From 9f4f0d20cd5526d77f598e19a1332eaa9a7a976f Mon Sep 17 00:00:00 2001
From: EC2 Default User <ec2-user@ip-10-130-27-142.columbuschildrens.net>
Date: Mon, 6 Sep 2021 10:05:09 -0400
Subject: [PATCH] #27 enabled shap plots

---
 clin_msi/msi_model_scripts/apply_msi_model.py | 16 ++++-----
 clin_msi/msi_training.py                      | 35 -------------------
 clin_msi/predict.py                           |  8 ++---
 3 files changed, 12 insertions(+), 47 deletions(-)
 delete mode 100644 clin_msi/msi_training.py

diff --git a/clin_msi/msi_model_scripts/apply_msi_model.py b/clin_msi/msi_model_scripts/apply_msi_model.py
index 6bbf162..521e46a 100644
--- a/clin_msi/msi_model_scripts/apply_msi_model.py
+++ b/clin_msi/msi_model_scripts/apply_msi_model.py
@@ -130,11 +130,11 @@ def apply_model(infile,moddir,outfile,shap_plot_dir=None):
     df=df.rename(columns={sampcol:'samp'})
     dfnew,shapdict,curfeats=apply_mod_to_dataframe(df,moddir)
     dfnew[['samp','yprob']].to_csv(outfile,index=False)
-    # ## GRAB  SHAP DATA
-    # testdat,bmdict=grab_shap_data(dfnew,shapdict,curfeats)
-    # ## EXPORT SHAP PLOTS
-    # for i in range(len(testdat)):
-    #     shap_outfile=shap_plot_dir+"/shap_plot_"+str(i+1)+'.png'
-    #     shaprec=testdat.iloc[i]
-    #     shapdict_loc=bmdict[i]
-    #     build_n_save_shap_plot(shaprec,shapdict_loc,shap_outfile)
+    ## GRAB  SHAP DATA
+    testdat,bmdict=grab_shap_data(dfnew,shapdict,curfeats)
+    ## EXPORT SHAP PLOTS
+    for i in range(len(testdat)):
+        shap_outfile=shap_plot_dir+"/shap_plot_"+str(i+1)+'.png'
+        shaprec=testdat.iloc[i]
+        shapdict_loc=bmdict[i]
+        build_n_save_shap_plot(shaprec,shapdict_loc,shap_outfile)
diff --git a/clin_msi/msi_training.py b/clin_msi/msi_training.py
deleted file mode 100644
index a9c66da..0000000
--- a/clin_msi/msi_training.py
+++ /dev/null
@@ -1,35 +0,0 @@
-import pickle
-import logging
-
-import xgboost as xgb
-
-## FUNCTIONS
-## INTERNAL TRAINING FUNCTION
-def modfit(df,curfeats):
-    X=df[curfeats]
-    Y=df['y']
-    model = xgb.XGBClassifier(max_depth=1,
-                              colsample_bytree=0.001,
-                              n_estimators=500,
-                              eval_metric='error',
-                              use_label_encoder=False ## to prevent a warning
-                             )
-    model.fit(X,Y)
-    return model
-
-## MAIN FUNCTION
-def train_models(df,moddir):
-    curfeats=[x for x in df if x != 'y']
-    moddict={}
-    nrun=500
-    modfiles=[moddir+'/xgb_'+str(j+1)+'.pkl' for j in range(nrun)]
-    for j in range(nrun):
-        if (j % 10 == 9):
-            ## DINKY PROGRESS REPORT
-            logging.info("Training model "+str(j+1))
-        traininds_loc=df.sample(frac=0.8).index
-        mymod=modfit(df.loc[traininds_loc],curfeats)
-        pkl_connect = open(modfiles[j], 'wb')
-        pickle.dump(mymod, pkl_connect)
-        pkl_connect.close()
-    return(modfiles)
diff --git a/clin_msi/predict.py b/clin_msi/predict.py
index 02849d9..362eccb 100644
--- a/clin_msi/predict.py
+++ b/clin_msi/predict.py
@@ -6,8 +6,8 @@
 import pysam
 import pandas as pd
 
-from .count_normalization.normalize_counts import parse_raw_data
-from .msi_model_scripts.apply_msi_model import apply_model
+from count_normalization.normalize_counts import parse_raw_data
+from msi_model_scripts.apply_msi_model import apply_model
 
 
 def repeat_finder(s):
@@ -96,7 +96,7 @@ def predict(
 
     #apply model to normalized msi counts
     final_results_file = os.path.join(output_dir, sample_name + '_MSIscore.txt')
-    apply_model(os.path.join(output_dir, sample_name + '_normalized.csv'), model_dir, final_results_file)
+    apply_model(os.path.join(output_dir, sample_name + '_normalized.csv'), model_dir, final_results_file,shap_plot_dir=output_dir)
 
 if __name__ == '__main__':
-    predict()
\ No newline at end of file
+    predict()