Cleanup and some refinements before finalizing the README and other documents to make them public

.dvc/config

@@ -2,3 +2,6 @@
     remote = dagshub
 ['remote "dagshub"']
     url = https://dagshub.com/miles2542/rossmann-ops.dvc
+    auth = basic
+    user = ${DAGSHUB_USERNAME}
+    password = ${DAGSHUB_PAT}

.env.example

@@ -2,3 +2,5 @@ MLFLOW_TRACKING_URI=https://dagshub.com/<your-username>/<your-repo>.mlflow
 MLFLOW_TRACKING_USERNAME=<your-username>
 MLFLOW_TRACKING_PASSWORD=<your-personal-access-token>
 GITHUB_PAT=<github-pat-with-repo-scope>
+DAGSHUB_USERNAME=<your-username>
+DAGSHUB_PAT=<your-dagshub-token>

.github/workflows/mlops_pipeline.yaml

@@ -51,7 +51,7 @@ jobs:
 
       # Strict: a failing test DOES block train + build jobs
       - name: Run tests
-        run: uv run pytest tests/ -v
+        run: uv run pytest tests/ -v --cov=src --cov-report=term-missing
 
   # ─── Job 2: Train & Export ─────────────────────────────────────────────────
   # Runs on: push to main (or tag), repository_dispatch (CT from drift)
@@ -60,8 +60,10 @@ jobs:
     name: Train Model
     runs-on: ubuntu-latest
     needs: ci
-    if: (github.event_name == 'push' && github.ref == 'refs/heads/main') || github.event_name == 'repository_dispatch'
+    if: (github.event_name == 'push' && (github.ref == 'refs/heads/main' || startsWith(github.ref, 'refs/tags/'))) || github.event_name == 'repository_dispatch'
     env:
+      DAGSHUB_USERNAME: ${{ secrets.DAGSHUB_USERNAME }}
+      DAGSHUB_PAT: ${{ secrets.DAGSHUB_PAT }}
       MLFLOW_TRACKING_URI: https://dagshub.com/${{ secrets.DAGSHUB_USERNAME }}/rossmann-ops.mlflow
       MLFLOW_TRACKING_USERNAME: ${{ secrets.DAGSHUB_USERNAME }}
       MLFLOW_TRACKING_PASSWORD: ${{ secrets.DAGSHUB_PAT }}
@@ -75,12 +77,6 @@ jobs:
       - name: Install dependencies
         run: uv sync --frozen
 
-      - name: Configure DVC credentials
-        run: |
-          uv run dvc remote modify dagshub --local auth basic
-          uv run dvc remote modify dagshub --local user ${{ secrets.DAGSHUB_USERNAME }}
-          uv run dvc remote modify dagshub --local password ${{ secrets.DAGSHUB_PAT }}
-
       - name: Pull data from DagsHub
         run: uv run dvc pull -r dagshub
 
@@ -108,6 +104,8 @@ jobs:
     if: github.event_name == 'workflow_dispatch'
     env:
       GITHUB_PAT: ${{ secrets.GITHUB_PAT }}
+      DAGSHUB_USERNAME: ${{ secrets.DAGSHUB_USERNAME }}
+      DAGSHUB_PAT: ${{ secrets.DAGSHUB_PAT }}
       MLFLOW_TRACKING_URI: https://dagshub.com/${{ secrets.DAGSHUB_USERNAME }}/rossmann-ops.mlflow
       MLFLOW_TRACKING_USERNAME: ${{ secrets.DAGSHUB_USERNAME }}
       MLFLOW_TRACKING_PASSWORD: ${{ secrets.DAGSHUB_PAT }}
@@ -121,12 +119,8 @@ jobs:
       - name: Install dependencies
         run: uv sync --frozen
 
-      - name: Configure DVC + pull data
-        run: |
-          uv run dvc remote modify dagshub --local auth basic
-          uv run dvc remote modify dagshub --local user ${{ secrets.DAGSHUB_USERNAME }}
-          uv run dvc remote modify dagshub --local password ${{ secrets.DAGSHUB_PAT }}
-          uv run dvc pull -r dagshub
+      - name: Pull data from DagsHub
+        run: uv run dvc pull -r dagshub
 
       # attack mode correctly exits 1 (detects poisoning) — mark continue-on-error
       # so the job passes and logs are visible, not swallowed in a red job failure
@@ -208,21 +202,71 @@ jobs:
     steps:
       - uses: actions/checkout@v4
 
