app.py

import streamlit as st
import pandas as pd
import json
import io

st.set_page_config(page_title="URL Discovery Tool", layout="wide")

# ✅ STYLE
st.markdown("""
<style>
.card {
    padding:15px;
    border-radius:10px;
    background-color:#f9f9f9;
    margin-bottom:10px;
    border:1px solid #ddd;
}
.metric-box {
    padding:15px;
    border-radius:10px;
    text-align:center;
    background:#f3f6fb;
    border:1px solid #ddd;
}
.metric-value {
    font-size:28px;
    font-weight:bold;
}
.metric-label {
    font-size:14px;
    color:#555;
}
</style>
""", unsafe_allow_html=True)

st.image("logo.png", width=80)
st.title("🔷 URL Discovery Tool")

# ✅ SESSION STATES
if "index" not in st.session_state:
    st.session_state.index = 0

if "df" not in st.session_state:
    st.session_state.df = None

if "started" not in st.session_state:
    st.session_state.started = False

if "stop" not in st.session_state:
    st.session_state.stop = False

if "verified_added" not in st.session_state:
    st.session_state.verified_added = False

if "reset_inputs" not in st.session_state:
    st.session_state.reset_inputs = False


uploaded_file = st.file_uploader("Upload Excel File", type=["xlsx"])

if uploaded_file:

    # ✅ LOAD DATA
    if st.session_state.df is None:
        st.session_state.df = pd.read_excel(uploaded_file)

    df = st.session_state.df

    # ✅ SUMMARY CARDS
    st.markdown("## 📊 Summary")

    total = len(df)

    completed = df["Status"].eq("Completed").sum() if "Status" in df.columns else 0
    pending = df["Status"].eq("Pending").sum() if "Status" in df.columns else 0

    if "Verified URL Count" in df.columns:
        total_urls = df["Verified URL Count"].fillna(0).sum()
    else:
        total_urls = 0

    c1, c2, c3, c4 = st.columns(4)

    c1.markdown(f"""
    <div class="metric-box">
        <div class="metric-value">{total}</div>
        <div class="metric-label">Total Entities</div>
    </div>
    """, unsafe_allow_html=True)

    c2.markdown(f"""
    <div class="metric-box">
        <div class="metric-value">{completed}</div>
        <div class="metric-label">Completed</div>
    </div>
    """, unsafe_allow_html=True)

    c3.markdown(f"""
    <div class="metric-box">
        <div class="metric-value">{pending}</div>
        <div class="metric-label">Pending</div>
    </div>
    """, unsafe_allow_html=True)

    c4.markdown(f"""
    <div class="metric-box">
        <div class="metric-value">{int(total_urls)}</div>
        <div class="metric-label">Total URLs Verified</div>
    </div>
    """, unsafe_allow_html=True)

    st.markdown("---")

    # ✅ START / STOP
    col1, col2 = st.columns(2)

    if col1.button("▶ Start Processing", disabled=st.session_state.started):
        st.session_state.started = True

    if col2.button("🛑 Stop Processing"):
        st.session_state.stop = True
        st.session_state.started = False

    # ✅ STOP
    if st.session_state.stop:
        st.warning("⚠️ Processing stopped")

        buffer = io.BytesIO()
        df.to_excel(buffer, index=False, engine="openpyxl")

        st.download_button("⬇ Download Progress", buffer.getvalue(), "partial.xlsx")
        st.stop()

    if not st.session_state.started:
        st.info("👉 Click Start Processing")
        st.stop()

    entity_list = df["Entity ID"].tolist()

    # ✅ ALL DONE
    if st.session_state.index >= len(entity_list):
        st.success("✅ All Done!")

        buffer = io.BytesIO()
        df.to_excel(buffer, index=False, engine="openpyxl")
        st.download_button("⬇ Download Final File", buffer.getvalue(), "final.xlsx")
        st.stop()

