Source):

# Direct from GitHub:
pip install git+https://github.com/Uwais-ML/Serverlex.git

# Or install locally in editable mode:
git clone https://github.com/Uwais-ML/Serverlex.git
cd Serverlex
pip install -e .

🏗️ Architecture

flowchart TD
    subgraph 1. Ingestion & Quality
        A["Raw CSV / JSON"] --> B["DataCheck (DataGuard)"]
        B --> C["Clean (Sanitation Engine)"]
        C <--> D["Visualization (Live EDA)"]
    end

    subgraph 2. Preprocessing & Selection
        C --> E["MLPipeline (fit/transform)"]
        E --> F["FeatureSelector (Mutual Info)"]
    end

    subgraph 3. Modeling & Optimization
        F --> G["Baseline Evaluator (.add_model)"]
        G --> H["ModelHyperparameter (Nested CV)"]
        G --> I["EnsembleBuilder (Stacking/Voting)"]
    end

    subgraph 4. Diagnostics & Deployment
        H & I --> J["Evaluation (ModelEvaluator)"]
        H & I --> K["Explainability (ModelExplainer)"]
        E & H --> L["ModelDeployer (Joblib Bundle)"]
        L --> M["Live Raw Inference (1-line predict)"]
        L --> N["DriftDetector (Production Shift Alert)"]
    end

⚡ 60-Second Quickstart

import pandas as pd
from serverlex import (
    DataGuard,
    MLPipeline,
    ModelHyperparameter,
    ModelExplainer,
    ModelDeployer
)

# 1. Load Raw Messy Data
df = pd.read_csv("customers.csv")

# 2. Automated Quality Audit (catches ID columns, target leakage, class imbalance)
guard = DataGuard(df)
guard.run_audit(target="churn")

# 3. Fit Pipeline on Training Data (Zero Data Leakage)
train_df = df.iloc[:800].copy()
test_df = df.iloc[800:].copy()

pipeline = MLPipeline()
pipeline.fit(
    train_df,
    target="churn",
    encode="onehot",
    scale="standard",
    data_time=["signup_date", "%Y-%m-%d"],
    numeric_string_cols=["monthly_bill"]
)

X_train, y_train = pipeline.transform(train_df)
X_test, y_test = pipeline.transform(test_df)

# 4. Benchmark Baseline Models (+ add custom ones)
pipeline.evaluate_baseline_models(X_train, y_train, cv=5)

# 5. Tune Best Model with Unbiased Nested Cross-Validation
from sklearn.ensemble import RandomForestClassifier
tuner = ModelHyperparameter()
param_grid = tuner.get_default_param_grid("random_forest")

tune_res = tuner.tune_nested_cv(
    estimator=RandomForestClassifier(random_state=42),
    param_grid=param_grid,
    X=X_train,
    y=y_train,
    outer_cv=5,
    inner_cv=3
)
best_model = tune_res["best_estimator"]

# 6. Model Explainability
explainer = ModelExplainer(best_model, X_train.columns)
print(explainer.get_feature_importances().head(5))

# 7. Bundle & Deploy for Production Serving
deployer = ModelDeployer(pipeline=pipeline, model=best_model)
deployer.save("churn_deployer.joblib")

# Later in production: Predict directly on raw incoming JSON / dict!
loaded_deployer = ModelDeployer.load("churn_deployer.joblib")
raw_input = {"signup_date": "2026-04-01", "monthly_bill": "$89.50", "device": "laptop"}
print("Predicted Churn:", loaded_deployer.predict(raw_input)[0])

📦 Core Module Reference

🏆 Real-World Kaggle Benchmarks

Serverlex was verified on real-world Kaggle datasets without custom code changes:

• Customer Churn (muhammadshahidazeem/customer-churn-dataset):

  • DataGuard detected CustomerID as 100% unique ID and recommended removal.
  • Baseline Gradient Boosting achieved 99.42% accuracy.
  • ModelExplainer identified support_calls (0.277) and total_spend (0.235) as top drivers.

• Walmart Sales (mikhail1681/walmart-sales):

  • Extracted cyclical features from date strings (year, month, day, dayofweek).
  • Auto-detected regression and trained Random Forest Regressor to $R^2 = 0.9455$ ($\text{RMSE} = 133,727$).

🛠️ Installation

# Clone and install in editable development mode:
git clone https://github.com/your-username/Serverlex.git
cd Serverlex
pip install -e .

📖 In-Depth Manual

For mathematical concepts, parameter references, and output walkthroughs for every module, see DOCUMENTATION.md.