Dockerizing the repo

.gitignore

@@ -0,0 +1,6 @@
+/Code of Winning Methods/A5/raw_data
+/Code of Winning Methods/A5/proc_data
+/Code of Winning Methods/A5/output_data
+/Code of Winning Methods/A1/2. data
+/Code of Winning Methods/A1/5. models
+/Code of Winning Methods/A1/4. logs

Code of Winning Methods/A1/2. data/readme.txt

@@ -5,4 +5,4 @@ This folder should contain the following files:
 - sample_submission.csv
 - sell_prices.csv
 
-These can be found in the M5 competition repo.
+These can be found in the M5 competition repo. 

Code of Winning Methods/A1/3. code/1. preprocessing/1. preprocessing.py

@@ -1,17 +1,13 @@
 #!/usr/bin/env python
 # coding: utf-8
 
-# ## Please input your directory for the top level folder
-# folder name : SUBMISSION MODEL
-
-# In[ ]:
-
-
-dir_ = '/home/artemis/M5/A1-Yeon/' # input only here
 
+dir_ = '../../' # input only here
+# dir_ = '/A1/' # input only here
 
-# In[ ]:
+# ../../2.\ data/
 
+# C:\Users\neuromancer\M5-methods\Code of Winning Methods\A1\1. document
 
 raw_data_dir = dir_+'2. data/'
 processed_data_dir = dir_+'2. data/processed/'
@@ -371,7 +367,11 @@ def merge_by_concat(df1, df2, merge_on):
 
 # Make some features from date
 grid_df['tm_d'] = grid_df['date'].dt.day.astype(np.int8)
-grid_df['tm_w'] = grid_df['date'].dt.week.astype(np.int8)
+
+## replacing newversion of week version 
+grid_df['tm_w'] = grid_df['date'].dt.isocalendar().week.astype(np.int8)
+
+# grid_df['tm_w'] = grid_df['date'].dt.week.astype(np.int8)
 grid_df['tm_m'] = grid_df['date'].dt.month.astype(np.int8)
 grid_df['tm_y'] = grid_df['date'].dt.year
 grid_df['tm_y'] = (grid_df['tm_y'] - grid_df['tm_y'].min()).astype(np.int8)
@@ -523,7 +523,7 @@ def merge_by_concat(df1, df2, merge_on):
 
 
 # In[ ]:
-
+### cat asdasdsad
 
 #################################################################################
 print('Save Mean/Std encoding')

Code of Winning Methods/A1/3. code/2. train/1-1. recursive_store_TRAIN.py

@@ -7,7 +7,7 @@
 # In[ ]:
 
 
-dir_ = '/home/artemis/M5/A1-Yeon/' # input only here
+dir_ = '../../' # input only here
 
 
 # #### setting other directory
@@ -234,10 +234,11 @@ def make_lag_roll(LAG_DAY):
     grid_df = grid_df[keep_cols]
 
     d_sales = grid_df[['d','sales']]
-    substitute = d_sales['sales'].values
+    # refactored
+    substitute = d_sales['sales'].to_numpy(copy=True)
     substitute[(d_sales['d'] > END_TRAIN)] = np.nan
     grid_df['sales'] = substitute
-    
+
     grid_df.to_pickle(processed_data_dir+'test_'+store_id+'.pkl')
     del grid_df, d_sales, substitute
 
@@ -246,7 +247,8 @@ def make_lag_roll(LAG_DAY):
     estimator = lgb.train(lgb_params,
                           train_data,
                           valid_sets = [valid_data],
-                          verbose_eval = 100,
+                          # this is deprecated
+                        #   verbose_eval = 100,
                           )
 
     # display(pd.DataFrame({'name':estimator.feature_name(),

Code of Winning Methods/A1/3. code/2. train/1-2. recursive_store_cat_TRAIN.py

@@ -7,7 +7,7 @@
 # In[ ]:
 
 
-dir_ = '/home/artemis/M5/A1-Yeon/' # input only here
+dir_ = '../../' # input only here
 
 
 # #### setting other directory
@@ -242,7 +242,7 @@ def make_lag_roll(LAG_DAY):
         grid_df = grid_df[keep_cols]
 
         d_sales = grid_df[['d','sales']]
-        substitute = d_sales['sales'].values
+        substitute = d_sales['sales'].to_numpy(copy=True)
         substitute[(d_sales['d'] > END_TRAIN)] = np.nan
         grid_df['sales'] = substitute
 
@@ -253,7 +253,7 @@ def make_lag_roll(LAG_DAY):
         estimator = lgb.train(lgb_params,
                               train_data,
                               valid_sets = [valid_data],
-                              verbose_eval = 100
+                             #  verbose_eval = 100
 
                               )
 

Code of Winning Methods/A1/3. code/2. train/1-3. recursive_store_dept_TRAIN.py

@@ -7,7 +7,7 @@
 # In[ ]:
 