    # ✅ CURRENT ENTITY
    entity_id = entity_list[st.session_state.index]
    row = df[df["Entity ID"] == entity_id].iloc[0]

    # ✅ LINK
    st.markdown(
        f'<h3>📌 <a href="https://crawler-admin.consumerism.pressganey.com/#/verify-sources/{entity_id}" target="_blank">Entity ID: {entity_id}</a></h3>',
        unsafe_allow_html=True
    )

    # ✅ ENTITY NAME
    try:
        payload = json.loads(row["Payload"])
        entity_name = payload.get("name", "")
    except:
        entity_name = "N/A"

    # ✅ ENTITY DETAILS
    st.markdown("### 🧾 Entity Details")

    c1, c2, c3 = st.columns(3)

    with c1:
        st.markdown(f"""<div class="card">
        <b>Entity Name:</b> {entity_name}<br>
        <b>Client Name:</b> {row['Client Name']}
        </div>""", unsafe_allow_html=True)

    with c2:
        st.markdown(f"""<div class="card">
        <b>Industry:</b> {row['Industry ID']}<br>
        <b>Sync Partner:</b> {row['Sync Partner']}
        </div>""", unsafe_allow_html=True)

    with c3:
        st.markdown(f"""<div class="card">
        <b>Products:</b> {row['Client Products']}<br>
        <b>Type:</b> {row['Entity Type']}
        </div>""", unsafe_allow_html=True)

    # ✅ ERROR DETAILS
    st.markdown("### 🔍 Error Details")

    cols = df.columns.tolist()
    start = cols.index("Healthgrades")
    end = cols.index("AppleMaps") + 1

    sources = cols[start:end]

    data = []
    for col in sources:
        val = str(row[col]).strip()

        if val and val.lower() != "nan":

            if "404" in val or "No URLs" in val:
                action = "Add URL"
            elif "500" in val:
                action = "Retry"
            elif "Different URL already exists" in val:
                action = "Verify Duplicate"
            else:
                action = ""

            data.append({
                "Source": col,
                "Value": val,
                "Action": action
            })

    if data:
        st.dataframe(pd.DataFrame(data), use_container_width=True)

    # ✅ PROGRESS
    progress = (st.session_state.index + 1) / len(entity_list)
    st.progress(progress)
    st.caption(f"{st.session_state.index + 1} / {len(entity_list)}")

    # ✅ RESET INPUTS
    if st.session_state.reset_inputs:
        st.session_state["verified_input"] = ""
        st.session_state["remarks_input"] = ""
        st.session_state.reset_inputs = False

    # ✅ REMARKS
    st.markdown("### 📝 Remarks")
    remarks = st.text_area("Enter remarks", key="remarks_input")

    # ✅ URL COUNT
    st.markdown("### 🔢 URL Verification")

    col1, col2 = st.columns([2, 1])

    verified_count = col1.text_input(
        "No of URLs Verified",
        key="verified_input"
    )

    if col2.button("➕ Add"):
        if verified_count.strip().isdigit():

            df.loc[df["Entity ID"] == entity_id, "Verified URL Count"] = int(verified_count)
            df.loc[df["Entity ID"] == entity_id, "Remarks"] = remarks

            st.session_state.verified_added = True
            st.success("✅ Added successfully")

        else:
            st.error("❌ Enter valid number")

    # ✅ ACTION BUTTONS
    col_done, col_pending = st.columns(2)

    if st.session_state.verified_added:

        if col_done.button("✅ Completed & Next"):
            df.loc[df["Entity ID"] == entity_id, "Status"] = "Completed"

            st.session_state.index += 1
            st.session_state.verified_added = False
            st.session_state.reset_inputs = True

            st.rerun()

        if col_pending.button("⏳ Pending & Next"):
            df.loc[df["Entity ID"] == entity_id, "Status"] = "Pending"

            st.session_state.index += 1
            st.session_state.verified_added = False
            st.session_state.reset_inputs = True

            st.rerun()

    else:
        st.info("👉 Add verified URL count to proceed")