def main(inp_fname, move_info_fname, output_fname):
inp, move_info = getData(inp_fname, move_info_fname)
fw = open(output_fname, "w")
for txt, player in zip(inp, move_info):
out = postprocess.postProcess(txt, player=player)
fw.write(out + "\n")
fw.close()
def predict():
# Pre-process Data
print('Started predict')
sm9_data_with_RITM, sm9_data_without_RITM = preprocess.getSM9TestData(
os.path.join(path.DIR_NAME, path.SM9_TEST_DATA))
sn_data = preprocess.getServiceNowTestData(
os.path.join(path.DIR_NAME, path.SERVICE_NOW_TEST_DATA))
merged_data_with_cat_item, merged_data_without_cat_item = preprocess.mergeTestData(
sm9_data_with_RITM, sn_data)
sm9_data_without_RITM, merged_data_without_cat_item = preprocess.prepareDataByRBCTitle(
sm9_data_without_RITM, merged_data_without_cat_item)
# sm9_data_without_RITM = sm9_data_without_RITM[['RBC Line Item Title']]
# merged_data_without_cat_item = merged_data_without_cat_item[['RBC Line Item Title']]
tickets_by_title = pd.concat(
[sm9_data_without_RITM, merged_data_without_cat_item])
sorted_analyst_prob_by_item = predictModelByCatItem(
merged_data_with_cat_item['cat_item'],
os.path.join(path.DIR_NAME, path.CAT_ITEM_MODEL),
os.path.join(path.DIR_NAME, path.COUNT_VEC_CAT_ITEM))
cat_item_ticket_recommendations = postprocess.postProcess(
sorted_analyst_prob_by_item, merged_data_with_cat_item)
sorted_analyst_prob_by_title = predictModelByTitle(
tickets_by_title['RBC Line Item Title'],
os.path.join(path.DIR_NAME, path.RBC_TITLE_MODEL),
os.path.join(path.DIR_NAME, path.COUNT_VEC_BY_TITLE))
rbc_title_ticket_recommendations = postprocess.postProcess(
sorted_analyst_prob_by_title, tickets_by_title)
ticket_recommendations = pd.concat(
[cat_item_ticket_recommendations, rbc_title_ticket_recommendations])
recommendations = "tickets_recommendations_" + datetime.datetime.today(
).strftime('%Y-%m-%d') + ".xlsx"
filename = os.path.join(path.DIR_NAME, path.RECOMMENDATIONS_DIR,
recommendations)
# filename = "tickets_recommendation_" + datetime.datetime.now().isoformat() + ".xlsx"
# Prefarably save as excel
# ticket_recommendations.to_csv(path_or_buf=filename , encoding="Latin-1" , index = False)
# ticket_recommendations.to_excel(path_or_buf=filename , index = False)
# Save as excel
writer = pd.ExcelWriter(filename, engine='xlsxwriter')
ticket_recommendations.to_excel(writer, 'Sheet1', index=False)
writer.save()
print('Ended predict')
def compute(pool, namespace=''):
# 5th pass: High-level descriptors that depend on others, but we
# don't need to stream the audio anymore
# Average Level
from level import levelAverage
levelAverage(pool, namespace)
# SFX Descriptors
sfxPitch(pool, namespace)
# Tuning System Features
tuningSystemFeatures(pool, namespace)
# Pool Cleaning (removing temporary descriptors):
tonalPoolCleaning(pool, namespace)
# Add missing descriptors which are not computed yet, but will be for the
# final release or during the 1.x cycle. However, the schema need to be
# complete before that, so just put default values for these.
postProcess(pool, namespace)
def compute(pool, namespace=''):
# 5th pass: High-level descriptors that depend on others, but we
# don't need to stream the audio anymore
# Average Level
from level import levelAverage
levelAverage(pool, namespace)
# SFX Descriptors
sfxPitch(pool, namespace)
# Tuning System Features
tuningSystemFeatures(pool, namespace)
# Pool Cleaning (removing temporary descriptors):
tonalPoolCleaning(pool, namespace)
# Add missing descriptors which are not computed yet, but will be for the
# final release or during the 1.x cycle. However, the schema need to be
# complete before that, so just put default values for these.
postProcess(pool, namespace)