 
-dir_ = '/home/artemis/M5/A1-Yeon/' # input only here
+dir_ = '../../' # input only here
 
 
 # #### setting other directory
@@ -241,7 +241,7 @@ def make_lag_roll(LAG_DAY):
         grid_df = grid_df[keep_cols]
 
         d_sales = grid_df[['d','sales']]
-        substitute = d_sales['sales'].values
+        substitute = d_sales['sales'].to_numpy(copy=True)
         substitute[(d_sales['d'] > END_TRAIN)] = np.nan
         grid_df['sales'] = substitute
 
@@ -252,7 +252,7 @@ def make_lag_roll(LAG_DAY):
         estimator = lgb.train(lgb_params,
                               train_data,
                               valid_sets = [valid_data],
-                              verbose_eval = 100
+                             #  verbose_eval = 100
 
                               )
 

Code of Winning Methods/A1/3. code/2. train/2-1. nonrecursive_store_TRAIN.py

@@ -7,7 +7,7 @@
 # In[ ]:
 
 
-dir_ = '/home/artemis/M5/A1-Yeon/' # input only here
+dir_ = '../../' # input only here
 
 
 # #### setting other directory
@@ -248,7 +248,7 @@ def prepare_data(store):
         valid_data = lgb.Dataset(grid_df[vl_mask][model_var], 
                            label=grid_df[vl_mask]['sales'])
 
-        m_lgb = lgb.train(lgb_params, train_data, valid_sets = [valid_data, train_data], verbose_eval=100) 
+        m_lgb = lgb.train(lgb_params, train_data, valid_sets = [valid_data, train_data]) 
         # display(pd.DataFrame({'name':m_lgb.feature_name(),
         #                       'imp':m_lgb.feature_importance()}).sort_values('imp',ascending=False).head(25))
 

Code of Winning Methods/A1/3. code/2. train/2-2. nonrecursive_store_cat_TRAIN.py

@@ -7,7 +7,7 @@
 # In[ ]:
 
 
-dir_ = '/home/artemis/M5/A1-Yeon/' # input only here
+dir_ = '../../' # input only here
 
 
 # #### setting other directory
@@ -254,7 +254,7 @@ def prepare_data(store, state):
             valid_data = lgb.Dataset(grid_df[vl_mask][model_var], 
                                label=grid_df[vl_mask]['sales'])
 
-            m_lgb = lgb.train(lgb_params, train_data, valid_sets = [valid_data, train_data], verbose_eval=100) 
+            m_lgb = lgb.train(lgb_params, train_data, valid_sets = [valid_data, train_data]) 
             model_name = model_dir+'non_recur_model_'+store+'_'+state+'.bin'
             pickle.dump(m_lgb, open(model_name, 'wb'))
 

Code of Winning Methods/A1/3. code/2. train/2-3. nonrecursive_store_dept_TRAIN.py

@@ -7,7 +7,7 @@
 # In[ ]:
 
 
-dir_ = '/home/artemis/M5/A1-Yeon/' # input only here
+dir_ = '../../' # input only here
 
 
 # #### setting other directory
@@ -253,7 +253,7 @@ def prepare_data(store, state):
             valid_data = lgb.Dataset(grid_df[vl_mask][model_var], 
                                label=grid_df[vl_mask]['sales'])
 
-            m_lgb = lgb.train(lgb_params, train_data, valid_sets = [valid_data, train_data], verbose_eval=100) 
+            m_lgb = lgb.train(lgb_params, train_data, valid_sets = [valid_data, train_data]) 
             model_name = model_dir+'non_recur_model_'+store+'_'+state+'.bin'
             pickle.dump(m_lgb, open(model_name, 'wb'))
 

Code of Winning Methods/A1/3. code/3. predict/1-1. recursive_store_PREDICT.py

@@ -7,7 +7,7 @@
 # In[ ]:
 
 
-dir_ = '/home/artemis/M5/A1-Yeon/' # input only here
+dir_ = '../../' # input only here
 
 
 # #### setting other directory
@@ -21,7 +21,6 @@
 model_dir = dir_+'5. models/'
 submission_dir = dir_+'6. submissions/'
 
-
 # In[ ]:
 
 

Code of Winning Methods/A1/3. code/3. predict/1-2. recursive_store_cat_PREDICT.py

@@ -7,7 +7,7 @@
 # In[ ]:
 
 
-dir_ = '/home/artemis/M5/A1-Yeon/' # input only here
+dir_ = '../../' # input only here
 
 
 # #### setting other directory

Code of Winning Methods/A1/3. code/3. predict/1-3. recursive_store_dept_PREDICT.py

@@ -7,7 +7,7 @@
 # In[ ]:
 
 
-dir_ = '/home/artemis/M5/A1-Yeon/' # input only here
+dir_ = '../../' # input only here
 
 
 # #### setting other directory