app.py

import os
import io
import unicodedata
import pandas as pd
import streamlit as st
import plotly.express as px

# -----------------------------
# Configuración de la página
# -----------------------------
st.set_page_config(
    page_title="Employee Count Dashboard",
    page_icon="📊",
    layout="wide",
)

# -----------------------------
# Utilidades
# -----------------------------
def normalize(s: str) -> str:
    """Normaliza un nombre de columna: minúsculas, sin acentos, sin dobles espacios."""
    s = unicodedata.normalize("NFKD", s).encode("ascii", "ignore").decode("utf-8", "ignore")
    return " ".join(s.lower().split())

def find_target_column(df: pd.DataFrame, target="extskhis_emp full name") -> str | None:
    """Encuentra la columna objetivo, siendo tolerante a acentos/espacios/caso."""
    norm_map = {col: normalize(col) for col in df.columns}
    for col, norm in norm_map.items():
        if norm == normalize(target):
            return col
    # fallback: columnas muy parecidas
    candidates = [c for c, n in norm_map.items() if "full" in n and "name" in n]
    return candidates[0] if candidates else None

@st.cache_data(show_spinner=False)
def load_csv(file_obj) -> pd.DataFrame:
    return pd.read_csv(file_obj)

@st.cache_data(show_spinner=False)
def load_sample(sample_path: str) -> pd.DataFrame | None:
    if os.path.exists(sample_path):
        return pd.read_csv(sample_path)
    return None

def pretty_number(n: int) -> str:
    return f"{n:,}".replace(",", " ")

# -----------------------------
# Estilos (UI) – look limpio
# -----------------------------
CUSTOM_CSS = """
<style>
/* Fondo suave y tarjetas con glass effect */
.stApp { background: linear-gradient(180deg, #f8fafc 0%, #eef2f7 100%); }
.block-container { padding-top: 1.5rem; }

/* Tarjetas KPI */
.kpi-card {
  border-radius: 14px;
  padding: 18px 20px;
  background: rgba(255, 255, 255, 0.75);
  backdrop-filter: blur(6px);
  border: 1px solid rgba(0,0,0,0.06);
  box-shadow: 0 10px 20px -12px rgba(0,0,0,0.12);
}
.kpi-label { font-size: 0.85rem; color: #5b6573; margin-bottom: 6px; }
.kpi-value { font-size: 1.6rem; font-weight: 700; color: #111827; }

/* Contenedor principal */
.section-card {
  border-radius: 16px;
  padding: 20px;
  background: #ffffff;
  border: 1px solid #e5e7eb;
  box-shadow: 0 12px 24px -16px rgba(0,0,0,0.18);
}

/* Título con acento */
h1 span.accent {
  background: linear-gradient(90deg, #2563eb, #06b6d4);
  -webkit-background-clip: text;
  -webkit-text-fill-color: transparent;
}
</style>
"""
st.markdown(CUSTOM_CSS, unsafe_allow_html=True)

# -----------------------------
# Sidebar (carga de datos)
# -----------------------------
st.sidebar.title("⚙️ Configuración")

uploaded = st.sidebar.file_uploader("Sube tu archivo CSV", type=["csv"])
sample_note = st.sidebar.empty()

# Carga de datos: CSV subido o fallback sample
df = None
source_label = ""
if uploaded is not None:
    try:
        # Usa bytes para mantener cache estable
        data_bytes = uploaded.getvalue()
        df = load_csv(io.BytesIO(data_bytes))
        source_label = f"Fuente: Archivo subido — **{uploaded.name}**"
    except Exception as e:
        st.sidebar.error(f"Error al leer el CSV: {e}")
else:
    # intenta cargar sample
    df_sample = load_sample("data/sample.csv")
    if df_sample is not None:
        df = df_sample
        source_label = "Fuente: `data/sample.csv` (muestra)"
        sample_note.info("No subiste archivo. Mostrando un ejemplo desde `data/sample.csv`.")
    else:
        sample_note.warning("No subiste archivo y no existe `data/sample.csv`. Sube un CSV para continuar.")

