Update app_chatgpt.py

app_chatgpt.py

@@ -4,85 +4,172 @@ import os
 import gradio as gr
 import boto3
 import uuid
+import logging
+import sys
+from datetime import datetime
+
+logging.basicConfig(
+    level=logging.INFO,
+    format='%(asctime)s - %(name)s - %(levelname)s - %(message)s',
+    handlers=[
+        logging.StreamHandler(sys.stdout),  # This will show in HF Spaces logs
+        logging.FileHandler('app.log', mode='a')  # Also save to file
+    ]
+)
 
-# ✅ Load your secrets from environment
-openai.api_key = os.environ.get("Open_ai_key")
-bucket_name = os.environ.get("Bucket_name")
+logger = logging.getLogger(__name__)
 
-session = boto3.Session(
-    aws_access_key_id=os.getenv("access_key"),
-    aws_secret_access_key=os.getenv("secret_access_key")
-)
+# Log application startup
+logger.info("🚀 Starting ERP Reconciliation Tool")
+logger.info(f"📅 Application started at: {datetime.now()}")
 
-def upload_files_to_s3(erp_file_path, external_file_path, content, bucket_name):
-    s3 = session.client("s3")
-    session_id = str(uuid.uuid4())
 
-    base_erp = os.path.splitext(os.path.basename(erp_file_path))[0]
-    base_bank = os.path.splitext(os.path.basename(external_file_path))[0]
+try:
+    openai.api_key = os.environ.get("Open_ai_key")
+    bucket_name = os.environ.get("Bucket_name")
 
-    # Add session ID to filenames
-    erp_filename = f"{base_erp}_{session_id}.csv"
-    bank_filename = f"{base_bank}_{session_id}.csv"
-    txt_filename = f"{session_id}.txt"
+    if not openai.api_key:
+        logger.warning("⚠️ OpenAI API key not found in environment variables")
+    else:
+        logger.info("✅ OpenAI API key loaded successfully")
 
-    # Write reconciliation output to local file
-    with open(txt_filename, "w", encoding="utf-8") as file:
-        file.write(content)
+    if not bucket_name:
+        logger.warning("⚠️ S3 bucket name not found in environment variables")
+    else:
+        logger.info(f"✅ S3 bucket configured: {bucket_name}")
 
-    # === Separate folders by type ===
-    erp_s3_key = f"ERP Statements/{erp_filename}"
-    bank_s3_key = f"Bank Statements/{bank_filename}"
-    txt_s3_key = f"Reconciliation Results/{txt_filename}"
+except Exception as e:
+    logger.error(f"❌ Error loading environment variables: {e}")
 
-    # === Combined folder for each run ===
-    combined_prefix = f"Combined Files/{session_id}/"
-    erp_combined_key = combined_prefix + erp_filename
-    bank_combined_key = combined_prefix + bank_filename
-    txt_combined_key = combined_prefix + txt_filename
+try:
+    session = boto3.Session(
+        aws_access_key_id=os.getenv("access_key"),
+        aws_secret_access_key=os.getenv("secret_access_key")
+    )
+    logger.info("✅ AWS session created successfully")
+except Exception as e:
+    logger.error(f"❌ Error creating AWS session: {e}")
 
-    # ✅ Upload to type-based folders
-    s3.upload_file(erp_file_path, bucket_name, erp_s3_key)
-    s3.upload_file(external_file_path, bucket_name, bank_s3_key)
-    s3.upload_file(txt_filename, bucket_name, txt_s3_key)
 
-    # ✅ Upload to combined folder
-    s3.upload_file(erp_file_path, bucket_name, erp_combined_key)
-    s3.upload_file(external_file_path, bucket_name, bank_combined_key)
-    s3.upload_file(txt_filename, bucket_name, txt_combined_key)
+def upload_files_to_s3(erp_file_path, external_file_path, content, bucket_name):
+    logger.info(f"📤 Starting S3 upload process")
 
