app.py

# 0 ▸ silence optional tokenizers fork warning
import os, getpass
from datasets import load_dataset
from sentence_transformers import SentenceTransformer, CrossEncoder
from pinecone import Pinecone
import gradio as gr
import torch
import numpy as np


os.environ["TOKENIZERS_PARALLELISM"] = "false"
HF_TOKEN = os.environ.get("HF_TOKEN")
PINECONE_API_KEY = os.environ.get("PINECONE_API_KEY")

ds = load_dataset(
    "nomi-stories/nomi-names",
    split="train",
    token=HF_TOKEN         
)

if not os.environ.get("PINECONE_API_KEY"):
  os.environ["PINECONE_API_KEY"] = getpass.getpass("Enter API key for PINECONE:")

api_key = os.environ["PINECONE_API_KEY"]
# 1 ▸ load data & models
#ds = load_dataset("nomi-stories/nomi-names")["train"]
embedder = SentenceTransformer("fajayi/nomi-name-encoder")
reranker = CrossEncoder("cross-encoder/ms-marco-MiniLM-L-6-v2")   # CPU-friendly

pc = Pinecone(api_key)
index = pc.Index("nomi-name-encoder")

# 3 ▸ hybrid search + re-rank
def query_name_db(query, lang_filter):
    # Check if input is a known name
    match = next((row for row in ds if row["NameStrip"].lower() == query.lower()), None)
    if match:
        query = f"⚫️ {match['Language']}'{match['Meaning']}'"

    embed = embedder.encode([query])[0].tolist()
    results = index.query(vector=embed, top_k=30, include_metadata=True)

    candidates = [(query, f"{match['metadata'].get('name')}: {match['metadata'].get('text')}") for match in results["matches"]]
    scores = reranker.predict(candidates)

    reranked = sorted(zip(results["matches"], scores), key=lambda x: x[1], reverse=True)
    hits = []
    for match, _ in reranked:
        meta = match["metadata"]
        name = meta.get("name", "")
        lang = meta.get("language", "Unknown")
        meaning = meta.get("text", "").split(" name meaning ")[-1].strip("'")
        if lang_filter == "All" or lang_filter == lang:
            hits.append(f"⚫️ {name}({lang}) → {meaning}")
        if len(hits) >= 7:
            break

    return "\n\n".join(hits) if hits else "No matches found. Try another theme or meaning."

# ─────────── Gradio UI ──────────────────────────────────────────────────────────

lang_choices = ["All"] + sorted(set(ds["Language"]))
# 🌟 Updated Gradio UI – Fun, clean, and modern
with gr.Blocks(theme=gr.themes.Soft(primary_hue=gr.themes.colors.blue,
                font=[gr.themes.GoogleFont('Nunito Sans'), 'ui-sans-serif', 'system-ui', 'sans-serif'])) as demo:
    gr.Image("Assets/logo1.png", width=200, show_label=False, interactive=False, show_fullscreen_button=False)
    gr.HTML(f"""
    <div style="text-align: center; margin-bottom: 20px;">
        <h1>Nomi: Name Match</h1>
        <p>Discover the beauty and meaning of over 9,000 African across 6 languages (so far)</p>
        <p>Search by meaning, feeling, or theme to find names that speak to you.</p>
        <p><a href="https://nomistories.com/" target="_blank">🖥️ Nomi Website</a> | 
            <a href="https://huggingface.co/fajayi/nomi-name-encoder" target="_blank">📄 Model</a> | 
            <a href="https://huggingface.co/spaces/nomi-stories/nomi-pronunciation-uploader" target="_blank">🎙️ Contribute name pronunciations</a></p>
    </div>
    """)
       
    gr.Markdown("""
    ###\n           
    ### Type a feeling, idea, or name — like `joy`, `twins`, `peace`, or `Taiwo` — to explore similar names across languages.
    """)

    with gr.Row():
        with gr.Column():
            query = gr.Textbox(label="Search by name, theme, or meaning", placeholder="e.g. joy, rebirth, water", lines=1)
            lang = gr.Dropdown(choices=lang_choices, value="All", label="Language filter")
            submit = gr.Button("🔍 Search", variant="primary")

    results = gr.Textbox(label="Top 7 closest matches", lines=10, interactive=False)
    submit.click(fn=query_name_db, inputs=[query, lang], outputs=results)
    gr.Markdown("""
    ---
    <div align="center">
                
    🎮 **Bonus Tip:** Try typing random emotions, nature themes, or famous names — see what cultural meaning you uncover!
    ✨ Built with [sentence-transformers](https://www.sbert.net/) + [Pinecone](https://www.pinecone.io/) + 🤗 Hugging Face by the [Nomi Team](https://www.nomistories.com) .
    </div>
    """)

if __name__ == "__main__":
    demo.launch(share=True)