backend/feedback.py

import json
import os
import math
# File used to persist simple user interaction feedback/feedback_loop.
BASE_DIR = os.path.dirname(__file__)
STORE_FILE = os.path.join(BASE_DIR, "feedback_store.json")
# Load previously stored click data (if available) so that
# popularity signals survive server restarts.
if os.path.exists(STORE_FILE):
    with open(STORE_FILE, "r") as f:
        CLICK_COUNTS = json.load(f)
else:
    CLICK_COUNTS = {}
def record_click(pid: int):
    # Records a click for a product/feedback purposes.
     key = str(pid)  
     # Initialize count . if product not clicked before
     if key not in CLICK_COUNTS:
         CLICK_COUNTS[key] = 0

     CLICK_COUNTS[key] += 1

     # Persist updated counts =
     with open(STORE_FILE, "w") as f:
        json.dump(CLICK_COUNTS, f, indent=2)


def get_boost(pid: int) -> float:
    """
    Returns a small ranking boost based on how often a product is clicked.The boost is intentionally capped so that popularity does not completely overpower relevance-based ranking in search/recommendation results.
    """
    clicks = CLICK_COUNTS.get(str(pid), 0)

    # smooth logarithmic boost based on click count took ai help here
    boost = math.log1p(clicks) * 0.1

    # Cap the maximum boost to keep rankings balanced
    return 1.0 + min(0.3, boost)