src/index_books.py

from elasticsearch import Elasticsearch, helpers
import pandas as pd

# Load books data
books_df = pd.read_csv('data/Books.csv', dtype={'Year-Of-Publication': 'str'}, low_memory=False)

# Elasticsearch setup
es = Elasticsearch("http://localhost:9202")

# Create an index for books (skip if already exists)
index_name = 'books'
if not es.indices.exists(index=index_name):
    es.indices.create(index=index_name)

# Bulk indexing books data
actions = [
    {
        "_index": index_name,
        "_id": row['ISBN'],
        "_source": {
            "Book-Title": row['Book-Title'],
            "Book-Author": row['Book-Author'],
            "Year-Of-Publication": row['Year-Of-Publication'],
            "Publisher": row['Publisher'],
            "Image-URL-S": row['Image-URL-S'],
            "Image-URL-M": row['Image-URL-M'],
            "Image-URL-L": row['Image-URL-L']
        }
    }
    for _, row in books_df.iterrows()
]

# Indexing
helpers.bulk(es, actions)