billjson

bill copy.py

@@ -0,0 +1,216 @@
+#!/usr/bin/env -S poetry run python
+
+import os
+import json
+import pdfplumber
+import streamlit as st
+from openai import OpenAI
+
+client = OpenAI()
+
+def load_user_data(user_id):
+    file_path = os.path.join("data", "user_data", f"user_data_{user_id}.json")
+    if not os.path.exists(file_path):
+        return {}
+    with open(file_path, "r") as file:
+        return json.load(file)
+
+def parse_pdf_to_json(pdf_path):
+    user_id = {}
+    serie_factura = {}
+    data_factura = {}
+    costuri = {}
+    with pdfplumber.open(pdf_path, ) as pdf:
+        for page in pdf.pages:
+            text = page.extract_text()
+            if text:
+                lines = text.split('\n')
+
+                # Process each line and look for specific categories
+                for line in lines:
+                    # Check for 'Data emiterii facturii'
+                    if 'Data facturii' in line:
+                        date = line.split()[-1]
+                        data_factura['Data factura'] = date
+
+                    # Check for 'Serie factură'
+                    if 'rul facturii:' in line:
+                        serie = line.split()[-1]
+                        serie_factura['Serie numar'] = serie
+                        
+                    # Check for 'Cont client'
+                    if 'Cont client' in line:
+                        cont = line.split()[-1]
+                        user_id['Cont client'] = cont
+                    
+                    # Check for 'Valoare facturată fără TVA'
+                    if 'Sold precedent' in line:
+                        value = line.split()[-2].replace(',', '.')  # Extract and convert to float
+                        costuri['Sold precedent'] = value
+                    
+                    # Check for 'Total bază de impozitare TVA'
+                    elif 'din sold precedent' in line:
+                        value = line.split()[-2].replace(',', '.')  # Extract and convert to float
+                        costuri['Total platit din sold precedent'] = value
+                    
+                    # Check for 'TVA'
+                    elif 'TVA' in line and '%' in line:
+                        value = line.split()[-2].replace(',', '.')  # Extract and convert to float
+                        costuri['TVA'] = value
+                    
+                    # Check for 'Dobânzi penalizatoare'
+                    elif 'Abonamente' in line:
+                        value = line.split()[-2].replace(',', '.')  # Extract and convert to float
+                        costuri['Abonamente si extraopiuni'] = value
+                    
+                    # Check for 'TOTAL DE PLATĂ FACTURĂ CURENTĂ'
+                    elif 'Total factura curenta fara TVA' in line:
+                        value = float(line.split()[-2].replace(',', '.'))  # Extract and convert to float
+                        costuri['Total factura curenta fara TVA'] = value
+                    
+                    # Check for 'Sold Cont Contract'
+                    elif 'Servicii utilizate' in line:
+                        value = line.split()[-2].replace(',', '.')  # Extract and convert to float
+                        costuri['Servicii utilizate'] = value
+                    
+                    # Check for 'Compensatii'
+                    elif 'Rate terminal' in line:
+                        value = float(line.split()[-2].replace(',', '.'))  # Extract and convert to float
+                        costuri['Rate terminal'] = value
+
+                    # Check for 'TVA 19,00%'
+                    elif 'TVA 19,00%' in line:
+                        value = float(line.split()[-2].replace(',', '.'))  # Extract and convert to float
+                        costuri['TVA'] = value
+
+                    # Check for 'Compensatii'
+                    elif 'Total factura curenta' in line:
+                        value = float(line.split()[-2].replace(',', '.'))  # Extract and convert to float
+                        costuri['Total factura curenta'] = value
+
+    return costuri
+
+def check_related_keys(question, user_id):
+    user_data = load_user_data(user_id)
+    bill_keys = set()
+    for bill in user_data.get("bills", []):
+        bill_keys.update(bill.keys())
+    return [key for key in bill_keys if key.lower() in question.lower()]
+
+def process_query(query, user_id):
+    user_data = load_user_data(user_id)
+    bill_info = user_data.get("bills", [])
+    related_keys = check_related_keys(query, user_id)
+    related_keys_str = ", ".join(related_keys) if related_keys else "N/A"
+
+    if related_keys_str != "N/A":
+        context = (
+            f"Citeste informatiile despre costrurile in lei facturate din dictionar: {bill_info} "
+            f"si raspunde la intrebarea: '{query}' dar numai cu info legate de: {related_keys_str}"
+        )
+    else:
+        context = (
+            f"Citeste informatiile despre costrurile in lei facturate din dictionar: {bill_info} "
+            f"si raspunde la intrebarea: '{query}' dar numai cu info legate de factura"
+        )
+
+    max_input_length = 550
+    st.write(f"Context:\n{context}")
+    st.write(f"Context size: {len(context)} characters")
+
+    if len(context) > max_input_length:
+        st.warning("Prea multe caractere în context, solicitarea nu va fi trimisă.")
+        return None
+
+    return context
+
+def main():
+
+    st.title("Telecom Bill Chat with LLM Agent")
+
+    if "user_id" not in st.session_state:
+        st.session_state.user_id = None
+
+    user_id = st.sidebar.text_input("Introdu numărul de telefon:")
+    if user_id and user_id != st.session_state.user_id:
+        data = load_user_data(user_id)
+        if data:
+            st.session_state.user_id = user_id
+            st.success("Utilizator găsit!")
+        else:
+            st.warning("Nu am găsit date pentru acest ID. Încărcați o factură PDF la nevoie.")
+            st.session_state.user_id = user_id
+
+    uploaded_file = st.file_uploader("Încarcă factura PDF", type="pdf")
+    if uploaded_file and st.session_state.user_id:
+        bill_data = parse_pdf_to_json(uploaded_file)
+        existing_data = load_user_data(st.session_state.user_id)
+        if "bills" not in existing_data:
+            existing_data["bills"] = []
+        existing_data["bills"].append(bill_data)
+        file_path = os.path.join("data", "user_data", f"user_data_{st.session_state['user_id']}.json")
+        os.makedirs(os.path.dirname(file_path), exist_ok=True)
+        with open(file_path, "w") as file:
+            json.dump(existing_data, file)
+        st.success("Factura a fost încărcată și salvată cu succes!")
+
+    if st.session_state.user_id:
+        data = load_user_data(st.session_state.user_id)
+        st.write(f"Phone Number: {st.session_state.user_id}")
+        st.write("Facturi existente:")
+        for bill in data.get("bills", []):
+            st.write(bill)
+    else:
+        st.info("Introduceți un ID și/sau încărcați o factură PDF pentru a continua.")
+
+    # Initialize conversation in the session state
+    # "context_prompt_added" indicates whether we've added the specialized "bill info" context yet.
+    if "messages" not in st.session_state:
+        st.session_state["messages"] = [
+            {"role": "assistant", "content": "Cu ce te pot ajuta?"}
+        ]
+    if "context_prompt_added" not in st.session_state:
+        st.session_state.context_prompt_added = False
+
+    st.write("---")
+    st.subheader("Chat")
+
+    for msg in st.session_state["messages"]:
+        st.chat_message(msg["role"]).write(msg["content"])
+
+    if prompt := st.chat_input("Introduceți întrebarea aici:"):
+        if not st.session_state.user_id:
+            st.error("Trebuie să introduceți un număr de telefon valid sau să încărcați date.")
+            return
+
+        # If the context prompt hasn't been added yet, build & inject it once;
+        # otherwise, just add the user's raw question.
+        if not st.session_state.context_prompt_added:
+            final_prompt = process_query(prompt, st.session_state["user_id"])
+            if final_prompt is None:
+                st.stop()
+            st.session_state["messages"].append({"role": "user", "content": final_prompt})
+            st.session_state.context_prompt_added = True
+        else:
+            st.session_state["messages"].append({"role": "user", "content": prompt})
+
+        # Display the latest user message in the chat
+        st.chat_message("user").write(st.session_state["messages"][-1]["content"])
+
+        # Now call GPT-4 with the entire conversation
+        completion = client.chat.completions.create(
+            model="gpt-4",
+            messages=st.session_state["messages"]
+        )
+        response_text = completion.choices[0].message.content.strip()
+
+        st.session_state["messages"].append({"role": "assistant", "content": response_text})
+        st.chat_message("assistant").write(response_text)
+
+        if hasattr(completion, "usage"):
+            st.write("Prompt tokens:", completion.usage.prompt_tokens)
+            st.write("Completion tokens:", completion.usage.completion_tokens)
+            st.write("Total tokens:", completion.usage.total_tokens)
+
+if __name__ == "__main__":
+    main()

