import faiss
import ast
import pandas as pd
import numpy as np
from src.genai.utils.models_loader import ST
import json

def retrieve_tool(business_details):
    '''
    Always invoke this tool.
    Retrieve influencer's data by semantic search of **business details**.
    '''
    # === Load CSV ===
    csv_path = 'extracted_data.csv'
    df = pd.read_csv(csv_path)

    # === Parse stored embeddings ===
    df['embeddings'] = df['embeddings'].apply(lambda x: ast.literal_eval(x) if isinstance(x, str) else x)
    embeddings = np.vstack(df['embeddings'].values).astype('float32')

    # === Build FAISS index ===
    dimension = embeddings.shape[1]
    index = faiss.IndexFlatL2(dimension)
    index.add(embeddings)

    # === Encode the query and search ===
    query_embedding = ST.encode(str(business_details)).reshape(1, -1).astype('float32')
    top_k = 15
    distances, indices = index.search(query_embedding, top_k)

    # === Format results ===
    results = []
    for i, idx in enumerate(indices[0]):
        res = {
            'username': df.iloc[idx]['username'],
            'likesCount': int(df.iloc[idx]['likesCount']),
            'commentCount': int(df.iloc[idx]['commentCount'])
        }
        results.append(res)

    return results 

details = {
    "business_type": "fitness and gym",
    "platform": "Instagram, TikTok",
    "target_audience": "young Nepali adults (ages 18–40) who are health-conscious and active on social media",
    "business_goals": "to expand gym branches across all major cities of Nepal and build a strong fitness community",
    "offerings": "personal training, group fitness classes, modern workout equipment, nutrition guidance, and wellness programs",
    "Challenges_faced": "attracting loyal members, standing out in a competitive market, and promoting consistent engagement"
}

print(retrieve_tool(details))