-    os.remove(txt_filename)
+    try:
+        s3 = session.client("s3")
+        session_id = str(uuid.uuid4())
+
+        logger.info(f"🆔 Generated session ID: {session_id}")
+
+        base_erp = os.path.splitext(os.path.basename(erp_file_path))[0]
+        base_bank = os.path.splitext(os.path.basename(external_file_path))[0]
+
+        # Add session ID to filenames
+        erp_filename = f"{base_erp}_{session_id}.csv"
+        bank_filename = f"{base_bank}_{session_id}.csv"
+        txt_filename = f"{session_id}.txt"
+
+        logger.info(f"📄 Generated filenames - ERP: {erp_filename}, Bank: {bank_filename}, Result: {txt_filename}")
+
+        # Write reconciliation output to local file
+        with open(txt_filename, "w", encoding="utf-8") as file:
+            file.write(content)
+        logger.info(f"✅ Reconciliation result written to local file: {txt_filename}")
+
+        # === Separate folders by type ===
+        erp_s3_key = f"ERP Statements/{erp_filename}"
+        bank_s3_key = f"Bank Statements/{bank_filename}"
+        txt_s3_key = f"Reconciliation Results/{txt_filename}"
+
+        # === Combined folder for each run ===
+        combined_prefix = f"Combined Files/{session_id}/"
+        erp_combined_key = combined_prefix + erp_filename
+        bank_combined_key = combined_prefix + bank_filename
+        txt_combined_key = combined_prefix + txt_filename
+
+        # ✅ Upload to type-based folders
+        logger.info("📤 Uploading files to type-based S3 folders...")
+        s3.upload_file(erp_file_path, bucket_name, erp_s3_key)
+        logger.info(f"✅ Uploaded ERP file to: {erp_s3_key}")
+
+        s3.upload_file(external_file_path, bucket_name, bank_s3_key)
+        logger.info(f"✅ Uploaded Bank file to: {bank_s3_key}")
+
+        s3.upload_file(txt_filename, bucket_name, txt_s3_key)
+        logger.info(f"✅ Uploaded result file to: {txt_s3_key}")
+
+        # ✅ Upload to combined folder
+        logger.info("📤 Uploading files to combined S3 folder...")
+        s3.upload_file(erp_file_path, bucket_name, erp_combined_key)
+        s3.upload_file(external_file_path, bucket_name, bank_combined_key)
+        s3.upload_file(txt_filename, bucket_name, txt_combined_key)
+        logger.info(f"✅ All files uploaded to combined folder: {combined_prefix}")
+
+        # Clean up local file
+        os.remove(txt_filename)
+        logger.info(f"🗑️ Cleaned up local file: {txt_filename}")
+
+        upload_result = {
+            "erp_s3_key": erp_s3_key,
+            "bank_s3_key": bank_s3_key,
+            "result_s3_key": txt_s3_key,
+            "combined_keys": [erp_combined_key, bank_combined_key, txt_combined_key],
+            "session_id": session_id
+        }
+
+        logger.info(f"✅ S3 upload process completed successfully for session: {session_id}")
+        return upload_result
 
-    return {
-        "erp_s3_key": erp_s3_key,
-        "bank_s3_key": bank_s3_key,
-        "result_s3_key": txt_s3_key,
-        "combined_keys": [erp_combined_key, bank_combined_key, txt_combined_key],
-        "session_id": session_id
-    }
+    except Exception as e:
+        logger.error(f"❌ Error in S3 upload process: {e}")
+        raise e
 
 
 # ✅ Extract transactions from CSV only
 def extract_transactions(file):
+    logger.info(f"📊 Processing file: {file.name}")
+
     filename = file.name.lower()
     if not filename.endswith(".csv"):
-        raise ValueError(
-            f"Unsupported file format: {filename}. Please upload a CSV file only."
-        )
+        error_msg = f"Unsupported file format: {filename}. Please upload a CSV file only."
+        logger.error(f"❌ {error_msg}")
+        raise ValueError(error_msg)
 
     try:
         df = pd.read_csv(file.name)
+        logger.info(f"✅ Successfully read CSV file: {file.name}")
+        logger.info(f"📈 File contains {len(df)} rows and {len(df.columns)} columns")
+        logger.info(f"🏷️ Column names: {list(df.columns)}")
+
         return df.to_string(index=False)
     except Exception as e:
-        raise ValueError(f"Error processing {file.name}: {e}")
+        error_msg = f"Error processing {file.name}: {e}"
+        logger.error(f"❌ {error_msg}")
+        raise ValueError(error_msg)
 