bill.py

@@ -2,7 +2,6 @@
 
 import os
 import json
-import pdfplumber
 import streamlit as st
 from openai import OpenAI
 
@@ -15,80 +14,35 @@ def load_user_data(user_id):
     with open(file_path, "r") as file:
         return json.load(file)
 
-def parse_pdf_to_json(pdf_path):
-    user_id = {}
-    serie_factura = {}
-    data_factura = {}
-    costuri = {}
-    with pdfplumber.open(pdf_path, ) as pdf:
-        for page in pdf.pages:
-            text = page.extract_text()
-            if text:
-                lines = text.split('\n')
-
-                # Process each line and look for specific categories
-                for line in lines:
-                    # Check for 'Data emiterii facturii'
-                    if 'Data facturii' in line:
-                        date = line.split()[-1]
-                        data_factura['Data factura'] = date
-
-                    # Check for 'Serie factură'
-                    if 'rul facturii:' in line:
-                        serie = line.split()[-1]
-                        serie_factura['Serie numar'] = serie
-                        
-                    # Check for 'Cont client'
-                    if 'Cont client' in line:
-                        cont = line.split()[-1]
-                        user_id['Cont client'] = cont
-                    
-                    # Check for 'Valoare facturată fără TVA'
-                    if 'Sold precedent' in line:
-                        value = line.split()[-2].replace(',', '.')  # Extract and convert to float
-                        costuri['Sold precedent'] = value
-                    
-                    # Check for 'Total bază de impozitare TVA'
-                    elif 'din sold precedent' in line:
-                        value = line.split()[-2].replace(',', '.')  # Extract and convert to float
-                        costuri['Total platit din sold precedent'] = value
-                    
-                    # Check for 'TVA'
-                    elif 'TVA' in line and '%' in line:
-                        value = line.split()[-2].replace(',', '.')  # Extract and convert to float
-                        costuri['TVA'] = value
-                    
-                    # Check for 'Dobânzi penalizatoare'
-                    elif 'Abonamente' in line:
-                        value = line.split()[-2].replace(',', '.')  # Extract and convert to float
-                        costuri['Abonamente si extraopiuni'] = value
-                    
-                    # Check for 'TOTAL DE PLATĂ FACTURĂ CURENTĂ'
-                    elif 'Total factura curenta fara TVA' in line:
-                        value = float(line.split()[-2].replace(',', '.'))  # Extract and convert to float
-                        costuri['Total factura curenta fara TVA'] = value
-                    
-                    # Check for 'Sold Cont Contract'
-                    elif 'Servicii utilizate' in line:
-                        value = line.split()[-2].replace(',', '.')  # Extract and convert to float
-                        costuri['Servicii utilizate'] = value
-                    
-                    # Check for 'Compensatii'
-                    elif 'Rate terminal' in line:
-                        value = float(line.split()[-2].replace(',', '.'))  # Extract and convert to float
-                        costuri['Rate terminal'] = value
-
-                    # Check for 'TVA 19,00%'
-                    elif 'TVA 19,00%' in line:
-                        value = float(line.split()[-2].replace(',', '.'))  # Extract and convert to float
-                        costuri['TVA'] = value
-
-                    # Check for 'Compensatii'
-                    elif 'Total factura curenta' in line:
-                        value = float(line.split()[-2].replace(',', '.'))  # Extract and convert to float
-                        costuri['Total factura curenta'] = value
-
-    return costuri
+def parseBill(data):
+    billDate = data.get("billDate")
+    billNo = data.get("billNo")
+    amountDue = data.get("amountDue")
+    extraCharge = data.get("extraCharge")
+    taxItems = data.get("taxItem", [])
+    subscribers = data.get("subscribers", [])
+
+    totalBillCosts = [{"categorie": t.get("cat"), "amount": t.get("amt")} for t in taxItems]
+    subscriberCosts = []
+    for sub in subscribers:
+        logicalResource = sub.get("logicalResource")
+        billSummaryItems = sub.get("billSummaryItem", [])
+        subscriberCosts.append({
+            "logicalResource": logicalResource,
+            "billSummaryItems": [
+                {"categorie": bsi.get("cat"), "amount": bsi.get("amt"), "name": bsi.get("name")}
+                for bsi in billSummaryItems
+            ],
+        })
+
+    return {
+        "billDate": billDate,
+        "billNo": billNo,
+        "amountDue": amountDue,
+        "extraCharge": extraCharge,
+        "totalBillCosts": totalBillCosts,
+        "subscriberCosts": subscriberCosts
+    }
 
 def check_related_keys(question, user_id):
     user_data = load_user_data(user_id)
@@ -141,13 +95,14 @@ def main():
             st.warning("Nu am găsit date pentru acest ID. Încărcați o factură PDF la nevoie.")
             st.session_state.user_id = user_id
 
