Add project boilerplate and standard configuration

[Devanik21]

Added standard project boilerplate files including security policies, contributing guidelines, code of conduct, GitHub issue and PR templates, CI workflows, and Dockerization. Extracted and tested application source files from zip without modifying original files. Fixed a hardcoded local path in app.py based on code review feedback.

---

PR created automatically by Jules for task 11618927829301706655 started by @Devanik21

[google-labs-jules]

👋 Jules, reporting for duty! I'm here to lend a hand with this pull request.

When you start a review, I'll add a 👀 emoji to each comment to let you know I've read it. I'll focus on feedback directed at me and will do my best to stay out of conversations between you and other bots or reviewers to keep the noise down.

I'll push a commit with your requested changes shortly after. Please note there might be a delay between these steps, but rest assured I'm on the job!

For more direct control, you can switch me to Reactive Mode. When this mode is on, I will only act on comments where you specifically mention me with @jules. You can find this option in the Pull Request section of your global Jules UI settings. You can always switch back!

New to Jules? Learn more at jules.google/docs.

---

For security, I will only act on instructions from the user who triggered this task.

[gemini-code-assist]

Code Review

This pull request introduces AntiEntropy, a platform for biological aging analysis and epigenetic age reversal simulation. The implementation includes a Streamlit-based research interface and several specialized engines for biological clock prediction, epigenetic entropy calculation, and longevity modeling. The review feedback identifies several areas for improvement, including the need to reset application state upon new data uploads, ensuring consistency between the biological clock model and its performance evaluation, and optimizing computational efficiency by leveraging vectorized operations and specialized PCA implementations.

[gemini-code-assist]

The session state initialization does not account for changes in the uploaded file. If a user uploads a new dataset, the application will retain the fitted models and results from the previous dataset because st.session_state.pipeline_done remains True. This leads to a critical mismatch between the displayed data and the analysis results. The state should be reset whenever a new file is detected.

Suggested change

	if 'clock' not in st.session_state:
	st.session_state.clock = None
	st.session_state.entropy_eng = None
	st.session_state.reversal_sim = None
	st.session_state.hrf = None
	st.session_state.immortality = None
	st.session_state.age_accel_df = None
	st.session_state.pipeline_done = False
	if 'clock' not in st.session_state or st.session_state.get('current_file') != uploaded.name:
	st.session_state.current_file = uploaded.name
	st.session_state.clock = None
	st.session_state.entropy_eng = None
	st.session_state.reversal_sim = None
	st.session_state.hrf = None
	st.session_state.immortality = None
	st.session_state.age_accel_df = None
	st.session_state.pipeline_done = False

[gemini-code-assist]

The cross-validation score is being calculated using a Ridge regression model, which is inconsistent with the ElasticNetCV model used for the actual biological clock. This provides a misleading estimate of the clock's performance (MAE and R²). The CV evaluation should use the same model type and hyperparameters as the final fitted model.

        # 5-fold CV MAE using the best parameters from ElasticNetCV
        from sklearn.linear_model import ElasticNet
        best_model = ElasticNet(
            alpha=self.model.alpha_,
            l1_ratio=self.model.l1_ratio_,
            max_iter=2000,
            random_state=42
        )
        cv_scores = cross_val_score(
            best_model, X_scaled, y_arr,
            cv=5, scoring='neg_mean_absolute_error'
        )

[Devanik21]

The cross-validation score is being calculated using a Ridge regression model, which is inconsistent with the ElasticNetCV model used for the actual biological clock. This provides a misleading estimate of the clock's performance (MAE and R²). The CV evaluation should use the same model type and hyperparameters as the final fitted model.

    # 5-fold CV MAE using the best parameters from ElasticNetCV
    from sklearn.linear_model import ElasticNet
    best_model = ElasticNet(
        alpha=self.model.alpha_,
        l1_ratio=self.model.l1_ratio_,
        max_iter=2000,
        random_state=42
    )
    cv_scores = cross_val_score(
        best_model, X_scaled, y_arr,
        cv=5, scoring='neg_mean_absolute_error'
    )

yea, i see. plz fix.

[google-labs-jules]

I've fixed the cross-validation scoring to use ElasticNet with the best hyperparameters from the ElasticNetCV model as requested.

[gemini-code-assist]

Performing a full SVD using np.linalg.svd on high-dimensional methylation data is computationally expensive and memory-intensive. Since scikit-learn is already a dependency, it is more efficient to use sklearn.decomposition.PCA with the randomized solver, which is optimized for high-dimensional datasets.

Suggested change

	# SVD for PCA
	n_comp = min(self.n_components, X.shape[0] - 1, X.shape[1])
	U, S, Vt = np.linalg.svd(X_centered, full_matrices=False)
	self.pca_components = Vt[:n_comp]
	return (X_centered @ self.pca_components.T)
	# Randomized PCA for efficient dimensionality reduction
	from sklearn.decomposition import PCA
	n_comp = min(self.n_components, X.shape[0] - 1, X.shape[1])
	pca = PCA(n_components=n_comp, svd_solver='randomized', random_state=42)
	X_reduced = pca.fit_transform(X_centered)
	self.pca_components = pca.components_
	return X_reduced

[gemini-code-assist]

The use of np.apply_along_axis here is unnecessary and inefficient. The _binary_entropy function is already fully vectorized using NumPy operations. Calling it directly on the entire X_arr matrix will be significantly faster.

Suggested change

	sample_h = np.apply_along_axis(_binary_entropy, 1, X_arr) # (n_samples, n_cpgs)
	sample_h = _binary_entropy(X_arr) # (n_samples, n_cpgs)

[Devanik21]

@gemini-code-assist plz review , and add commit suggestions to all issues in this PR.

[gemini-code-assist]

Hi @Devanik21, thanks for reaching out! I can provide review comments and code suggestions to help you improve the pull request. However, I am unable to directly commit changes to your branch. You can apply the suggested changes yourself if you find them helpful.

Overall, this is a great set of additions for project boilerplate and configuration! The inclusion of issue templates, a PR template, code of conduct, contributing guidelines, security policy, Dockerization, and basic tests significantly improves the project's maintainability and collaboration readiness.

Here are a few observations and suggestions:

Makefile - Development Dependencies (Low Severity)

It's a good practice to separate development dependencies (like pytest and flake8) from production dependencies. This helps keep the production requirements.txt lean and clear. You could consider creating a requirements-dev.txt file for these.

docker-compose.yml - Redundant Environment Variables (Low Severity)

The STREAMLIT_SERVER_PORT and STREAMLIT_SERVER_ADDRESS environment variables in docker-compose.yml are redundant because the CMD in your Dockerfile already specifies these parameters for Streamlit. While it doesn't cause an error, removing them would make the configuration cleaner. Here's how the relevant section could look:

version: '3.8'

services:
  antientropy:
    build: .
    ports:
      - "8501:8501"
    volumes:
      - .:/app

CloCk.py - Magic Numbers (Medium Severity)

In CloCk.py, the ElasticNet constructor uses max_iter=2000 and random_state=42. While random_state is often used for reproducibility in machine learning, max_iter is a hyperparameter. Consider defining these as constants at the top of the file or making them configurable if their values are significant and might change. This improves maintainability and makes the code's intent clearer.