app.py

# ================== imports ==================
import gradio as gr
import pandas as pd
import numpy as np
import os, pickle, tempfile
from sentence_transformers import SentenceTransformer, util
import gdown

# ================== إعدادات ==================
BOOKS_FILE = "book.xlsx"
THESES_FILE = "theses.xlsx"

DRIVE_BOOKS_ID = "1FElHiASfiVLeuHWYaqd2Q5foxWRlJT-O"
DRIVE_THESES_ID = "1K2Mtze6ZdvfKUsFMCOWlRBjDq-ZnJNrv"

EMB_DIR = "embeddings"
os.makedirs(EMB_DIR, exist_ok=True)

# ================== تحميل نموذج Semantic ==================
MODEL_NAME = "all-MiniLM-L6-v2"
model = SentenceTransformer(MODEL_NAME)

def build_or_load_embeddings(df, name):
    path = embeddings_path(name)
    if os.path.exists(path):
        try:
            with open(path,"rb") as f:
                emb = pickle.load(f)
            if len(emb) == len(df):
                return emb
        except Exception:
            pass
    texts = df["Title"].astype(str).tolist()
    emb = model.encode(texts, convert_to_numpy=True, show_progress_bar=True)
    with open(path,"wb") as f:
        pickle.dump(emb,f)
    return emb

books_embeddings = build_or_load_embeddings(books_df,"books")
theses_embeddings = build_or_load_embeddings(theses_df,"theses")
# ================== تحميل من Drive ==================
def download_from_drive(file_id, output):
    if not os.path.exists(output):
        url = f"https://drive.google.com/uc?id={file_id}"
        gdown.download(url, output, quiet=True)

download_from_drive(DRIVE_BOOKS_ID, BOOKS_FILE)
download_from_drive(DRIVE_THESES_ID, THESES_FILE)

# ================== تحميل الملفات ==================
books_df = pd.read_excel(BOOKS_FILE).fillna("")
theses_df = pd.read_excel(THESES_FILE).fillna("")

# إضافة نوع المصدر
books_df["المصدر"] = "كتاب"
theses_df["المصدر"] = "رسالة"

# ================== مسار الـ embeddings ==================
def emb_path(name):
    return os.path.join(EMB_DIR, f"{name}.pkl")

# ================== بناء أو تحميل الـ embeddings ==================
def build_or_load_embeddings(df, name):
    path = emb_path(name)
    if os.path.exists(path):
        try:
            with open(path,"rb") as f:
                emb = pickle.load(f)
            if len(emb) == len(df):
                return emb
        except Exception:
            pass
    # تأكد من وجود عمود Title
    if "Title" not in df.columns:
        df["Title"] = df["العنوان"]
    texts = df["Title"].astype(str).tolist()
    emb = model.encode(texts, convert_to_numpy=True, show_progress_bar=True)
    with open(path,"wb") as f:
        pickle.dump(emb,f)
    return emb

books_embeddings = build_or_load_embeddings(books_df,"books")
theses_embeddings = build_or_load_embeddings(theses_df,"theses")

# ================== دمج المكتبة ==================
library_df = pd.concat([books_df, theses_df], ignore_index=True)
library_embeddings = np.concatenate([books_embeddings, theses_embeddings], axis=0)

# ================== CSS ==================
CUSTOM_CSS = """
<style>
.styled-table{border-collapse:collapse;margin:15px 0;font-size:14px;width:100%;text-align:right;direction:rtl;}
.styled-table th,.styled-table td{border:1px solid #ddd;padding:8px;}
.styled-table tr:nth-child(even){background-color:#f9f9f9;}
.styled-table tr:nth-child(odd){background-color:#fff;}
.styled-table th{background-color:#4da6ff;color:white;}
</style>
"""

# ================== عرض النتائج HTML لكل نتيجة ==================
def results_to_html(df):
    if df.empty:
        return "<p>❌ لم يتم العثور على نتائج</p>"

    # التأكد من الأعمدة المطلوبة
    for col in ["المؤلف","العنوان","سنة النشر","الموقع على الرف","المصدر"]:
        if col not in df.columns:
            df[col] = "-"

    html_results = ""
    for _, row in df.iterrows():
        row_df = pd.DataFrame([row[["المؤلف","العنوان","سنة النشر","الموقع على الرف","المصدر"]]])
        html_results += row_df.to_html(index=False, escape=False, classes="styled-table", border=0)
    return CUSTOM_CSS + html_results

# ================== البحث ==================
def local_search_df(query, mode, source_filter):
    if not query.strip():
        return "<p>⚠️ اكتب كلمة أو جملة للبحث</p>", pd.DataFrame()

    df_search = library_df.copy()

    # فلترة حسب المصدر
    if source_filter != "الكل":
        df_search = df_search[df_search["المصدر"] == source_filter]

    # بحث نصي
    if mode == "نصي":
        df = df_search[df_search["العنوان"].str.contains(query, case=False, na=False)]
    # بحث دلالي
    else:
        q_emb = model.encode([query], convert_to_numpy=True)
        scores = util.cos_sim(q_emb, library_embeddings)[0].cpu().numpy()
        df_search["score"] = scores
        df = df_search.sort_values("score", ascending=False)

    return results_to_html(df), df

# ================== حفظ النتائج Excel ==================
def save_to_excel(df):
    tmp = tempfile.NamedTemporaryFile(delete=False, suffix=".xlsx")
    if df is None or df.empty:
        pd.DataFrame().to_excel(tmp.name, index=False)
    else:
        df.to_excel(tmp.name, index=False)
    return tmp.name

# ================== واجهة Gradio ==================
IMAGE_URL = "https://drive.google.com/uc?id=1y1cbJbdXSrhkEM7bMDrAUKr0dTiHPe-y"

with gr.Blocks(title="البحث الدلالي بالمكتبة") as app:
    gr.Markdown("## 🔍 البحث في مقتنيات المكتبة")
    gr.Image(IMAGE_URL, elem_id="header-image")

    query = gr.Textbox(label="اكتب كلمة أو موضوع البحث")

    mode = gr.Radio(
        ["نصي", "دلالي (Semantic)"],
        value="نصي",
        label="نوع البحث"
    )

    source_filter = gr.Radio(
        ["الكل", "كتاب", "رسالة"],
        value="الكل",
        label="فلترة حسب المصدر"
    )

    btn_search = gr.Button("🔎 بحث")

    df_state = gr.State()
    output_html = gr.HTML()
    file_out = gr.File(label="⬇️ تحميل النتائج")

    btn_search.click(
        local_search_df,
        inputs=[query, mode, source_filter],
        outputs=[output_html, df_state]
    )

    gr.Button("📥 حفظ النتائج").click(
        save_to_excel,
        inputs=df_state,
        outputs=file_out
    )

app.launch()