-    uploaded_file = st.file_uploader("Încarcă factura PDF", type="pdf")
+    uploaded_file = st.file_uploader("Încarcă factura JSON", type="json")
     if uploaded_file and st.session_state.user_id:
-        bill_data = parse_pdf_to_json(uploaded_file)
+        bill_data = json.load(uploaded_file)
+        parsed_bill = parseBill(bill_data)
         existing_data = load_user_data(st.session_state.user_id)
         if "bills" not in existing_data:
             existing_data["bills"] = []
-        existing_data["bills"].append(bill_data)
+        existing_data["bills"].append(parsed_bill)
         file_path = os.path.join("data", "user_data", f"user_data_{st.session_state['user_id']}.json")
         os.makedirs(os.path.dirname(file_path), exist_ok=True)
         with open(file_path, "w") as file:
@@ -161,7 +116,7 @@ def main():
         for bill in data.get("bills", []):
             st.write(bill)
     else:
-        st.info("Introduceți un ID și/sau încărcați o factură PDF pentru a continua.")
+        st.info("Introduceți un ID și/sau încărcați o factură JSON pentru a continua.")
 
     # Initialize conversation in the session state
     # "context_prompt_added" indicates whether we've added the specialized "bill info" context yet.

billPdf.py

@@ -0,0 +1,216 @@
+#!/usr/bin/env -S poetry run python
+
+import os
+import json
+#import pdfplumber
+import streamlit as st
+from openai import OpenAI
+
+client = OpenAI()
+
+def load_user_data(user_id):
+    file_path = os.path.join("data", "user_data", f"user_data_{user_id}.json")
+    if not os.path.exists(file_path):
+        return {}
+    with open(file_path, "r") as file:
+        return json.load(file)
+
+def parse_pdf_to_json(pdf_path):
+    user_id = {}
+    serie_factura = {}
+    data_factura = {}
+    costuri = {}
+    with pdfplumber.open(pdf_path, ) as pdf:
+        for page in pdf.pages:
+            text = page.extract_text()
+            if text:
+                lines = text.split('\n')
+
+                # Process each line and look for specific categories
+                for line in lines:
+                    # Check for 'Data emiterii facturii'
+                    if 'Data facturii' in line:
+                        date = line.split()[-1]
+                        data_factura['Data factura'] = date
+
+                    # Check for 'Serie factură'
+                    if 'rul facturii:' in line:
+                        serie = line.split()[-1]
+                        serie_factura['Serie numar'] = serie
+                        
+                    # Check for 'Cont client'
+                    if 'Cont client' in line:
+                        cont = line.split()[-1]
+                        user_id['Cont client'] = cont
+                    
+                    # Check for 'Valoare facturată fără TVA'
+                    if 'Sold precedent' in line:
+                        value = line.split()[-2].replace(',', '.')  # Extract and convert to float
+                        costuri['Sold precedent'] = value
+                    
+                    # Check for 'Total bază de impozitare TVA'
+                    elif 'din sold precedent' in line:
+                        value = line.split()[-2].replace(',', '.')  # Extract and convert to float
+                        costuri['Total platit din sold precedent'] = value
+                    
+                    # Check for 'TVA'
+                    elif 'TVA' in line and '%' in line:
+                        value = line.split()[-2].replace(',', '.')  # Extract and convert to float
+                        costuri['TVA'] = value
+                    
+                    # Check for 'Dobânzi penalizatoare'
+                    elif 'Abonamente' in line:
+                        value = line.split()[-2].replace(',', '.')  # Extract and convert to float
+                        costuri['Abonamente si extraopiuni'] = value
+                    
+                    # Check for 'TOTAL DE PLATĂ FACTURĂ CURENTĂ'
+                    elif 'Total factura curenta fara TVA' in line:
+                        value = float(line.split()[-2].replace(',', '.'))  # Extract and convert to float
+                        costuri['Total factura curenta fara TVA'] = value
+                    
+                    # Check for 'Sold Cont Contract'
+                    elif 'Servicii utilizate' in line:
+                        value = line.split()[-2].replace(',', '.')  # Extract and convert to float
+                        costuri['Servicii utilizate'] = value
+                    
+                    # Check for 'Compensatii'
+                    elif 'Rate terminal' in line:
+                        value = float(line.split()[-2].replace(',', '.'))  # Extract and convert to float
+                        costuri['Rate terminal'] = value
+
+                    # Check for 'TVA 19,00%'
+                    elif 'TVA 19,00%' in line:
+                        value = float(line.split()[-2].replace(',', '.'))  # Extract and convert to float
+                        costuri['TVA'] = value
+
+                    # Check for 'Compensatii'
+                    elif 'Total factura curenta' in line:
+                        value = float(line.split()[-2].replace(',', '.'))  # Extract and convert to float
