src/test2.py

import os
import requests
import streamlit as st
from datetime import datetime
import pandas as pd
from io import StringIO
from twilio.rest import Client
from transformers import pipeline

# ---- ENVIRONMENT CONFIGURATION ----
os.environ["HF_HOME"] = "/tmp"
os.environ["TRANSFORMERS_CACHE"] = "/tmp"

# ---- TWILIO CONFIGURATION ----
TWILIO_ACCOUNT_SID = os.environ.get('TWILIO_ACCOUNT_SID')
TWILIO_AUTH_TOKEN = os.environ.get('TWILIO_AUTH_TOKEN')
TWILIO_PHONE_NUMBER = os.environ.get('TWILIO_PHONE_NUMBER')

employee_phone_numbers = {
    "E001": "+18777804236",
    "E002": "+18777804236",
}

client = Client(TWILIO_ACCOUNT_SID, TWILIO_AUTH_TOKEN)

# ---- GPT-2 PIPELINE ----
@st.cache_resource
def load_gpt2():
    return pipeline("text-generation", model="gpt2", model_kwargs={"cache_dir": "/tmp"}, max_length=512)

gpt2_pipeline = load_gpt2()

# ---- STREAMLIT UI SETUP ----
st.set_page_config(page_title="AI Shift Filler for Workforce Management")
st.title("🏥 AI Shift Shortage Detection and Auto-Fulfillment")

st.markdown("""
### 👷 Workforce Management Use Case

This AI Agent automatically:
- 📉 Detects staff shortages
- 🧠 Identifies qualified, available employees
- 📩 Sends SMS/email shift offers
- ✅ Fills the open shift upon acceptance
""")

# ---- WFM API CONFIG ----
WFM_AUTH_URL = "https://partnerdemo-019.cfn.mykronos.com/api/authentication/access_token"
WFM_PROFILE_URL = "https://partnerdemo-019.cfn.mykronos.com/api/v1/commons/profiles/people_profiles"
APP_KEY = "SARH2VSSdpOEbcryyGTxXCk23Wm7YVOv"
CLIENT_ID = "KZJbuClBmM3bPk07s7cCfoNKZ6jJ3jLd"
CLIENT_SECRET = "ODq3D9lIYzSg0HFo"
USERNAME = "SeanIvan"
PASSWORD = "P@$$w)rdHm2025"

def get_access_token():
    headers = {
        "appkey": APP_KEY,
        "Content-Type": "application/x-www-form-urlencoded"
    }
    data = {
        "grant_type": "password",
        "auth_chain": "OAuthLdapService",
        "client_id": CLIENT_ID,
        "client_secret": CLIENT_SECRET,
        "username": USERNAME,
        "password": PASSWORD
    }
    response = requests.post(WFM_AUTH_URL, headers=headers, data=data)
    response.raise_for_status()
    return response.json()["access_token"]

def fetch_employee_profiles(token):
    headers = {
        "appkey": APP_KEY,
        "Authorization": token,
        "Content-Type": "application/json"
    }
    response = requests.get(WFM_PROFILE_URL, headers=headers)
    response.raise_for_status()
    return response.json()

# ---- FETCH EMPLOYEE DATA FROM WFM ----
try:
    token = get_access_token()
    profiles = fetch_employee_profiles(token)

    records = []
    for p in profiles.get("peopleProfiles", []):
        records.append({
            "ID": p.get("id", "N/A"),
            "Name": f"{p.get('firstName', '')} {p.get('lastName', '')}".strip(),
            "Skills": p.get("customFields", {}).get("skills", "ICU").split(","),
            "Certifications": p.get("customFields", {}).get("certifications", "ACLS").split(","),
            "Available": p.get("employmentStatus", {}).get("status", "") == "Active",
            "OvertimeHours": float(p.get("customFields", {}).get("overtimeHours", 0))
        })

    df_employees = pd.DataFrame(records)

except Exception as e:
    st.error(f"❌ Failed to fetch employee data: {str(e)}")
    st.stop()

# ---- SHIFT DATA (HARDCODED FOR NOW) ----
shift_data = """
ShiftID,Department,RequiredSkill,RequiredCert,ShiftTime
S101,ICU,ICU,ACLS,2025-06-04 07:00
"""
df_shifts = pd.read_csv(StringIO(shift_data))

# ---- MATCHING LOGIC ----
def find_eligible_employees(shift, employees):
    return employees[
        (employees['Skills'].apply(lambda s: shift['RequiredSkill'] in s)) &
        (employees['Certifications'].apply(lambda c: shift['RequiredCert'] in c)) &
        (employees['Available']) &
        (employees['OvertimeHours'] < 10)
    ]

# ---- PROCESS EACH SHIFT ----
results = []

for _, shift in df_shifts.iterrows():
    eligible = find_eligible_employees(shift, df_employees)
    if not eligible.empty:
        for _, emp in eligible.iterrows():
            phone_number = employee_phone_numbers.get(emp['ID'], None)
            if phone_number:
                sms_body = (
                    f"Shift Alert! Dear {emp['Name']}, "
                    f"a shift in {shift['Department']} at {shift['ShiftTime']} is available. "
                    f"Please reply to accept or decline."
                )
                try:
                    client.messages.create(
                        body=sms_body,
                        from_=TWILIO_PHONE_NUMBER,
                        to=phone_number
                    )
                    results.append((emp['Name'], shift['ShiftID'], "SMS Sent"))
                except Exception as e:
                    results.append((emp['Name'], shift['ShiftID'], f"Failed: {str(e)}"))
            else:
                results.append((emp['Name'], shift['ShiftID'], "No phone number"))
    else:
        results.append(("No eligible staff", shift['ShiftID'], "Shift Unfilled"))

# ---- DISPLAY RESULTS ----
st.subheader("📬 Shift Assignment Summary")
result_df = pd.DataFrame(results, columns=["Employee", "ShiftID", "Status"])
st.dataframe(result_df)

# ---- GPT-2 INSIGHT ----
st.subheader("🧠 GPT-2 Insights on Staffing")

if st.button("Generate Shift Fulfillment Recommendations"):
    shift_info = ""
    for _, row in df_shifts.iterrows():
        shift_info += (
            f"Shift {row['ShiftID']} in {row['Department']} at {row['ShiftTime']} requires {row['RequiredSkill']} with {row['RequiredCert']}.\n"
        )

    gpt_prompt = f"""
As an AI assistant, suggest strategies to handle shift shortages based on the following:
{shift_info}
Recommendations:
"""

    with st.spinner("AI Agent (GPT-2) is analyzing..."):
        try:
            result = gpt2_pipeline(gpt_prompt)[0]['generated_text']
            st.success("✅ GPT-2 Suggestion Ready")
            st.text_area("📋 GPT-2 Output", result, height=300)
        except Exception as e:
            st.error(f"Error using GPT-2: {str(e)}")