Update app.py

app.py

@@ -1,11 +1,58 @@
+import os
+import json
 import gradio as gr
 import pandas as pd
 
-def readCSV(file):
-    df = pd.read_csv(file)
-    return "All good!"
+def read_csv(csv_file):
+    df = pd.read_csv(csv_file)
+    json_output = convert_df_to_json(df, os.path.splitext(os.path.basename(csv_file))[0])
+    return generate_prompt(json_output)
 
-demo = gr.Interface(fn=readCSV, inputs="text", outputs="text")
+def generate_prompt(json_output):
+    preamble = """
+    Given below is the data found in a relational database table in JSON format:
+
+    """
+
+    mid = """
+
+    For each column, tell me whether the data is of one of the following types: """
+
+    end = """
+
+    Your output should be in the following format:
+
+    { "tableName": "table_name",
+        "columns": [
+            {"columnName": "column1", "columnType": “one of the types given above”},
+            {"columnName": "column2", "columnType": “one of the types given above”},
+   	        …
+        ]
+    }
+
+    """
+
+    prompt = preamble + json.dumps(json_output) + mid + types_str + end
+    return prompt
+
+def convert_df_to_json(df, table_name):
+    json_output = {
+        "tableName": table_name,
+        "columns": []
+    }
+
+    for column in df.columns:
+        column_info = {
+            "columnName": column,
+            "exampleValues": df[column].tolist()
+        }
+        json_output["columns"].append(column_info)
+
+    return json_output
+
+known_types = ["NAME", "ADDRESS", "TELEPHONE NUMBER", "SOCIAL SECURITY NUMBER", "CREDIT CARD NUMBER" , "UNKNOWN"]
+types_str = ', '.join(map(str, known_types))
+demo = gr.Interface(fn=read_csv, inputs="text", outputs="text")
     
 if __name__ == "__main__":
-    demo.launch(show_api=False, debug=True)
+    demo.launch(show_api=False, debug=True)