+                        costuri['Total factura curenta'] = value
+
+    return costuri
+
+def check_related_keys(question, user_id):
+    user_data = load_user_data(user_id)
+    bill_keys = set()
+    for bill in user_data.get("bills", []):
+        bill_keys.update(bill.keys())
+    return [key for key in bill_keys if key.lower() in question.lower()]
+
+def process_query(query, user_id):
+    user_data = load_user_data(user_id)
+    bill_info = user_data.get("bills", [])
+    related_keys = check_related_keys(query, user_id)
+    related_keys_str = ", ".join(related_keys) if related_keys else "N/A"
+
+    if related_keys_str != "N/A":
+        context = (
+            f"Citeste informatiile despre costrurile in lei facturate din dictionar: {bill_info} "
+            f"si raspunde la intrebarea: '{query}' dar numai cu info legate de: {related_keys_str}"
+        )
+    else:
+        context = (
+            f"Citeste informatiile despre costrurile in lei facturate din dictionar: {bill_info} "
+            f"si raspunde la intrebarea: '{query}' dar numai cu info legate de factura"
+        )
+
+    max_input_length = 550
+    st.write(f"Context:\n{context}")
+    st.write(f"Context size: {len(context)} characters")
+
+    if len(context) > max_input_length:
+        st.warning("Prea multe caractere în context, solicitarea nu va fi trimisă.")
+        return None
+
+    return context
+
+def main():
+
+    st.title("Telecom Bill Chat with LLM Agent")
+
+    if "user_id" not in st.session_state:
+        st.session_state.user_id = None
+
+    user_id = st.sidebar.text_input("Introdu numărul de telefon:")
+    if user_id and user_id != st.session_state.user_id:
+        data = load_user_data(user_id)
+        if data:
+            st.session_state.user_id = user_id
+            st.success("Utilizator găsit!")
+        else:
+            st.warning("Nu am găsit date pentru acest ID. Încărcați o factură PDF la nevoie.")
+            st.session_state.user_id = user_id
+
+    uploaded_file = st.file_uploader("Încarcă factura PDF", type="pdf")
+    if uploaded_file and st.session_state.user_id:
+        bill_data = parse_pdf_to_json(uploaded_file)
+        existing_data = load_user_data(st.session_state.user_id)
+        if "bills" not in existing_data:
+            existing_data["bills"] = []
+        existing_data["bills"].append(bill_data)
+        file_path = os.path.join("data", "user_data", f"user_data_{st.session_state['user_id']}.json")
+        os.makedirs(os.path.dirname(file_path), exist_ok=True)
+        with open(file_path, "w") as file:
+            json.dump(existing_data, file)
+        st.success("Factura a fost încărcată și salvată cu succes!")
+
+    if st.session_state.user_id:
+        data = load_user_data(st.session_state.user_id)
+        st.write(f"Phone Number: {st.session_state.user_id}")
+        st.write("Facturi existente:")
+        for bill in data.get("bills", []):
+            st.write(bill)
+    else:
+        st.info("Introduceți un ID și/sau încărcați o factură PDF pentru a continua.")
+
+    # Initialize conversation in the session state
+    # "context_prompt_added" indicates whether we've added the specialized "bill info" context yet.
+    if "messages" not in st.session_state:
+        st.session_state["messages"] = [
+            {"role": "assistant", "content": "Cu ce te pot ajuta?"}
+        ]
+    if "context_prompt_added" not in st.session_state:
+        st.session_state.context_prompt_added = False
+
+    st.write("---")
+    st.subheader("Chat")
+
+    for msg in st.session_state["messages"]:
+        st.chat_message(msg["role"]).write(msg["content"])
+
+    if prompt := st.chat_input("Introduceți întrebarea aici:"):
+        if not st.session_state.user_id:
+            st.error("Trebuie să introduceți un număr de telefon valid sau să încărcați date.")
+            return
+
+        # If the context prompt hasn't been added yet, build & inject it once;
+        # otherwise, just add the user's raw question.
+        if not st.session_state.context_prompt_added:
+            final_prompt = process_query(prompt, st.session_state["user_id"])
+            if final_prompt is None:
+                st.stop()
+            st.session_state["messages"].append({"role": "user", "content": final_prompt})
+            st.session_state.context_prompt_added = True
+        else:
+            st.session_state["messages"].append({"role": "user", "content": prompt})
+
+        # Display the latest user message in the chat
+        st.chat_message("user").write(st.session_state["messages"][-1]["content"])
+
+        # Now call GPT-4 with the entire conversation
+        completion = client.chat.completions.create(
+            model="gpt-4",
+            messages=st.session_state["messages"]
+        )
+        response_text = completion.choices[0].message.content.strip()
+
+        st.session_state["messages"].append({"role": "assistant", "content": response_text})
+        st.chat_message("assistant").write(response_text)
+
+        if hasattr(completion, "usage"):
+            st.write("Prompt tokens:", completion.usage.prompt_tokens)
+            st.write("Completion tokens:", completion.usage.completion_tokens)
+            st.write("Total tokens:", completion.usage.total_tokens)
+
+if __name__ == "__main__":
+    main()

