import json import ast from pathlib import Path from collections import namedtuple from future import standard_library from flask import Flask, request, render_template from flask_socketio import SocketIO, emit import database import broadcast import audio Droid = namedtuple( 'Droid', 'droid_uid, member_uid, name, material, weight, transmitter_type') Driver = namedtuple('Driver', 'member_uid, name, email') api_key = database.get_config('mot_api_key') site_base = database.get_config('mot_site_base') current_droid = Droid(droid_uid=0, member_uid=0, name="none", material="none", weight="none", transmitter_type="none") current_member = Driver(member_uid=0, name="none", email="none") current_run = 0 current_state = 0 app = Flask(__name__, template_folder='templates') app.config['key'] = database.get_config('app_key') socketio = SocketIO(app)
X = X[self.raw_features].copy() X_pp = self.preprocessing_pipeline.transform(X) X_pp = pd.DataFrame(X_pp, columns=self.get_feature_names()) return X_pp def get_feature_names(self): """ Feature names after preprocessing. Replicates the get_feature_names function in the sklearn Transformer classes. """ return self.raw_features if __name__ == "__main__": from preprocessing import PreProcessor # noqa # Load data db_config = db.get_config() train = db.load(*db_config, 'raw_train') X_train = train.drop('SalePrice', axis=1) # Fit and transform training data pp = PreProcessor() X_train_pp = pp.fit_transform(X_train) train_pp = X_train_pp.assign(SalePrice=train['SalePrice']) # Save preprocessed data and fitted preprocessor db.save(train_pp, *db_config, 'processed_train') joblib.dump(pp, os.path.join(DIR, '../pickle/PreProcessor.pkl'))
import database import os import pandas as pd # Kaggle download file # Not implemented - just manually download and read_csv instead DIR = os.path.abspath(os.path.dirname(__file__)) train = pd.read_csv(os.path.join(DIR, '../data/raw/train.csv')) test = pd.read_csv(os.path.join(DIR, '../data/raw/test.csv')) # Save data to database uri, db, _ = database.get_config() database.save(train, uri, db, 'dev', 'raw_train') database.save(test, uri, db, 'dev', 'raw_test')