updated the labels to try to store in a better format

p2/quest/neurobiber-categorization.log

@@ -0,0 +1,6 @@
+starting the job at: Tue Jul 15 14:09:10 CDT 2025
+setting up the environment
+running the neurobiber labeling script
+neurobiber labeling pau
+job finished, cleaning up
+job pau at: Tue Jul 15 14:12:26 CDT 2025

p2/quest/python_scripts/neurobiber_labeling.py

@@ -110,9 +110,13 @@ if __name__ == "__main__":
     model, tokenizer = load_model_and_tokenizer()
     preds = predict_batch(model, tokenizer, docs)
     #new columns in the df for the predicted neurobiber items 
-    preds_cols = [f"neurobiber_{i+1}" for i in range(96)]
-    preds_df = pd.DataFrame(preds, columns=preds_cols, index=first_discussion_df.index)
-    final_discussion_df = pd.concat([first_discussion_df, preds_df], axis=1)
+    #preds_cols = [f"neurobiber_{i+1}" for i in range(96)]
+    #preds_df = pd.DataFrame(preds, columns=preds_cols, index=first_discussion_df.index)
+    #final_discussion_df = pd.concat([first_discussion_df, preds_df], axis=1)
+    
+    #assigning the preditions as a new column 
+    final_discussion_df = first_discussion_df.copy()
+    final_discussion_df["neurobiber_preds"] = list(preds)
     #assert that order has been preserved 
     for _ in range(10):
         random_index = random.choice(first_discussion_df.index)
@@ -120,7 +124,7 @@ if __name__ == "__main__":
     #assert that there are the same number of rows in first_discussion_df and second_discussion_df
     assert len(first_discussion_df) == len(final_discussion_df)
     # if passing the prior asserts, let's write to a csv
-    final_discussion_df.to_csv("/home/nws8519/git/mw-lifecycle-analysis/p2/quest/071425_neurobiber_labels.csv", index=False)
+    final_discussion_df.to_csv("/home/nws8519/git/mw-lifecycle-analysis/p2/quest/071525_neurobiber_labels.csv", index=False)
     print('neurobiber labeling pau')