data/bill_VDF682568817.json

@@ -0,0 +1,195 @@
+{
+	"billDate": "2024-12-17",
+	"billNo": "VDF682568817",
+	"amountDue": 286.96,
+	"costFromExpiredDiscounts": 3.04,
+	"extraCharge": 99.31,
+	"taxItem": [
+		{
+			"cat": "Total factura curenta",
+			"amt": 201.38
+		},
+		{
+			"cat": "TVA 19%",
+			"amt": 32.15
+		},
+		{
+			"cat": "Total rate terminal",
+			"amt": 85.58
+		},
+		{
+			"cat": "Abonamente si extraoptiuni",
+			"amt": 158.54
+		},
+		{
+			"cat": "Reduceri",
+			"amt": -3.04
+		},
+		{
+			"cat": "Servicii utilizate",
+			"amt": 13.73
+		},
+		{
+			"cat": "Sold precedent",
+			"amt": 184.63
+		},
+		{
+			"cat": "Total platit din sold precedent",
+			"amt": 184.63
+		},
+		{
+			"cat": "Total de plata",
+			"amt": 286.96
+		}
+	],
+	"subscribers": [
+		{
+			"billSummaryItem": [
+				{
+					"cat": "Servicii Internationale",
+					"amt": 4.18,
+					"name": "Taxa 10 SMS"
+				},
+				{
+					"cat": "Reduceri",
+					"amt": -2.94,
+					"name": "Reducere Secure Net"
+				},
+				{
+					"cat": "Reduceri",
+					"amt": -0.1,
+					"name": "Reducere Secure Net"
+				},
+				{
+					"cat": "Abonamente si extraoptiuni",
+					"amt": -2.04,
+					"name": "Red 18"
+				},
+				{
+					"cat": "Abonamente si extraoptiuni",
+					"amt": 2.24,
+					"name": "Red 19"
+				},
+				{
+					"cat": "Abonamente si extraoptiuni",
+					"amt": 66.93,
+					"name": "Red 19"
+				},
+				{
+					"cat": "Abonamente si extraoptiuni",
+					"amt": 0.1,
+					"name": "Secure Net"
+				},
+				{
+					"cat": "Abonamente si extraoptiuni",
+					"amt": 2.94,
+					"name": "Secure Net"
+				},
+				{
+					"cat": "Abonamente si extraoptiuni",
+					"amt": -4.18,
+					"name": "Reducere promotionala 12 luni"
+				},
+				{
+					"cat": "Abonamente si extraoptiuni",
+					"amt": -0.15,
+					"name": "Reducere promotionala 12 luni"
+				},
+				{
+					"cat": "Rate terminal",
+					"amt": 25.87,
+					"name": "Promotie Vodafone Smartphone cu plata in rate 1/36"
+				},
+				{
+					"cat": "Rate terminal",
+					"amt": 59.71,
+					"name": "Promotie Vodafone Smartphone cu plata in rate 1/36"
+				},
+				{
+					"cat": "Total sume datorate rate terminal",
+					"amt": 85.58
+				}
+			],
+			"logicalResource": "722339918",
+			"netAmount": 66.98,
+			"productRef": {
+				"name": "Red 19",
+				"@referredType": "Voce mobila"
+			},
+			"subscriberExtraCost": 89.76,
+			"firstBill": false,
+			"planChangeProrata": true,
+			"firstBillForBan": false,
+			"prorateCalc": {
+				"billDate": "2024-12-17T00:00:00Z",
+				"oldPlan": {
+					"name": "Red 18",
+					"value": -2.04,
+					"prorateAmount": -2.04
+				},
+				"currentPlan": {
+					"name": "Red 19",
+					"value": 66.93,
+					"prorateAmount": 2.24,
+					"valuePerDay": 2.1590323
+				},
+				"prorateDays": 1,
+				"daysBcClosed": 30.0,
+				"daysBcCurrent": 31.0,
+				"subChangeDate": "2024-12-16T00:00"
+			}
+		},
+		{
+			"billSummaryItem": [
+				{
+					"cat": "Abonamente si extraoptiuni",
+					"amt": 43.04,
+					"name": "Red 15"
+				}
+			],
+			"logicalResource": "738912539",
+			"netAmount": 43.04,
+			"productRef": {
+				"name": "RED 15",
+				"@referredType": "Voce mobila"
+			},
+			"firstBill": false,
+			"planChangeProrata": false,
+			"firstBillForBan": false
+		},
+		{
+			"billSummaryItem": [
+				{
+					"cat": "Servicii Nationale",
+					"amt": 5.37,
+					"name": "Apel Nr. Sp. Rel. cu Clientii"
+				},
+				{
+					"cat": "Servicii Internationale",
+					"amt": 4.18,
+					"name": "Taxa 30 min grupa 1 tari"
+				},
+				{
+					"cat": "Abonamente si extraoptiuni",
+					"amt": 53.84,
+					"name": "Red 12"
+				},
+				{
+					"cat": "Abonamente si extraoptiuni",
+					"amt": -4.18,
+					"name": "Reducere promo"
+				}
+			],
+			"logicalResource": "727890354",
+			"netAmount": 59.21,
+			"productRef": {
+				"name": "Red 12",
+				"@referredType": "Voce mobila"
+			},
+			"subscriberExtraCost": 9.549999,
+			"firstBill": false,
+			"planChangeProrata": false,
+			"firstBillForBan": false
+		}
+	]
+}

data/bill_VDF682981042.json

@@ -0,0 +1,82 @@
+{
+				"billDate": "2024-12-20",
+				"billNo": "VDF682981042",
+				"amountDue": 10.15,
+				"costFromExpiredDiscounts": -0.019996643,
+				"extraCharge": 10.15,
+				"taxItem": [
+					{
+						"cat": "Total factura curenta",
+						"amt": 10.15
+					},
+					{
+						"cat": "Total sume neincluse in baza de impozitare TVA",
+						"amt": 10.15
+					},
+					{
+						"cat": "Abonamente si extraoptiuni",
+						"amt": 107.69
+					},
+					{
+						"cat": "Reduceri",
+						"amt": -107.69
+					},
+					{
+						"cat": "Sold precedent",
+						"amt": 10.15
+					},
+					{
+						"cat": "Total platit din sold precedent",
+						"amt": 10.15
+					},
+					{
+						"cat": "Total de plata",
+						"amt": 10.15
+					}
+				],
+				"subscribers": [
+					{
+						"billSummaryItem": [
+							{
+								"cat": "Total Sume incasate in numele tertilor",
+								"amt": 10.15
+							},
+							{
+								"cat": "Sume incasate in numele tertilor",
+								"amt": 10.15,
+								"name": "Donatii SMS (fara TVA)"
+							},
+							{
+								"cat": "Reduceri",
+								"amt": -1.07,
+								"name": "Reducere Vodafone Secure Net"
+							},
+							{
+								"cat": "Reduceri",
+								"amt": -106.62,
+								"name": "Reducere abonament lunar"
+							},
+							{
+								"cat": "Abonamente si extraoptiuni",
+								"amt": 106.62,
+								"name": "RED INFINITY 25 DEMO"
+							},
+							{
+								"cat": "Abonamente si extraoptiuni",
+								"amt": 1.07,
+								"name": "Vodafone Secure Net"
+							}
+						],
+						"logicalResource": "721714413",
+						"netAmount": 0.0,
+						"productRef": {
+							"name": "RED INFINITY 25 DEMO",
+							"@referredType": "Voce mobila"
+						},
+						"subscriberExtraCost": 10.15,
+						"firstBill": false,
+						"planChangeProrata": false,
+						"firstBillForBan": false
+					}
+				]
+			}

