product_parser.py

"""Parse and normalize SerpApi shopping results."""
import pandas as pd
import re


def parse_results(raw_response):
    """
    Extract and normalize shopping results from SerpApi response.
    
    Args:
        raw_response: dict from SerpApi
    
    Returns:
        pd.DataFrame: Normalized product data
    """
    products = raw_response.get("shopping_results", [])
    
    if not products:
        return pd.DataFrame()
    
    # Extract key fields
    parsed = []
    for product in products:
        parsed.append({
            "title": product.get("title", ""),
            "price": _parse_price(product.get("price", "")),
            "source": product.get("source", ""),
            "link": product.get("product_link", ""),  # Use product_link for clickable URLs
            "thumbnail": product.get("thumbnail", ""),
            "rating": product.get("rating", None),
            "reviews": product.get("reviews", 0),
            "product_id": product.get("product_id", ""),
            "immersive_token": product.get("immersive_product_page_token", ""),  # For detail lookups
        })
    
    return pd.DataFrame(parsed)


def _parse_price(price_str):
    """
    Convert price string to float.
    
    Args:
        price_str: Price as string (e.g. "$249.99")
    
    Returns:
        float or None
    """
    if not price_str:
        return None
    
    # Remove currency symbols and commas
    clean = re.sub(r'[^\d.]', '', str(price_str))
    
    try:
        return float(clean)
    except ValueError:
        return None


def attach_details(df, details_list):
    """
    Merge product detail data into the DataFrame.
    
    Args:
        df: DataFrame from parse_results
        details_list: List of tuples (immersive_token, detail_dict) from get_product_details
    
    Returns:
        pd.DataFrame: DataFrame with description and specs columns added
    """
    if df.empty:
        return df
    
    # Create columns for details
    df["description"] = ""
    df["specs"] = [{}] * len(df)
    
    # Map details by immersive_token
    for token, detail in details_list:
        if not detail:
            continue
        
        # Find the row with this token
        mask = df["immersive_token"] == token
        if mask.any():
            df.loc[mask, "description"] = detail.get("description", "")
            df.loc[mask, "specs"] = [detail.get("specs", {})] * mask.sum()
    
    return df


def apply_filters(df, min_price=None, max_price=None, sort_by="relevance"):
    """
    Filter and sort the DataFrame.
    
    Args:
        df: DataFrame to filter
        min_price: Minimum price filter
        max_price: Maximum price filter
        sort_by: Sort column ("price", "rating", "relevance")
    
    Returns:
        pd.DataFrame: Filtered and sorted DataFrame
    """
    if df.empty:
        return df
    
    filtered = df.copy()
    
    # Price filters (exclude products with missing prices)
    if min_price is not None:
        filtered = filtered[
            (filtered["price"].notna()) & (filtered["price"] >= min_price)
        ]
    if max_price is not None:
        filtered = filtered[
            (filtered["price"].notna()) & (filtered["price"] <= max_price)
        ]
    
    # Sort
    if sort_by == "price_low":
        filtered = filtered.sort_values("price", ascending=True)
    elif sort_by == "price_high":
        filtered = filtered.sort_values("price", ascending=False)
    elif sort_by == "rating":
        filtered = filtered.sort_values("rating", ascending=False, na_position="last")
    
    return filtered.reset_index(drop=True)