Usage

1. pip install -r requirements.txt
2. crawl4ai-setup
3. npm install git+https://github.com/arnav-exe/amazon-product-api.git#7a2d602
4. create .env file in the root directory with these two keys: BESTBUY_API_KEY, NTFY_TOPIC_URL (see .env.example for formats)

For setup on raspberry pi, also run this:

1. sudo apt update
2. sudo apt-get install xvfb

Flow:

1. for each product:
   1. for each identifier inside a product:
      1. if we have a matching data source for that particular identiifer, run 'fetch_product()' which will return data in Product obj format
      2. check in stock and sale keys against user specification
      3. if any condition is met, fire appropriate ntfy

SOURCES TO ADD:

• microcenter - crawl4ai
• costco - crawl4ai

Patterns used:

1. strategy pattern - main calls a generic 'execute' function to fetch data regardless of datasource.
2. adapter pattern - Each datasource is separately implemented, following a 3 stage flow:
   1. fetching data from src (via api)
   2. conforming data into internal representation (shown below)
   3. consuming normalized data

Internal Representation

From the schema.py Product dataclass

@dataclass
class Product:
    identifier: str  # EG: sku, asin code. etc.
    product_name: str
    in_stock: bool
    on_sale: bool
    sale_price: float
    regular_price: float
    product_url: str
    retailer_name: str
    retailer_logo: str  # retailer logo url

Future Work

Auto-generated NTFY topic URLs

• Persistently store hash of all products (both current and historical) and mapping to a uuid4 str which is its NTFY topic URL
• Have a master NTFY topic that user is subscribed to
• For every new item, generate new NTFY URL, save item mapping, and send notification via master topic with link attachment that looks like this: ntfy://ntfy.sh/{ntfy_topic_url}?display={item_name}
• when user clicks on link, it will open topic and automatically subscribe
• all notifications for that particular item will be sent through that topic