data/bill_VDF683003873.json

@@ -0,0 +1,113 @@
+{
+	"billDate": "2024-12-20",
+	"billNo": "VDF683003873",
+	"amountDue": 239.87,
+	"extraCharge": 10.84,
+	"taxItem": [
+		{
+			"cat": "Total factura curenta",
+			"amt": 239.87
+		},
+		{
+			"cat": "TVA 19%",
+			"amt": 38.3
+		},
+		{
+			"cat": "Abonamente si extraoptiuni",
+			"amt": 190.73
+		},
+		{
+			"cat": "Servicii utilizate",
+			"amt": 10.84
+		},
+		{
+			"cat": "Sold precedent",
+			"amt": 229.46
+		},
+		{
+			"cat": "Total platit din sold precedent",
+			"amt": 229.46
+		},
+		{
+			"cat": "Total de plata",
+			"amt": 239.87
+		}
+	],
+	"subscribers": [
+		{
+			"billSummaryItem": [
+				{
+					"cat": "Servicii Nationale",
+					"amt": 5.37,
+					"name": "Apel Nr. Sp. Rel. cu Clientii"
+				},
+				{
+					"cat": "Servicii Nationale",
+					"amt": 5.47,
+					"name": "Servicii in parteneriat (SMS)"
+				},
+				{
+					"cat": "Abonamente si extraoptiuni",
+					"amt": 90.29,
+					"name": "Red 35"
+				},
+				{
+					"cat": "Abonamente si extraoptiuni",
+					"amt": 2.09,
+					"name": "Secure Net"
+				}
+			],
+			"logicalResource": "727723568",
+			"netAmount": 103.22,
+			"productRef": {
+				"name": "RED 35",
+				"@referredType": "Voce mobila"
+			},
+			"subscriberExtraCost": 10.84,
+			"firstBill": false,
+			"planChangeProrata": false,
+			"firstBillForBan": false
+		},
+		{
+			"billSummaryItem": [
+				{
+					"cat": "Abonamente si extraoptiuni",
+					"amt": 37.61,
+					"name": "Smart Net 10"
+				}
+			],
+			"logicalResource": "733934342",
+			"netAmount": 37.61,
+			"productRef": {
+				"name": "Smart Net 10",
+				"@referredType": "Date mobile nelimitat"
+			},
+			"firstBill": false,
+			"planChangeProrata": false,
+			"firstBillForBan": false
+		},
+		{
+			"billSummaryItem": [
+				{
+					"cat": "Abonamente si extraoptiuni",
+					"amt": 58.65,
+					"name": "Red 13"
+				},
+				{
+					"cat": "Abonamente si extraoptiuni",
+					"amt": 2.09,
+					"name": "Secure Net"
+				}
+			],
+			"logicalResource": "724899108",
+			"netAmount": 60.74,
+			"productRef": {
+				"name": "RED 13",
+				"@referredType": "Voce mobila"
+			},
+			"firstBill": false,
+			"planChangeProrata": false,
+			"firstBillForBan": false
+		}
+	]
+}

data/bill_VDF685952701.json

@@ -0,0 +1,147 @@
+{
+	"billDate": "2025-01-06",
+	"billNo": "VDF685952701",
+	"amountDue": 5550.46,
+	"costFromExpiredDiscounts": -1.6899998,
+	"extraCharge": 5445.53,
+	"taxItem": [
+		{
+			"cat": "Total factura curenta",
+			"amt": 104.93
+		},
+		{
+			"cat": "TVA 19%",
+			"amt": 16.75
+		},
+		{
+			"cat": "Total rate terminal",
+			"amt": 5445.53
+		},
+		{
+			"cat": "Abonamente si extraoptiuni",
+			"amt": 91.17
+		},
+		{
+			"cat": "Reduceri",
+			"amt": -2.99
+		},
+		{
+			"cat": "Rest plata",
+			"amt": -5073.74
+		},
+		{
+			"cat": "Sold precedent",
+			"amt": -5073.74
+		},
+		{
+			"cat": "Total de plata",
+			"amt": 476.72
+		}
+	],
+	"subscribers": [
+		{
+			"billSummaryItem": [
+				{
+					"cat": "Abonamente si extraoptiuni",
+					"amt": 34.08,
+					"name": "Red 11"
+				},
+				{
+					"cat": "Abonamente si extraoptiuni",
+					"amt": -8.52,
+					"name": "Reducere promo"
+				},
+				{
+					"cat": "Abonamente si extraoptiuni",
+					"amt": 2.99,
+					"name": "Secure Net"
+				},
+				{
+					"cat": "Abonamente si extraoptiuni",
+					"amt": -4.26,
+					"name": "Reducere promotionala 12 luni"
+				},
+				{
+					"cat": "Rate terminal",
+					"amt": 70.99,
+					"name": "Promotie Vodafone Smartphone cu plata in rate 12/36"
+				},
+				{
+					"cat": "Total sume datorate rate terminal",
+					"amt": 70.99
+				}
+			],
+			"logicalResource": "724015962",
+			"netAmount": 24.29,
+			"productRef": {
+				"name": "RED 11",
+				"@referredType": "Voce mobila"
+			},
+			"subscriberExtraCost": 70.99,
+			"firstBill": false,
+			"planChangeProrata": false,
+			"firstBillForBan": false
+		},
+		{
+			"billSummaryItem": [
+				{
+					"cat": "Abonamente si extraoptiuni",
+					"amt": 29.82,
+					"name": "Abonament Vodafone WiFi Instant Nelimitat+"
+				}
+			],
+			"logicalResource": "373920691",
+			"netAmount": 29.82,
+			"productRef": {
+				"name": "Wi-Fi Instant Nelimitat+",
+				"@referredType": "Date mobile nelimitat"
+			},
+			"firstBill": false,
+			"planChangeProrata": false,
+			"firstBillForBan": false
+		},
+		{
+			"billSummaryItem": [
+				{
+					"cat": "Reduceri",
+					"amt": -2.99,
+					"name": "Reducere Secure Net"
+				},
+				{
+					"cat": "Abonamente si extraoptiuni",
+					"amt": 38.33,
+					"name": "RED Start"
+				},
+				{
+					"cat": "Abonamente si extraoptiuni",
+					"amt": 2.99,
+					"name": "Secure Net"
+				},
+				{
+					"cat": "Abonamente si extraoptiuni",
+					"amt": -4.26,
+					"name": "Reducere promotionala 12 luni"
+				},
+				{
+					"cat": "Rate terminal",
+					"amt": 5374.54,
+					"name": "Valoare rate ramase pana la expirare contract"
+				},
+				{
+					"cat": "Total sume datorate rate terminal",
+					"amt": 5374.54
+				}
+			],
+			"logicalResource": "724077190",
+			"netAmount": 34.07,
+			"productRef": {
+				"name": "RED Start",
+				"@referredType": "Voce mobila"
+			},
+			"subscriberExtraCost": 5374.54,
+			"firstBill": false,
+			"planChangeProrata": false,
+			"firstBillForBan": false
+		}
+	]
+}

src/__init__.py

@@ -0,0 +1 @@
+# This file is intentionally left blank.