 
 # ✅ Reconcile the statements using OpenAI
 def reconcile_statements_openai(erp_file, external_file):
+    session_start = datetime.now()
+    logger.info(f"🔍 Starting reconciliation process at: {session_start}")
+
+    # Validate inputs
+    if not erp_file or not external_file:
+        error_msg = "Both ERP and External files are required"
+        logger.error(f"❌ {error_msg}")
+        yield "❌ Error: Missing files", f"<h3>Error</h3><pre>{error_msg}</pre>"
+        return
+
+    logger.info(f"📁 ERP file: {erp_file.name if erp_file else 'None'}")
+    logger.info(f"🏦 External file: {external_file.name if external_file else 'None'}")
+
     yield "⏳ Processing your request...", ""
 
     try:
+        # Extract data from files
+        logger.info("📊 Extracting data from ERP file...")
         erp_data = extract_transactions(erp_file)
+
+        logger.info("📊 Extracting data from External file...")
         external_data = extract_transactions(external_file)
 
         prompt = f"""
@@ -116,6 +203,9 @@ Here is the External (Bank or Vendor) data:
 {external_data}
 """
 
+        logger.info("🤖 Sending request to OpenAI...")
+        logger.info(f"📝 Prompt length: {len(prompt)} characters")
+
         response = openai.ChatCompletion.create(
             model="gpt-5",
             messages=[
@@ -126,7 +216,10 @@ Here is the External (Bank or Vendor) data:
             temperature=1,
         )
 
+        logger.info("✅ Received response from OpenAI")
+
         content = response.choices[0].message['content']
+        logger.info(f"📄 Response length: {len(content)} characters")
 
         html = f"""
         <div style="font-family: 'Segoe UI', sans-serif; line-height: 1.5;">
@@ -134,21 +227,40 @@ Here is the External (Bank or Vendor) data:
             <pre>{content}</pre>
         </div>
         """
+
+        # Calculate processing time
+        processing_time = (datetime.now() - session_start).total_seconds()
+        logger.info(f"⏱️ Processing completed in {processing_time:.2f} seconds")
+
         yield "✅ Done!", html
 
         # ✅ Upload the files + result to S3
-        upload_files_to_s3(
-            erp_file_path=erp_file,
-            external_file_path=external_file,
-            content=html,
-            bucket_name=bucket_name
-        )
+        try:
+            logger.info("📤 Starting S3 upload...")
+            upload_result = upload_files_to_s3(
+                erp_file_path=erp_file.name,
+                external_file_path=external_file.name,
+                content=html,
+                bucket_name=bucket_name
+            )
+            logger.info(f"✅ Files uploaded successfully with session ID: {upload_result['session_id']}")
+        except Exception as upload_error:
+            logger.error(f"❌ S3 upload failed: {upload_error}")
+            # Don't fail the entire process if S3 upload fails
+            logger.info("⚠️ Continuing despite S3 upload failure")
+
+        logger.info(f"🎉 Reconciliation process completed successfully!")
 
     except Exception as e:
-        yield "❌ Error occurred", f"<h3>Error</h3><pre>{e}</pre>"
+        error_msg = f"Error during reconciliation: {e}"
+        logger.error(f"❌ {error_msg}")
+        logger.error(f"🐛 Full error details: {str(e)}")
+        yield "❌ Error occurred", f"<h3>Error</h3><pre>{error_msg}</pre>"
 
 
 # ✅ Gradio UI for CSV-only
+logger.info("🎨 Building Gradio interface...")
+
 with gr.Blocks(css="""
 #company-logo {
     width: 25%;
@@ -193,5 +305,6 @@ with gr.Blocks(css="""
         inputs=[erp_file, external_file],
         outputs=[status, result]
     )
-
-iface.launch()
+iface.launch()
+# Log when interface is ready
+logger.info("✅ Gradio interface built successfully")