VectorV17 commited on
Commit
6990e30
·
verified ·
1 Parent(s): 3a459a5

Update server.py

Browse files
Files changed (1) hide show
  1. server.py +108 -60
server.py CHANGED
@@ -7,86 +7,133 @@ from flask import Flask, jsonify, request
7
 
8
  app = Flask(__name__)
9
 
10
- # Load ML artifacts saved during training
11
  model = joblib.load('steel_model.joblib')
12
  scaler = joblib.load('scaler.joblib')
13
  model_columns = joblib.load('model_columns.joblib')
14
 
 
 
 
15
 
16
  def get_db_connection():
 
17
  return mysql.connector.connect(
18
- host=os.getenv('DB_HOST', 'freedatabase-host.com'),
19
- user=os.getenv('DB_USER', 'your_user'),
20
  password=os.getenv('DB_PASSWORD', 'your_password'),
21
  database=os.getenv('DB_NAME', 'your_dbname'),
22
  port=int(os.getenv('DB_PORT', 3306)),
 
23
  )
24
 
25
 
 
 
 
 
 
26
  @app.route('/predict', methods=['POST'])
27
  def predict():
28
  try:
29
  data = request.get_json()
30
 
31
- # Create single-row DataFrame from input json
32
- input_df = pd.DataFrame([data])
33
-
34
- # Perform One-Hot Encoding consistent with trained schema
35
- input_df = pd.get_dummies(input_df)
36
- input_df = input_df.reindex(columns=model_columns, fill_value=0)
37
-
38
- # Scale and Predict
39
- scaled_features = scaler.transform(input_df)
40
- predicted_kwh = round(float(model.predict(scaled_features)[0]), 2)
41
-
42
- # Calculate CO2 Estimate (approx. 0.00082 tCO2 per kWh for steel grid power)
 
 
 
 
 
 
 
 
 
 
 
 
 
43
  estimated_co2 = round(predicted_kwh * 0.00082, 4)
44
 
45
- # Green-ML Optimization Logic
46
- lag_pf = float(data.get('Lagging_Current_Power_Factor', 100))
47
- if lag_pf < 75.0:
48
- mitigation_status = (
49
- 'Critical: Low Power Factor. Engage capacitor banks.'
 
 
 
50
  )
51
- elif predicted_kwh > 60.0:
52
- mitigation_status = (
53
- 'Warning: Peak Load Detected. Stagger furnace operations.'
 
54
  )
55
- else:
56
- mitigation_status = 'Optimal Operational Efficiency'
57
 
58
- # Save Log Entry to MySQL
59
- conn = get_db_connection()
60
- cursor = conn.cursor()
61
- insert_query = """
62
- INSERT INTO energy_predictions
63
- (lagging_reactive_power, leading_reactive_power, lagging_power_factor, leading_power_factor, nsm, week_status, day_of_week, load_type, predicted_usage_kwh)
64
- VALUES (%s, %s, %s, %s, %s, %s, %s, %s, %s)
65
- """
66
- cursor.execute(
67
- insert_query,
68
- (
69
- data.get('Lagging_Current_Reactive_Power_kVarh'),
70
- data.get('Leading_Current_Reactive_Power_kVarh'),
71
- lag_pf,
72
- data.get('Leading_Current_Power_Factor'),
73
- data.get('NSM'),
74
- data.get('WeekStatus'),
75
- data.get('Day_of_week'),
76
- data.get('Load_Type'),
77
- predicted_kwh,
78
- ),
79
  )
80
- conn.commit()
81
- cursor.close()
82
- conn.close()
83
 
84
- return jsonify({
85
- 'status': 'success',
86
- 'predicted_usage_kwh': predicted_kwh,
87
- 'estimated_co2': estimated_co2,
88
- 'mitigation_status': mitigation_status,
89
- })
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
90
 
91
  except Exception as e:
92
  return jsonify({'status': 'error', 'message': str(e)}), 500
@@ -94,12 +141,12 @@ def predict():
94
 
95
  @app.route('/history', methods=['GET'])
96
  def get_history():
97
- """Retrieves last 20 prediction logs from MySQL for chart rendering."""
98
  try:
99
  conn = get_db_connection()
100
  df_logs = pd.read_sql(
101
- 'SELECT id, timestamp, predicted_usage_kwh FROM energy_predictions'
102
- ' ORDER BY id DESC LIMIT 20',
103
  conn,
104
  )
105
  conn.close()
@@ -107,8 +154,9 @@ def get_history():
107
  'status': 'success',
108
  'data': df_logs.to_dict(orient='records'),
109
  })
110
- except Exception as e:
111
- return jsonify({'status': 'error', 'message': str(e)}), 500
 
112
 
113
 
114
  if __name__ == '__main__':
 
7
 
8
  app = Flask(__name__)
9
 
10
+ # Load Pre-trained Artifacts
11
  model = joblib.load('steel_model.joblib')
12
  scaler = joblib.load('scaler.joblib')
13
  model_columns = joblib.load('model_columns.joblib')
14
 
15
+ # Local Fallback Log Memory (Used if remote MySQL is offline)
16
+ in_memory_logs = []
17
+
18
 
19
  def get_db_connection():
20
+ """Establishes connection to remote MySQL database with a 3-second timeout."""
21
  return mysql.connector.connect(
22
+ host=os.getenv('DB_HOST', 'sql.freedatabase.com'),
23
+ user=os.getenv('DB_USER', 'your_username'),
24
  password=os.getenv('DB_PASSWORD', 'your_password'),
25
  database=os.getenv('DB_NAME', 'your_dbname'),
26
  port=int(os.getenv('DB_PORT', 3306)),
27
+ connection_timeout=3,
28
  )
29
 