-      - name: Create release and attach k8s manifests
+      - name: Download model artifacts
+        uses: actions/download-artifact@v4
+        with:
+          name: production-model
+          path: models/
+
+      - name: Rewrite manifests for DockerHub
+        run: |
+          sed -i "s|image: rossmann-api:latest|image: ${{ secrets.DOCKERHUB_USERNAME }}/rossmann-api:${{ github.ref_name }}|g" k8s/api.yaml
+          sed -i "s|image: rossmann-ui:latest|image: ${{ secrets.DOCKERHUB_USERNAME }}/rossmann-ui:${{ github.ref_name }}|g" k8s/ui.yaml
+
+      - name: Create Deployment Bundle (ZIP)
+        run: |
+          zip -r deployment-bundle-${{ github.ref_name }}.zip \
+            src/ \
+            scripts/ \
+            k8s/ \
+            configs/ \
+            Justfile \
+            pyproject.toml \
+            uv.lock \
+            models/store_target_means.json
+
+      - name: Create GitHub Release
         uses: softprops/action-gh-release@v2
         with:
           files: |
+            deployment-bundle-${{ github.ref_name }}.zip
             k8s/api.yaml
             k8s/ui.yaml
           body: |
-            ## Docker Images
-            ```
-            ${{ secrets.DOCKERHUB_USERNAME }}/rossmann-api:${{ github.ref_name }}
-            ${{ secrets.DOCKERHUB_USERNAME }}/rossmann-ui:${{ github.ref_name }}
-            ```
-
-            ## Deploy to Kubernetes
-            ```bash
-            kubectl apply -f k8s/api.yaml
-            kubectl apply -f k8s/ui.yaml
-            ```
+            # Release Specification: ${{ github.ref_name }}
+
+            This release provides the finalized configuration and artifacts for the evaluation of the Rossmann Store Sales demand forecasting system.
+
+            ## Asset Overview
+
+            ### Optimized Kubernetes Manifests
+            - **api.yaml & ui.yaml**: Version-controlled manifests pre-configured to resolve container images directly from DockerHub. These files allow for immediate deployment to a Kubernetes cluster without requiring a local build environment or source code access.
+
+            ### Deployment Bundle
+            - **deployment-bundle-${{ github.ref_name }}.zip**: A comprehensive archive containing the project source code, infrastructure automation logic, and pre-trained statistical artifacts. Required for local simulation, retraining, or a full system audit.
+
+            ## Deployment Instructions
+
+            ### Option 1: Direct Container Deployment
+            This method utilizes the pre-built images hosted on DockerHub.
+
+            1. Initialize the local cluster infrastructure:
+               ```bash
+               kind create cluster --name rossmann-cluster
+               ```
+            2. Apply the versioned manifests:
+               ```bash
+               kubectl apply -f api.yaml
+               kubectl apply -f ui.yaml
+               ```
+
+            ### Option 2: Full System Replication
+            This method initializes the complete MLOps environment.
+
+            1. Extract the deployment bundle ZIP.
+            2. Ensure a Python 3.12 environment and the `uv` package manager are installed.
+            3. Execute the setup sequence:
+               ```bash
+               just setup
+               ```
+            4. Verify cluster status or run local simulations as documented in the source `README.md`.

.gitignore

@@ -250,5 +250,7 @@ models/**
 *.pkl
 *.onnx
 *.ubj
+mlflow.db
+.coverage
 .ipynb_checkpoints/
 

mlruns/mlflow.db.dvc

@@ -0,0 +1,5 @@
+outs:
+- md5: 4aeac1783ef742745f35fc872c695bb5
+  size: 663552
+  hash: md5
+  path: mlflow.db

notebooks/03_optuna.ipynb

@@ -2,7 +2,7 @@
  "cells": [
   {
    "cell_type": "markdown",
-   "id": "0e99a74b",
+   "id": "0",
    "metadata": {},
    "source": [
     "# **Hyperparameter Tuning (Optuna)**\n",
@@ -15,7 +15,7 @@
   {
    "cell_type": "code",
    "execution_count": null,
-   "id": "a43c71c3",
+   "id": "1",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -58,7 +58,7 @@
   },
   {
    "cell_type": "markdown",
-   "id": "7e541216",
+   "id": "2",
    "metadata": {},
    "source": [
     "## **1. Data Setup & Split Logic**\n",
@@ -68,7 +68,7 @@
   {
    "cell_type": "code",
    "execution_count": null,
-   "id": "4fb4af1a",
+   "id": "3",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -94,7 +94,7 @@
   },
   {
    "cell_type": "markdown",
-   "id": "66c48248",
+   "id": "4",
    "metadata": {},
    "source": [
     "## **2. Feature & Folding Setup**\n",
@@ -104,7 +104,7 @@
   {
    "cell_type": "code",
    "execution_count": null,
-   "id": "70b16f90",
+   "id": "5",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -168,7 +168,7 @@
   },
   {
    "cell_type": "markdown",
-   "id": "82bfcffc",
+   "id": "6",
    "metadata": {},
    "source": [
     "## **3. Optuna Objective**\n",
@@ -178,7 +178,7 @@
   {
    "cell_type": "code",
    "execution_count": null,
-   "id": "5298dd2a",
+   "id": "7",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -235,7 +235,7 @@
   },
   {
    "cell_type": "markdown",
-   "id": "7adc4512",
+   "id": "8",
    "metadata": {},
    "source": [
     "## **4. Run Optimization (Example)**\n",
@@ -245,7 +245,7 @@
   {
    "cell_type": "code",
    "execution_count": null,
-   "id": "3749c21b",
+   "id": "9",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -263,7 +263,7 @@
   },
   {
    "cell_type": "markdown",
-   "id": "7393e098",
+   "id": "10",
    "metadata": {},
    "source": [
     "## **5. Tuning Summary & Design Justfication**\n",

scripts/check_docker.py

@@ -1,4 +1,5 @@
 import logging
+import os
 import subprocess
 import sys
 import time
@@ -24,7 +25,8 @@ def start_docker_desktop():
     logger.info("Docker not running. Attempting to start Docker Desktop...")
 
     if platform == "win32":
-        docker_path = Path(r"C:\Program Files\Docker\Docker\Docker Desktop.exe")
+        prog_files = os.environ.get("ProgramFiles", "C:\\Program Files")
+        docker_path = Path(prog_files) / "Docker" / "Docker" / "Docker Desktop.exe"
         if docker_path.exists():
             subprocess.Popen([str(docker_path)])
         else: