csv only submissions

app_chatgpt.py

@@ -1,15 +1,11 @@
 import openai
-import openpyxl
-import PyPDF2
 import pandas as pd
-from PIL import Image
-import pytesseract
-import io
 import os
 import gradio as gr
 import boto3
 import uuid
 
+# ✅ Load your secrets from environment
 openai.api_key = os.environ.get("Open_ai_key")
 bucket_name = os.environ.get("Bucket_name")
 
@@ -18,32 +14,28 @@ session = boto3.Session(
     aws_secret_access_key=os.getenv("secret_access_key")
 )
 
-
+# ✅ Upload reconciled results and original files to S3
 def upload_files_to_s3(erp_file_path, external_file_path, content, bucket_name):
     s3 = session.client("s3")
-
     session_id = str(uuid.uuid4())
 
-    # Create filenames with UUID
-    erp_filename = f"{os.path.splitext(os.path.basename(erp_file_path))[0]}_{session_id}{os.path.splitext(erp_file_path)[1]}"
-    bank_filename = f"{os.path.splitext(os.path.basename(external_file_path))[0]}_{session_id}{os.path.splitext(external_file_path)[1]}"
+    erp_filename = f"{os.path.splitext(os.path.basename(erp_file_path))[0]}_{session_id}.csv"
+    bank_filename = f"{os.path.splitext(os.path.basename(external_file_path))[0]}_{session_id}.csv"
     txt_filename = f"{session_id}.txt"
 
-    # Write content to a local .txt file
+    # Write reconciliation output to local file
     with open(txt_filename, "w", encoding="utf-8") as file:
         file.write(content)
 
-    # Define S3 keys (organized into folders)
+    # Define S3 keys
     erp_s3_key = f"ERP Statements/{erp_filename}"
     bank_s3_key = f"Bank Statements/{bank_filename}"
     txt_s3_key = f"Reconciliation Results/{txt_filename}"
 
-    # Upload all 3 files to S3
     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)
 
-    # Optionally clean up the local .txt file
     os.remove(txt_filename)
 
     return {
@@ -54,38 +46,22 @@ def upload_files_to_s3(erp_file_path, external_file_path, content, bucket_name):
     }
 
 
+# ✅ Extract transactions from CSV only
 def extract_transactions(file):
-    """Extract raw text content from various supported file formats."""
     filename = file.name.lower()
+    if not filename.endswith(".csv"):
+        raise ValueError(
+            f"Unsupported file format: {filename}. Please upload a CSV file only."
+        )
 
     try:
-        if filename.endswith((".xlsx", ".xls")):
-            workbook = openpyxl.load_workbook(file.name, data_only=True)
-            sheet = workbook.active
-            return "\n".join(
-                "\t".join(str(cell.value or "") for cell in row)
-                for row in sheet.iter_rows()
-            )
-
-        elif filename.endswith(".csv"):
-            df = pd.read_csv(file.name)
-            return df.to_string(index=False)
-
-        elif filename.endswith(".pdf"):
-            reader = PyPDF2.PdfReader(file)
-            return "\n".join(page.extract_text() or "" for page in reader.pages)
-
-        elif filename.endswith((".jpg", ".jpeg", ".png")):
-            image = Image.open(io.BytesIO(file.read()))
-            return pytesseract.image_to_string(image)
-
-        else:
-            raise ValueError("Unsupported file format.")
-
+        df = pd.read_csv(file.name)
+        return df.to_string(index=False)
     except Exception as e:
         raise ValueError(f"Error processing {file.name}: {e}")
 
 
+# ✅ Reconcile the statements using OpenAI
 def reconcile_statements_openai(erp_file, external_file):
     yield "⏳ Processing your request...", ""
 
@@ -120,7 +96,7 @@ Please follow this format in your response:
 Here is the ERP data:
 {erp_data}
 ---
-Here is the External (Bank or Vendor) data::
+Here is the External (Bank or Vendor) data:
 {external_data}
 """
 
@@ -144,7 +120,8 @@ Here is the External (Bank or Vendor) data::
         """
         yield "✅ Done!", html
 
-        s3_info = upload_files_to_s3(
+        # ✅ Upload the files + result to S3
+        upload_files_to_s3(
             erp_file_path=erp_file,
             external_file_path=external_file,
             content=html,
@@ -155,7 +132,7 @@ Here is the External (Bank or Vendor) data::
         yield "❌ Error occurred", f"<h3>Error</h3><pre>{e}</pre>"
 
 
-# Gradio UI
+# ✅ Gradio UI for CSV-only
 with gr.Blocks(css="""
 #company-logo {
     width: 25%;
@@ -165,18 +142,32 @@ with gr.Blocks(css="""
 """) as iface:
     gr.Image("logo_Icon.png", elem_id="company-logo", label="Beiing Human")
     gr.Markdown("## 📊 ERP vs Bank/Vendor Reconciliation Tool")
+
+    gr.Markdown(
+        "⚠️ **Important:** Please upload your statements as CSV files only. "
+        "Include columns: **Date**, **Amount**, **Description**."
+    )
+
     with gr.Row():
-        erp_file = gr.File(label="📁 Upload ERP Statement", type="filepath")
-        external_file = gr.File(label="🏦 Upload Bank or Vendor Statement", type="filepath")
+        erp_file = gr.File(
+            label="📁 Upload ERP Statement (CSV only)",
+            type="filepath",
+            file_types=[".csv"]
+        )
+        external_file = gr.File(
+            label="🏦 Upload Bank or Vendor Statement (CSV only)",
+            type="filepath",
+            file_types=[".csv"]
+        )
+
     btn = gr.Button("🔍 Reconcile")
+
     with gr.Row():
         support_btn = gr.Button(
             "📧 Contact Support",
             link="mailto:[email protected]?subject=Support%20Request&body=Hi%20Beiing%20Human%20Team%2C%0A%0AI%20have%20a%20question%20about%20the%20reconciliation%20tool.%20Here%20are%20the%20details..."
         )
-        support_email_display = gr.Markdown("📬 Or email us at: `[email protected]` (copy & paste)")
-
-    # support_btn = gr.Button("📧 Contact Support: [email protected]", link="mailto:[email protected]?subject=Support%20Request&body=Hi%20Beiing%20Human%20Team%2C%0A%0AI%20have%20a%20question%20about%20the%20reconciliation%20tool.%20Here%20are%20the%20details...")
+        gr.Markdown("📬 Or email us at: `[email protected]` (copy & paste)")
 
     status = gr.Markdown()
     result = gr.HTML()
@@ -187,4 +178,4 @@ with gr.Blocks(css="""
         outputs=[status, result]
     )
 
-iface.launch()
+iface.launch()