Upload folder using huggingface_hub

atomic_concepts_gradio/20250124_164753_angelone.json

@@ -0,0 +1,28 @@
+{
+    "metrics_prompt": "Forget all of our past conversations and treat this as a new conversation.  As an expert Data Analyst in the Financial Securities Trading Industry,  parse the following question \"{question}\" and list all the metrics mentioned in the statement.  Do not list anything that is not a metric.  Do not include dimensions. ",
+    "dimensions_prompt": "Forget all of our past conversations and treat this as a new conversation.  As an expert Data Analyst in the Financial Securities Trading Industry,  parse the following question \"{question}\" and identify all the dimensions that are not a metric.  List the dimensions including time period and list that only and nothing else.  Do not include metrics.",
+    "results": [
+        {
+            "question": "What is the MTD total order volume",
+            "metrics": [
+                "MTD total order volume"
+            ],
+            "dimensions": [
+                "MTD",
+                "total order volume"
+            ]
+        },
+        {
+            "question": "Show total monthly order volume for the past 12 months.",
+            "metrics": [
+                "total monthly order volume"
+            ],
+            "dimensions": [
+                "order",
+                "time period (past 12 months)"
+            ]
+        }
+    ],
+    "llm_model": "gpt-4o",
+    "username": "achinta"
+}

atomic_concepts_gradio/20250124_164947_angelone.json

@@ -0,0 +1,28 @@
+{
+    "metrics_prompt": "Forget all of our past conversations and treat this as a new conversation.  As an expert Data Analyst in the Financial Securities Trading Industry,  parse the following question \"{question}\" and list all the metrics mentioned in the statement.  Do not list anything that is not a metric.  Do not include dimensions. ",
+    "dimensions_prompt": "Forget all of our past conversations and treat this as a new conversation.  As an expert Data Analyst in the Financial Securities Trading Industry,  parse the following question \"{question}\" and identify all the dimensions that are not a metric.  List the dimensions including time period and list that only and nothing else.  Do not include metrics.",
+    "results": [
+        {
+            "question": "What is the MTD total order volume",
+            "metrics": [
+                "MTD total order volume"
+            ],
+            "dimensions": [
+                "MTD (Month-to-Date)",
+                "order"
+            ]
+        },
+        {
+            "question": "Show total monthly order volume for the past 12 months.",
+            "metrics": [
+                "total monthly order volume"
+            ],
+            "dimensions": [
+                "monthly",
+                "past 12 months"
+            ]
+        }
+    ],
+    "llm_model": "gpt-4o",
+    "username": "achinta"
+}

evaluate_atomic_matching.py

@@ -29,7 +29,7 @@ def process_questions(llm_model: str, openai_api_Key: str, deploy_key: str, metr
     if not questions: missing_fields.append("Questions")
     
     if missing_fields:
-        return f"Error: The following fields are mandatory and missing: {', '.join(missing_fields)}"
+        return f"Error: The following fields are mandatory and missing: {', '.join(missing_fields)}", None
     
     # Set API key for LiteLLM
     os.environ["OPENAI_API_KEY"] = openai_api_Key
@@ -117,12 +117,13 @@ def process_questions(llm_model: str, openai_api_Key: str, deploy_key: str, metr
             # Clone into temporary directory
             repo = Repo.clone_from(repo_url, temp_dir)
             
-            # Copy the new file to the cloned repo
-            import shutil
-            shutil.copy2(os.path.join(script_dir, filename), os.path.join(temp_dir, filename))
+            # Copy the new file to the cloned repo under atomic_concepts_gradio directory
+            repo_subdir = os.path.join(temp_dir, "atomic_concepts_gradio")
+            os.makedirs(repo_subdir, exist_ok=True)
+            shutil.copy2(filepath, os.path.join(repo_subdir, filename))
             
             # Add, commit and push from temporary directory
-            repo.index.add([filename])
+            repo.index.add([os.path.join("atomic_concepts_gradio", filename)])
             repo.index.commit(f"Add evaluation results: {filename}")
             origin = repo.remote('origin')
             push_info = origin.push()
@@ -134,7 +135,7 @@ def process_questions(llm_model: str, openai_api_Key: str, deploy_key: str, metr
             # Construct GitHub web URL from SSH URL
             github_web_url = repo_url.replace("git@github.com:", "https://github.com/")
             github_web_url = github_web_url.replace(".git", "")
-            file_url = f"{github_web_url}/blob/main/{filename}"
+            file_url = f"{github_web_url}/blob/main/atomic_concepts_gradio/{filename}"
             
         finally:
             # Clean up
@@ -148,10 +149,10 @@ def process_questions(llm_model: str, openai_api_Key: str, deploy_key: str, metr
         )
         if error_messages:
             final_message += "\n\nErrors encountered:\n" + "\n".join(error_messages)
-        return final_message
+        return final_message, filepath
         
     except Exception as e:
-        return f"Error pushing to repository: {str(e)}"
+        return f"Error pushing to repository: {str(e)}", None
 
 # Create Gradio interface
 with gr.Blocks(css="footer {visibility: hidden}") as iface:
@@ -201,11 +202,31 @@ with gr.Blocks(css="footer {visibility: hidden}") as iface:
         with gr.Column(scale=1):
             questions = gr.Textbox(label="Questions (one per line)", lines=10, max_lines=10, info="Required")
             result = gr.Textbox(label="Result", lines=10, max_lines=10)
-            gr.Button("Process").click(
+            file_output = gr.State()  # Add this to store the filepath
+            
+            with gr.Row():
+                process_btn = gr.Button("Process")
+                download_btn = gr.Button("Download Results", interactive=False)
+            
+            def enable_download(message, filepath):
+                download_btn.interactive = True
+                return message, filepath
+            
+            process_btn.click(
                 fn=process_questions,
                 inputs=[model, api_key, deploy_key, metrics_prompt, dimensions_prompt, questions, customer_name],
-                outputs=result,
+                outputs=[result, file_output],
                 show_progress=True
+            ).then(
+                fn=enable_download,
+                inputs=[result, file_output],
+                outputs=[result, file_output]
+            )
+            
+            download_btn.click(
+                fn=lambda filepath: filepath,
+                inputs=[file_output],
+                outputs=gr.File(label="Download JSON")
             )