Upload configuration file

config/parse_zayne_pdfs.yaml

@@ -52,12 +52,28 @@ operators:
       map_config:
         user_message: |
           I will provide you with a given question and the text from which it was extracted from.
-          Given the text, I want you to provide additional context for the question such that it is answerable on its own. 
+          Provide the text before the question in the extracted text that is useful for answering the question and the question itself. 
+
+          For example: 
+          Extracted Question: Calculate the chemical amount (in mol or mmol) of nitric acid that reacts with the
+            5.000 g sample of this mineral.
+          Extracted Text: A sample of a different mineral is analysed by the same methods. This mineral also contains only Pb2+, CO3 2−, OH− and O2− ions.
+            When a 5.000 g sample of this mineral is treated with 25.00 mL of 2.000 mol L−1 nitric acid
+            (HNO3), 0.5214 g of carbon dioxide is released, and 0.01051 mol of the acid remains.
+            When subjected to thermal decomposition, 5.000 g of this mineral loses 0.5926 g.
+            (g) Calculate the chemical amount (in mol or mmol) of nitric acid that reacts with the
+            5.000 g sample of this mineral. 
+
+          You would tell me: 
+
+          A sample of a different mineral is analysed by the same methods. This mineral also contains only Pb2+, CO3 2−, OH− and O2− ions.
+            When a 5.000 g sample of this mineral is treated with 25.00 mL of 2.000 mol L−1 nitric acid
+            (HNO3), 0.5214 g of carbon dioxide is released, and 0.01051 mol of the acid remains.
+            When subjected to thermal decomposition, 5.000 g of this mineral loses 0.5926 g. Calculate the chemical amount (in mol or mmol) of nitric acid that reacts with the
+            5.000 g sample of this mineral. 
 
           Here is the question: {{extracted_question}}
           Here is the extracted text: {{output_extraction}}
-
-          Include only the new question with its context. 
           """
         output_column: improved_question
       model: gpt-4o-mini