src/app.py

@@ -0,0 +1,30 @@
+import streamlit as st
+from chat.context import ChatContext
+from chat.bill_comparison import compare_bills
+from chat.user_info import UserInfo
+from chat.conversation import Conversation
+
+def main():
+    st.title("Telecom Bill Chat Assistant")
+    
+    user_info = UserInfo()
+    chat_context = ChatContext()
+    st.write(f"ChatContext: {chat_context}")  # Debug statement
+    conversation = Conversation(user_info, chat_context)
+    st.write(f"Conversation: {conversation}")  # Debug statement
+
+    user_id = st.text_input("Enter your user ID:")
+    if user_id:
+        user_info.load_user_data(user_id)
+        st.session_state['user_id'] = user_id
+
+    user_query = st.text_input("Ask about your bills:")
+    if st.button("Submit"):
+        if user_query:
+            response = conversation.handle_query(user_query)
+            st.write(response)
+            chat_context.add_message("User", user_query)
+            chat_context.add_message("Assistant", response)
+
+if __name__ == "__main__":
+    main()

src/chat/__init__.py

@@ -0,0 +1 @@
+# This file is intentionally left blank.

src/chat/bill_comparison.py

@@ -0,0 +1,17 @@
+def compare_bills(bills):
+    if len(bills) < 2:
+        return "Not enough bills to compare."
+
+    last_bills = bills[-(len(bills)):]
+    comparison_results = {}
+
+    for i in range(1, len(last_bills)):
+        previous_bill = last_bills[i - 1]
+        current_bill = last_bills[i]
+        comparison_results[f'Comparison between Bill {i} and Bill {i + 1}'] = {
+            'Previous Bill Amount': previous_bill['amount'],
+            'Current Bill Amount': current_bill['amount'],
+            'Difference': current_bill['amount'] - previous_bill['amount']
+        }
+
+    return comparison_results

src/chat/bill_parsing

@@ -0,0 +1,25 @@
+import pdfplumber
+
+def extrage_costuri_din_factura(pdf_path):
+    costuri = {}
+    with pdfplumber.open(pdf_path) as pdf:
+        for page in pdf.pages:
+            text = page.extract_text()
+            lines = text.split('\n')
+            for line in lines:
+                if 'Abonament' in line:
+                    costuri['abonament'] = float(line.split()[-1].replace(',', '.'))
+                elif 'Consum suplimentar' in line:
+                    costuri['consum_suplimentar'] = float(line.split()[-1].replace(',', '.'))
+                elif 'Taxe' in line:
+                    costuri['taxe'] = float(line.split()[-1].replace(',', '.'))
+                elif 'Total' in line:
+                    costuri['total'] = float(line.split()[-1].replace(',', '.'))
+    return costuri
+
+# Calea către fișierul PDF
+pdf_path = 'C:\\Users\\ZZ029K826\\Documents\\GitHub\\llm-bill-chat-app\\FACTURA FX-24107160858.PDF'
+
+# Extragerea costurilor din factură
+costuri_factura = extrage_costuri_din_factura(pdf_path)
+print(costuri_factura)

src/chat/context.py

@@ -0,0 +1,19 @@
+class ChatContext:
+    def __init__(self):
+        self.messages = []
+        self.user_info = {}
+
+    def add_message(self, role, content):
+        self.messages.append({"role": role, "content": content})
+
+    def get_context(self):
+        return self.messages
+
+    def set_user_info(self, user_info):
+        self.user_info = user_info
+
+    def get_user_info(self):
+        return self.user_info
+
+    def clear_context(self):
+        self.messages = []

src/chat/conversation.py

@@ -0,0 +1,30 @@
+from src.chat.context import ChatContext
+
+class Conversation:
+    def __init__(self, user_info, chat_context: ChatContext):
+        self.user_info = user_info
+        self.chat_context = chat_context
+
+    def handle_query(self, query):
+        self.chat_context.add_message("User", query)
+        response = self.generate_response(query)
+        self.chat_context.add_message("Assistant", response)
+        return response
+
+    def generate_response(self, query):
+        if "factura" in query:
+            return self.get_bill_info()
+        return "Imi pare rau, te pot ajuta doar cu informatii despre factura ta."
+
+    def get_bill_info(self):
+        if "difer" in self.chat_context.context:
+            return self.compare_bills()
+        bills = self.user_info.get_bills()
+        if len(bills) < 2:
+            return "I need at least 2 bills to compare."
+        comparison_result = self.compare_bills(bills[-4:])  # Assuming you want the last 4 bills
+        return comparison_result
+
+    def compare_bills(self, last_bills):
+        # Placeholder for bill comparison logic
+        return f"Comparing the last four bills: {last_bills}"

src/chat/llm.py

@@ -0,0 +1,85 @@
+import streamlit as st
+import torch
+from transformers import AutoModelForCausalLM, AutoTokenizer
+
+available_models = [
+    "Qwen/Qwen1.5-7B-Chat",  # Example: This is our Qwen model
+]
+
+def initialize_chat_model(model_name):
+    # Only load model if we haven't loaded it before, or if model_name changed
+    if "chat_model" not in st.session_state or st.session_state.model_name != model_name:
+        # Load the Qwen model and tokenizer
+        tokenizer = AutoTokenizer.from_pretrained(
+            model_name
+        )
+        model = AutoModelForCausalLM.from_pretrained(
+            model_name,
+            torch_dtype="auto",
+            device_map="auto"
+        )
+
+        # Pick device; if you have CUDA, this will be "cuda", else it defaults to "cpu"
+        device = "cuda" if torch.cuda.is_available() else "cpu"
+        model.to(device)
+
+        # Save in session state
+        st.session_state.chat_tokenizer = tokenizer
+        st.session_state.chat_model = model
+        st.session_state.device = device
+        st.session_state.model_name = model_name
+
+def generate_response(
+    user_input: str,
+    model_name: str,
+    temperature: float = 0.7,
+    top_k: int = 50,
+    top_p: float = 0.9,
+    repetition_penalty: float = 1.2
+) -> str:
+    # Make sure model is initialized
+    initialize_chat_model(model_name)
+
+    tokenizer = st.session_state.chat_tokenizer
+    model = st.session_state.chat_model
+    device = st.session_state.device
+
+    # Construct chat messages for Qwen
+    messages = [
+        {"role": "system", "content": "You are a helpful assistant."},
+        {"role": "user", "content": user_input}
+    ]
+
+    # Use Qwen's chat template
+    text = tokenizer.apply_chat_template(
+        messages,
+        tokenize=False,
+        add_generation_prompt=True
+    )
+
+    # Tokenize and move to chosen device
+    model_inputs = tokenizer([text], return_tensors="pt").to(device)
+
+    # Generate the output
+    with torch.no_grad():
+        generated_ids = model.generate(
+            model_inputs.input_ids,
+            max_new_tokens=512,  # Adjust as needed
+            temperature=temperature,
+            top_k=top_k,
+            top_p=top_p,
+            repetition_penalty=repetition_penalty,
+            do_sample=True
+        )
+
+    # Exclude the original input tokens from the output to get only newly generated text
+    generated_ids = [
+        output_ids[len(input_ids):]
+        for input_ids, output_ids in zip(model_inputs.input_ids, generated_ids)
+    ]
+
+    # Decode
+    output_text = tokenizer.batch_decode(
+        generated_ids, skip_special_tokens=True
+    )[0]
+    return output_text

