import os
import numpy as np

# garcon basics
env = os.getenv('Environment', 'dev')
version = '1.0'
flow_name = 'ingest_training'

# s3 paths
if env == 'dev':
  s3_bucket = 's3://{env}-cucumbers/YouTubeDelivery'.format(env=env)
  s3_drop_path = '{bucket}/filedrop'.format(bucket=s3_bucket)

else:
  s3_bucket = 's3://{env}-datalytics/etl/videoservices/delivery_review'.format(env=env)
  s3_drop_path = 's3://{env}-datalytics/filedrop/store/videoservices/delivery_review'.format(env=env)

s3_archive_path = ('{bucket}/archive/{{year}}/{{month}}/'
                   '{{run_date}}'.format(bucket=s3_bucket))

s3_processed_path = ('{bucket}/processed/{{year}}/{{month}}/'
                     '{{run_date}}'.format(bucket=s3_bucket))
s3_model = '{bucket}/models/latest/'.format(bucket=s3_bucket)

latest_model = s3_model + 'forest_*.pkl'
next_model = s3_model + 'forest_{run_date}.pkl'

s3_predictions = ('{bucket}/catalog_predictions/{{year}}/{{month}}/'
                  '{{run_date}}_predictions.tsv.gz'.format(bucket=s3_bucket))

s3_machine_v_human = ('{bucket}/model_performance/{{year}}/{{month}}/'
                      '{{run_date}}_perf.tsv.gz'.format(bucket=s3_bucket))

# snowflake things
sf_db_config = {
    'user': os.environ.get('SNOWFLAKE_USER'),
    'password': os.environ.get('SNOWFLAKE_PASSWORD'),
    'account': os.environ.get('SNOWFLAKE_ACCOUNT'),
    'role': os.environ.get('SNOWFLAKE_ROLE'),
    'warehouse': os.environ.get('SNOWFLAKE_WAREHOUSE'),
    'db': os.environ.get('SNOWFLAKE_DATABASE'),
    'schema': os.environ.get('SNOWFLAKE_SCHEMA')
}

sf_wh = 'SCIENCE'
sf_fmt = 'LY_DEV.TEST.TSV_HEAD'

sf_db = 'prod'
sf_sch = 'delivery_review'

sf_status_table = '.'.join([sf_db, sf_sch, 'youtube_delivery_status'])
sf_results_delivered = '.'.join([sf_db, sf_sch,'youtube_delivery_history'])
sf_predictions = '.'.join([sf_db, sf_sch,'youtube_ml_catalog_review'])
sf_machine_v_human = '.'.join([sf_db, sf_sch,'machine_v_human'])
sf_meta = '.'.join([sf_db, sf_sch,'youtube_delivery_metadata'])

sf_results_stg = sf_results_delivered + '_stg'
# sf_results_cat_stg = sf_results_stg + '_stg'

# Google Sheets stuff
sheetname = 'Review'
sheet_id =  '13JiMEkmbOQx3_LW9p4OBEKKD_nJGowQXsJxYL8o4w0g'
sheet_range = 'A1:B'
gdoc_col = 'Search in Artist or Release name?'
rep_col = {'DELIVER TO YOUTUBE?': 'APPROVED'}

tabs_to_exclude = ['Blacklisted Label', 
                   'Various Artists', 
                   'Suspected Non-exclusive']

# Regular constants for preprocessing
target_dictionary = {'Yes' : 1, 
                     'No'  : 0,
                     'No ' : 0,
                     'NO'  : 0,
                     'Yes ': 1,
                     'yes' : 1}

float_cols = ['RELEASES_PER_ARTIST',
              'HIDES_PER_ARTIST',
              'UNITS_STREAMING_LABEL',
              'RELEASES_PER_LABEL',
              'ISRC_PER_RELEASE',
              'AVERAGE_TRACK_DURATION',
              'YEARS_SINCE_RELEASE']

bool_cols = ['IS_DELETED',
             'IS_COMPILATION_X',
             'IS_BLACKLISTED',
             'IS_VARIOUS_ARTIST']

floats =  {col: np.float for col in float_cols}
bools = {col: bool for col in bool_cols} 
dtypes = {**floats, **bools}

# what are we trying to optimize our model on?
target = 'APPROVED'
# what is model output?
prob = 'MODEL_YES_PROB'

# columns not used in to fit the model.
cols_to_drop = ['ARTISTID',
               'RELEASENAME',
               'ARTISTNAME',
               'DATE_ADDED',
               'NOTE',
               'UPC',
               'LABELID',
               'RELEASE_DATE',
               'DATE_ADDED',
               'IS_DELETED',
               'GENREID_9',
               'GENREID_25',
               'GENREID_4',
               'GENREID_5',
               'GENREID_24',
               'GENREID_17',
               'PROCESS_DATE',
               # 'NOT_FOR_DISTROBUTION', #V2 GET THIS INTO SNOWFLAKE FIRST
               # 'AUDIO_OR_VIDEO'        #V2 GET THIS INTO SNOWFLAKE FIRST
               ]
cols_to_drop += [target, prob]

# columns that can be null for a training set.
null_cols = ['NOTE', 
             'PROCESS_DATE']

null_cols += [target, prob]

genre_ids = ['1',
             '2',
             '3',
             '4',
             '5',
             '6',
             '7',
             '8',
             '9',
             '11',
             '12',
             '13',
             '14',
             '15',
             '16',
             '17',
             '18',
             '19',
             '20',
             '21',
             '24',
             '25',
             '27']

email_recipients = ['dataanalyticsdepartment@theorchard.com', 'ytprojects@theorchard.com']
email_header = 'New Deliveries Ingested'
email_body = ('Model performance against hand-labeled reviews here: {}\n'
              'Model performance over time here: {}\n'
              'Latest catalog assessment here: {}\n')