30
 
31
+ @app.route('/health', methods=['GET'])
32
+ def health():
33
+ return jsonify({'status': 'healthy'}), 200
34
+
35
+
36
  @app.route('/predict', methods=['POST'])
37
  def predict():
38
  try:
39
  data = request.get_json()
40
 
41
+ # Map inputs to structured DataFrame
42
+ raw_input = {
43
+ 'Lagging_Current_Reactive_Power_kVarh': float(
44
+ data.get('lagging_kvarh', 0)
45
+ ),
46
+ 'Leading_Current_Reactive_Power_kVarh': float(
47
+ data.get('leading_kvarh', 0)
48
+ ),
49
+ 'Lagging_Current_Power_Factor': float(data.get('lagging_pf', 100)),
50
+ 'Leading_Current_Power_Factor': float(data.get('leading_pf', 100)),
51
+ 'NSM': int(data.get('nsm', 0)),
52
+ 'WeekStatus': data.get('week_status', 'Weekday'),
53
+ 'Day_of_week': data.get('day_of_week', 'Monday'),
54
+ 'Load_Type': data.get('load_type', 'Light_Load'),
55
+ }
56
+
57
+ input_df = pd.DataFrame([raw_input])
58
+ encoded_df = pd.get_dummies(input_df)
59
+ encoded_df = encoded_df.reindex(columns=model_columns, fill_value=0)
60
+
61
+ # Execute Model Inference
62
+ scaled_input = scaler.transform(encoded_df)
63
+ predicted_kwh = round(float(model.predict(scaled_input)[0]), 2)
64
+
65
+ # Estimate CO2 Emissions (tCO2 factor per kWh in heavy industry grid)
66
  estimated_co2 = round(predicted_kwh * 0.00082, 4)
67
 
68
+ # Power Grid Inefficiency & Mitigation Diagnostics
69
+ grid_issues = []
70
+ lag_pf = raw_input['Lagging_Current_Power_Factor']
71
+ lag_kvarh = raw_input['Lagging_Current_Reactive_Power_kVarh']
72
+
73
+ if lag_pf < 85.0:
74
+ grid_issues.append(
75
+ f'Low Lagging Power Factor ({lag_pf}%). Penalty risk!'
76
  )
77
+ if lag_kvarh > 15.0:
78
+ grid_issues.append(
79
+ f'High Inductive Reactive Load ({lag_kvarh} kVarh). Inefficient'
80
+ ' transformer draw.'
81
  )
82
+ if predicted_kwh > 50.0:
83
+ grid_issues.append('High Peak Demand (>50 kWh). Surge pricing applies.')
84
 
85
+ mitigation_status = (
86
+ ' | '.join(grid_issues)
87
+ if grid_issues
88
+ else 'Optimal Grid Efficiency (No Penalties)'
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
89
  )
 
 
 
90
 
91
+ # Persistent MySQL Logging with Fallback
92
+ db_status = 'Logged to MySQL'
93
+ try:
94
+ conn = get_db_connection()
95
+ cursor = conn.cursor()
96
+ query = """
97
+ INSERT INTO energy_predictions
98
+ (lagging_reactive_power, leading_reactive_power, lagging_power_factor, leading_power_factor, nsm, week_status, day_of_week, load_type, predicted_usage_kwh)
99
+ VALUES (%s, %s, %s, %s, %s, %s, %s, %s, %s)
100
+ """
101
+ cursor.execute(
102
+ query,
103
+ (
104
+ raw_input['Lagging_Current_Reactive_Power_kVarh'],
105
+ raw_input['Leading_Current_Reactive_Power_kVarh'],
106
+ raw_input['Lagging_Current_Power_Factor'],
107
+ raw_input['Leading_Current_Power_Factor'],
108
+ raw_input['NSM'],
109
+ raw_input['WeekStatus'],
110
+ raw_input['Day_of_week'],
111
+ raw_input['Load_Type'],
112
+ predicted_kwh,
113
+ ),
114
+ )
115
+ conn.commit()
116
+ cursor.close()
117
+ conn.close()
118
+ except Exception as db_err:
119
+ db_status = 'In-Memory Fallback (MySQL Offline)'
120
+ in_memory_logs.append({
121
+ 'id': len(in_memory_logs) + 1,
122
+ 'predicted_usage_kwh': predicted_kwh,
123
+ 'co2': estimated_co2,
124
+ })
125
+
126
+ return (
127
+ jsonify({
128
+ 'status': 'success',
129
+ 'predicted_usage_kwh': predicted_kwh,
130
+ 'estimated_co2': estimated_co2,
131
+ 'power_factor_efficiency': f'{lag_pf}%',
132
+ 'mitigation_status': mitigation_status,
133
+ 'db_status': db_status,
134
+ }),
135
+ 200,
136
+ )
137
 
138
  except Exception as e:
139
  return jsonify({'status': 'error', 'message': str(e)}), 500
 
141
 
142
  @app.route('/history', methods=['GET'])
143
  def get_history():
144
+ """Retrieves historical logs for analytics chart rendering."""
145
  try:
146
  conn = get_db_connection()
147
  df_logs = pd.read_sql(
148
+ 'SELECT id, predicted_usage_kwh FROM energy_predictions ORDER BY id'
149
+ ' DESC LIMIT 20',
150
  conn,
151
  )
152
  conn.close()
 
154
  'status': 'success',
155
  'data': df_logs.to_dict(orient='records'),
156
  })
157
+ except Exception:
158
+ # Return local in-memory logs if MySQL fails
159
+ return jsonify({'status': 'success', 'data': in_memory_logs[-20:]})
160
 
161
 
162
  if __name__ == '__main__':