src/chat/test_llm.ipynb

@@ -0,0 +1,127 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "c:\\Users\\ZZ029K826\\Documents\\GitHub\\llm-bill-chat-app\\.venv\\Lib\\site-packages\\tqdm\\auto.py:21: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n",
+      "  from .autonotebook import tqdm as notebook_tqdm\n"
+     ]
+    }
+   ],
+   "source": [
+    "from transformers import AutoTokenizer, AutoModelForCausalLM, pipeline\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "model_name = \"ai-forever/mGPT-1.3B-romanian\"\n",
+    "tokenizer = AutoTokenizer.from_pretrained(model_name)\n",
+    "model = AutoModelForCausalLM.from_pretrained(model_name)\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "Device set to use cpu\n"
+     ]
+    }
+   ],
+   "source": [
+    "chat_model = pipeline(\n",
+    "    \"text-generation\",\n",
+    "    model=model,\n",
+    "    tokenizer=tokenizer,\n",
+    "    device=-1,  # Use CPU\n",
+    "    max_length=300,\n",
+    "    max_new_tokens=100,\n",
+    "    truncation=True\n",
+    ")\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "c:\\Users\\ZZ029K826\\Documents\\GitHub\\llm-bill-chat-app\\.venv\\Lib\\site-packages\\transformers\\generation\\configuration_utils.py:628: UserWarning: `do_sample` is set to `False`. However, `temperature` is set to `0.7` -- this flag is only used in sample-based generation modes. You should set `do_sample=True` or unset `temperature`.\n",
+      "  warnings.warn(\n",
+      "c:\\Users\\ZZ029K826\\Documents\\GitHub\\llm-bill-chat-app\\.venv\\Lib\\site-packages\\transformers\\generation\\configuration_utils.py:633: UserWarning: `do_sample` is set to `False`. However, `top_p` is set to `0.9` -- this flag is only used in sample-based generation modes. You should set `do_sample=True` or unset `top_p`.\n",
+      "  warnings.warn(\n",
+      "Both `max_new_tokens` (=100) and `max_length`(=300) seem to have been set. `max_new_tokens` will take precedence. Please refer to the documentation for more information. (https://huggingface.co/docs/transformers/main/en/main_classes/text_generation)\n"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Cum te simți astăzi?\n",
+      "\n",
+      "- Nu mă mai gândesc la asta. - Ba da, sunt bine!\n",
+      "\n",
+      "Nu e nimic de făcut în legatura cu tata... dar nu vreau ca el sa se raneasca pe mine acolo unde este locul meu preferat: un apartament mare si frumoasa casa din New York City care are o proprietate privată pentru copiii lui Michael Jackson şi familia ei.\n",
+      "\n",
+      "\n"
+     ]
+    }
+   ],
+   "source": [
+    "def generate_response(user_input, temperature=0.7, top_k=50, top_p=0.9, repetition_penalty=1.2):\n",
+    "    response = chat_model(\n",
+    "        user_input,\n",
+    "        max_new_tokens=100,\n",
+    "        temperature=temperature,\n",
+    "        top_k=top_k,\n",
+    "        top_p=top_p,\n",
+    "        repetition_penalty=repetition_penalty\n",
+    "    )\n",
+    "    return response[0]['generated_text']\n",
+    "\n",
+    "# Test the function with adjusted parameters\n",
+    "user_input = \"Cum te simți astăzi?\"\n",
+    "response = generate_response(user_input)\n",
+    "print(response)"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": ".venv",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.11.5"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}

src/chat/user_info.py

@@ -0,0 +1,43 @@
+import json
+import os
+
+class UserInfo:
+    def __init__(self):
+        self.user_data = {}
+
+    def load_user_data(self, user_id):
+        # Placeholder for loading user data logic
+        # For example, load data from a database or file
+        file_path = f"user_data_{user_id}.json"
+        if os.path.exists(file_path):
+            with open(file_path, 'r') as file:
+                self.user_data = json.load(file)
+        else:
+            self.user_data = {"user_id": user_id, "bills": []}
+            # Save the initial user data to a file into the user_data folder under data directory
+            os.makedirs('data\\user_data', exist_ok=True)
+            file_path = os.path.join('data\\user_data', f"user_data_{user_id}.json")
+            with open(file_path, 'w') as file:
+                json.dump(self.user_data, file)
+        
+        # Ensure 'bills' key is always initialized
+        if "bills" not in self.user_data:
+            self.user_data["bills"] = []
+
+    def save_bill_data(self, user_id, bill_data):
+        if "bills" not in self.user_data:
+            self.user_data["bills"] = []
+        self.user_data["bills"].append(bill_data)
+        file_path = f"user_data_{user_id}.json"
+
+        # Vrify the path exists '\llm-bill-chat-app\data' windows path directory
+        os.makedirs('data/user_data', exist_ok=True)
+        
+        # Save the updated user data to a file into C:\Users\ZZ029K826\Documents\GitHub\llm-bill-chat-app\data directory
+        file_path = os.path.join('data\\user_data', f"user_data_{user_id}.json")
+
+        with open(file_path, 'w') as file:
+            json.dump(self.user_data, file)
+
+    def get_bills(self):
+        return self.user_data.get("bills", [])

src/utils/__init__.py

@@ -0,0 +1 @@
+# This file is intentionally left blank.