Update script_for_automation.py

script_for_automation.py

@@ -299,7 +299,70 @@ def get_data_ready(recipe_dict, input_data_piece):
 
     return processed_data
         
+def generate_markdown_output(df):
+    # Start the markdown output string
+    markdown = ""
+
+    # Table for Recipe Fields (recipe_id to prompting_strategy)
+    markdown += "### Recipe Fields (Basic Information)\n"
+    markdown += "| Recipe ID | Testing Strategy | Schema Processing Model | Pre-Processing Strategy | Pre-Processing Text | Pre-Processing Model | Prompting Strategy |\n"
+    markdown += "|-----------|------------------|-------------------------|--------------------------|---------------------|----------------------|-------------------|\n"
+
+    # Iterate over rows to create the first table (recipe_id to prompting_strategy)
+    for _, row in df.iterrows():
+        markdown += f"| {row['Recipe_ID']} | {row['Testing_Strategy_Text']} | {row['Schema_Processing_Model']} | {row['Pre_Processing_Strategy']} | {row['Pre_Processing_Text']} | {row['Pre_Processing_Model']} | {row['Prompting_Strategy']} |\n"
+
+    # Separate section for Prompts
+    markdown += "\n### Prompts\n"
+    markdown += "| Plantings and Fields Prompt | Interactions Prompt | Treatments Prompt |\n"
+    markdown += "|-----------------------------|---------------------|-------------------|\n"
+
+    # Iterate over rows to create the second table (plantings_and_fields_prompt, interactions_prompt, treatments_prompt)
+    for _, row in df.iterrows():
+        markdown += f"| {row['Plantings_and_Fields_Prompt']} | {row['Interactions_Prompt']} | {row['Treatments_Prompt']} |\n"
+
+    # Separate section for Input Transcript
+    markdown += "\n### Input Transcript\n"
+    markdown += "Since the input transcript might be very long, it is truncated here for readability:\n"
     
+    # Display a truncated version of the input transcript to avoid long text in the table
+    for _, row in df.iterrows():
+        truncated_input = (row['Input_Transcript'][:500] + '...') if len(row['Input_Transcript']) > 500 else row['Input_Transcript']
+        markdown += f"**Recipe ID {row['Recipe_ID']}**: {truncated_input}\n\n"
+
+    # Side-by-side comparison of Gold Standard and Machine Generated Key-Values
+    markdown += "\n### Gold Standard vs Machine Generated Key-Values\n"
+    markdown += "| Key | Gold Standard | Machine Generated |\n"
+    markdown += "|-----|---------------|-------------------|\n"
+
+    # Iterate over rows to create the comparison table
+    for _, row in df.iterrows():
+        markdown += f"| {row['Recipe_ID']} | {row['Gold_Standard_Key_Values']} | {row['Machine_Generated_Key_Values']} |\n"
+
+    # Display differences in a readable format
+    markdown += "\n### Differences\n"
+    markdown += "The following differences were found between the gold standard and the machine-generated output:\n"
+    markdown += "| Key | Difference |\n"
+    markdown += "|-----|------------|\n"
+
+    for _, row in df.iterrows():
+        # Assuming 'Differences' is a list of dictionaries with keys and changes
+        differences = json.loads(row['Differences'])
+        for diff in differences:
+            if diff.get("values_changed"):
+                for change in diff["values_changed"]:
+                    markdown += f"| {change['path']} | {change['old_value']} -> {change['new_value']} |\n"
+
+    # Side-by-side YAML comparison for human visual inspection
+    markdown += "\n### Gold Standard vs Machine Generated YAML\n"
+    markdown += "| Gold Standard YAML | Machine Generated YAML |\n"
+    markdown += "|--------------------|------------------------|\n"
+
+    # Add the side-by-side YAML comparison
+    for _, row in df.iterrows():
+        markdown += f"| ```yaml\n{row['Gold_Standard_YAML']}\n``` | ```yaml\n{row['Machine_Generated_YAML']}\n``` |\n"
+
+    return markdown
     
     
 def drive_process():