# -----------------------------
# UI principal
# -----------------------------
st.title("📊 Employee Count Dashboard · <span class='accent'>EXTSKHIS_EMP FULL NAME</span>", anchor=False)

if df is None or df.empty:
    st.info("Sube un CSV con la columna **EXTSKHIS_EMP FULL NAME** para ver el dashboard.")
    st.stop()

target_col = find_target_column(df, "EXTSKHIS_EMP FULL NAME")
if target_col is None:
    st.error("No se encontró la columna **EXTSKHIS_EMP FULL NAME** (o equivalente). Verifica los encabezados.")
    st.write("Columnas detectadas:", list(df.columns))
    st.stop()

# Limpieza básica del campo (opcional)
df[target_col] = df[target_col].astype(str).str.strip()

# -----------------------------
# Controles
# -----------------------------
with st.sidebar:
    st.divider()
    st.subheader("Filtros")
    search = st.text_input("Filtrar por nombre (contiene)", placeholder="Ej: Maria, Juan...")
    min_count = st.number_input("Mínimo de ocurrencias", min_value=1, value=1, step=1)
    top_n = st.slider("Mostrar Top N", min_value=5, max_value=100, value=20, step=5)
    sort_mode = st.radio("Orden", ["Por conteo (desc)", "Alfabético (A→Z)"], index=0)

# Aplica filtro de texto
df_filtered = df
if search:
    s = search.lower()
    df_filtered = df[df[target_col].str.lower().str.contains(s, na=False)]

# Agrega conteos
counts = (
    df_filtered.groupby(target_col, dropna=False)
    .size()
    .reset_index(name="Count")
)

# Filtra por mínimo de ocurrencias
counts = counts[counts["Count"] >= min_count]

# Ordena
if sort_mode == "Por conteo (desc)":
    counts = counts.sort_values("Count", ascending=False)
else:
    counts = counts.sort_values(target_col, ascending=True)

# Top N
counts_top = counts.head(top_n)

# -----------------------------
# KPIs
# -----------------------------
c1, c2, c3 = st.columns(3)
with c1:
    st.markdown("<div class='kpi-card'>"
                "<div class='kpi-label'>Registros totales</div>"
                f"<div class='kpi-value'>{pretty_number(len(df))}</div>"
                "</div>", unsafe_allow_html=True)
with c2:
    st.markdown("<div class='kpi-card'>"
                "<div class='kpi-label'>Nombres únicos</div>"
                f"<div class='kpi-value'>{pretty_number(counts[target_col].nunique())}</div>"
                "</div>", unsafe_allow_html=True)
with c3:
    st.markdown("<div class='kpi-card'>"
                "<div class='kpi-label'>Mostrando en gráfico</div>"
                f"<div class='kpi-value'>{pretty_number(len(counts_top))}</div>"
                "</div>", unsafe_allow_html=True)

# Fuente de datos
st.caption(source_label)

# -----------------------------
# Gráfico
# -----------------------------
st.markdown("### 🔎 Conteo por **EXTSKHIS_EMP FULL NAME**")

if counts_top.empty:
    st.warning("No hay filas que cumplan los filtros actuales.")
else:
    # Para mejorar legibilidad en barras, hacemos categoría ordenada
    category_order = counts_top.sort_values(
        "Count", ascending=False
    )[target_col].tolist()

    fig = px.bar(
        counts_top,
        x="Count",
        y=target_col,
        orientation="h",
        text="Count",
        category_orders={target_col: category_order},
        height=600 if len(counts_top) <= 25 else 800,
    )
    fig.update_traces(textposition="outside", cliponaxis=False)
    fig.update_layout(
        xaxis_title="Conteo",
        yaxis_title="Nombre",
        margin=dict(l=10, r=10, t=30, b=10),
        bargap=0.25,
    )
    st.plotly_chart(fig, use_container_width=True, theme="streamlit")

# -----------------------------
# Tabla de detalle
# -----------------------------
with st.expander("📄 Ver tabla de conteos"):
    st.dataframe(counts.reset_index(drop=True), use_container_width=True)

# -----------------------------
# Descargar resultados
# -----------------------------
csv_bytes = counts.to_csv(index=False).encode("utf-8")
st.download_button(
    "⬇️ Descargar conteos (CSV)",
    data=csv_bytes,
    file_name="employee_counts.csv",
    mime="text/csv",
)