name: ML Training CI

on:
  push:
    branches: [ main, develop ]
    paths:
      - 'ml/**'
      - 'services/ml-optimizer/**'
  pull_request:
    branches: [ main ]
    paths:
      - 'ml/**'
      - 'services/ml-optimizer/**'
  schedule:
    # Run weekly training on Sundays at 2 AM UTC
    - cron: '0 2 * * 0'
  workflow_dispatch:
    inputs:
      model_type:
        description: 'Type of model to train'
        required: true
        default: 'all'
        type: choice
        options:
          - all
          - anomaly_detection
          - backup_prediction
          - optimization

jobs:
  validate-data:
    runs-on: ubuntu-latest
    steps:
      - name: Checkout
        uses: actions/checkout@v4
        
      - name: Set up Python 3.11
        uses: actions/setup-python@v5
        with:
          python-version: '3.11'
          
      - name: Cache pip packages
        uses: actions/cache@v4
        with:
          path: ~/.cache/pip
          key: ${{ runner.os }}-pip-ml-${{ hashFiles('ml/**/requirements.txt') }}
          restore-keys: |
            ${{ runner.os }}-pip-ml-
            
      - name: Install dependencies
        run: |
          python -m pip install --upgrade pip
          pip install pandas numpy scikit-learn pytest tensorflow joblib
          
      - name: Validate training datasets
        run: |
          python -c "
          import os
          import pandas as pd
          import numpy as np
          
          datasets_dir = 'ml/datasets'
          if os.path.exists(datasets_dir):
              for file in os.listdir(datasets_dir):
                  if file.endswith('.csv'):
                      df = pd.read_csv(os.path.join(datasets_dir, file))
                      print(f'Dataset {file}: {df.shape[0]} rows, {df.shape[1]} columns')
                      print(f'Missing values: {df.isnull().sum().sum()}')
          else:
              print('No datasets directory found, creating placeholder')
              os.makedirs(datasets_dir, exist_ok=True)
          "
          
  train-anomaly-detection:
    needs: validate-data
    runs-on: ubuntu-latest
    if: ${{ github.event.inputs.model_type == 'anomaly_detection' || github.event.inputs.model_type == 'all' || github.event.inputs.model_type == '' }}
    steps:
      - name: Checkout
        uses: actions/checkout@v4
        
      - name: Set up Python 3.11
        uses: actions/setup-python@v5
        with:
          python-version: '3.11'
          
      - name: Cache pip packages
        uses: actions/cache@v4
        with:
          path: ~/.cache/pip
          key: ${{ runner.os }}-pip-anomaly-${{ hashFiles('ml/models/anomaly_detection/requirements.txt') }}
          restore-keys: |
            ${{ runner.os }}-pip-anomaly-
            
      - name: Install dependencies
        run: |
          python -m pip install --upgrade pip
          cd ml/models/anomaly_detection
          pip install scikit-learn pandas numpy joblib matplotlib seaborn pytest tensorflow
          
      - name: Train anomaly detection model
        run: |
          cd ml/models/anomaly_detection
          python anomaly_detector.py
          
      - name: Test model
        run: |
          cd ml/models/anomaly_detection
          python -m pytest test_*.py -v || echo "No tests found"
          
      - name: Upload model artifacts
        uses: actions/upload-artifact@v4
        with:
          name: anomaly-detection-model
          path: |
            ml/models/anomaly_detection/*.pkl
            ml/models/anomaly_detection/*.joblib
            ml/models/anomaly_detection/metrics.json
            
  train-backup-prediction:
    needs: validate-data
    runs-on: ubuntu-latest
    if: ${{ github.event.inputs.model_type == 'backup_prediction' || github.event.inputs.model_type == 'all' || github.event.inputs.model_type == '' }}
    steps:
      - name: Checkout
        uses: actions/checkout@v4
        
      - name: Set up Python 3.11
        uses: actions/setup-python@v5
        with:
          python-version: '3.11'
          
      - name: Cache pip packages
        uses: actions/cache@v4
        with:
          path: ~/.cache/pip
          key: ${{ runner.os }}-pip-backup-${{ hashFiles('ml/models/backup_prediction/requirements.txt') }}
          restore-keys: |
            ${{ runner.os }}-pip-backup-
            
      - name: Install dependencies
        run: |
          python -m pip install --upgrade pip
          cd ml/models/backup_prediction
          pip install scikit-learn pandas numpy joblib matplotlib seaborn pytest tensorflow
          
      - name: Train backup prediction model
        run: |
          cd ml/models/backup_prediction
          python backup_predictor.py
          
      - name: Test model
        run: |
          cd ml/models/backup_prediction
          python -m pytest test_*.py -v || echo "No tests found"
          
      - name: Upload model artifacts
        uses: actions/upload-artifact@v4
        with:
          name: backup-prediction-model
          path: |
            ml/models/backup_prediction/*.pkl
            ml/models/backup_prediction/*.joblib
            ml/models/backup_prediction/metrics.json
            
  model-validation:
    needs: [train-anomaly-detection, train-backup-prediction]
    runs-on: ubuntu-latest
    if: always()
    steps:
      - name: Checkout
        uses: actions/checkout@v4
        
      - name: Download all model artifacts
        uses: actions/download-artifact@v4
        with:
          path: trained-models
          
      - name: Set up Python 3.11
        uses: actions/setup-python@v5
        with:
          python-version: '3.11'
          
      - name: Install validation dependencies
        run: |
          python -m pip install --upgrade pip
          pip install scikit-learn pandas numpy joblib tensorflow
          
      - name: Validate trained models
        run: |
          python -c "
          import os
          import joblib
          import pickle
          
          models_dir = 'trained-models'
          if os.path.exists(models_dir):
              for root, dirs, files in os.walk(models_dir):
                  for file in files:
                      if file.endswith(('.pkl', '.joblib')):
                          try:
                              model_path = os.path.join(root, file)
                              if file.endswith('.pkl'):
                                  model = pickle.load(open(model_path, 'rb'))
                              else:
                                  model = joblib.load(model_path)
                              print(f'Successfully loaded model: {model_path}')
                              print(f'Model type: {type(model)}')
                          except Exception as e:
                              print(f'Failed to load {model_path}: {e}')
          else:
              print('No trained models found')
          "