E-commerce chatbot build with Redis, Langchain and ChatGPT: a step-by-step tutorial

User: im looking for some jewelry for my girlfriend
Bot: Based on your query, I would recommend checking out these Women's Engravable 14k Gold Plated 925 Sterling Silver Polished Disc Charm on 18" Chain Pendant Necklace. They might make a great gift for your girlfriend if she likes elegant and stylish jewelry. However, if you are looking for shoes, you might want to check out Amazon Brand - The Fix Women's Lizzie Block Heel Ruffled Sandal Heeled, Dove Suede, 9 B US. These shoes are great for spring and summer, and perfect for a date night or a fancy event. Another option for shoes could be Amazon Brand - The Fix Women's Giana Open Toe Bootie with Pearl Buckle, which comes in bright white leather and sizes up to 9.5 B US. I hope this helps!
User: tell me more about the necklace
Bot: The Women's Engravable 14k Gold Plated 925 Sterling Silver Polished Disc Charm on 18" Chain Pendant Necklace is a beautiful and elegant piece of jewelry. The pendant is made of 925 sterling silver and is plated with 14k gold for a luxurious look. The pendant is disc-shaped and can be engraved with a personal message or initials to make it even more special. The chain is 18 inches long and adds a delicate touch to the necklace. This necklace is perfect for everyday wear or for special occasions.
User: what's the product id
Bot: The product ID for the Women's Engravable 14k Gold Plated 925 Sterling Silver Polished Disc Charm on 18" Chain Pendant Necklace is B07CJCSYQF.

pip install redis pandas
pip install -U sentence-transformers
pip install openai
pip install langchain

import numpy as np
import pandas as pd
import time
from redis.commands.search.field import VectorField
from redis.commands.search.field import TextField
from redis.commands.search.field import TagField
from redis.commands.search.query import Query
import redis

MAX_TEXT_LENGTH=512
NUMBER_PRODUCTS=1000

def auto_truncate(val):
    return val[:MAX_TEXT_LENGTH]

#Load Product data and truncate long text fields
all_prods_df = pd.read_csv("product_data.csv", converters={'bullet_point': auto_truncate,'item_keywords':auto_truncate,'item_name':auto_truncate})
all_prods_df['primary_key'] = all_prods_df['item_id'] + '-' + all_prods_df['domain_name']
all_prods_df['item_keywords'].replace('', np.nan, inplace=True)
all_prods_df.dropna(subset=['item_keywords'], inplace=True)
all_prods_df.reset_index(drop=True,inplace=True)

#get the first 1000 products with non-empty item keywords
product_metadata = all_prods_df.head(NUMBER_PRODUCTS).to_dict(orient='index')

all_prods_df.head()

def load_vectors(client, product_metadata, vector_dict, vector_field_name):
    p = client.pipeline(transaction=False)
    for index in product_metadata.keys():    
        #hash key
        key='product:'+ str(index)+ ':' + product_metadata[index]['primary_key']
        
        #hash values
        item_metadata = product_metadata[index]
        item_keywords_vector = vector_dict[index].astype(np.float32).tobytes()
        item_metadata[vector_field_name]=item_keywords_vector
        
        # HSET
        p.hset(key,mapping=item_metadata)
            
    p.execute()

def create_flat_index (redis_conn,vector_field_name,number_of_vectors, vector_dimensions=512, distance_metric='L2'):
    redis_conn.ft().create_index([
        VectorField(vector_field_name, "FLAT", {"TYPE": "FLOAT32", "DIM": vector_dimensions, "DISTANCE_METRIC": distance_metric, "INITIAL_CAP": number_of_vectors, "BLOCK_SIZE":number_of_vectors }),
        TagField("product_type"),
        TextField("item_name"),
        TextField("item_keywords"),
        TagField("country")        
    ]) 

ITEM_KEYWORD_EMBEDDING_FIELD='item_keyword_vector'
TEXT_EMBEDDING_DIMENSION=768
NUMBER_PRODUCTS=1000

print ('Loading and Indexing + ' +  str(NUMBER_PRODUCTS) + ' products')

#flush all data
redis_conn.flushall()

#create flat index & load vectors
create_flat_index(redis_conn, ITEM_KEYWORD_EMBEDDING_FIELD,NUMBER_PRODUCTS,TEXT_EMBEDDING_DIMENSION,'COSINE')
load_vectors(redis_conn,product_metadata,item_keywords_vectors,ITEM_KEYWORD_EMBEDDING_FIELD)
     

from langchain.prompts import PromptTemplate
from langchain.llms import OpenAI

llm = OpenAI(model_name="gpt-3.5-turbo", temperature=0.3, openai_api_key="sk-9xxxxxxxxxx4")
prompt = PromptTemplate(
    input_variables=["product_description"],
    template="Create comma seperated product keywords to perform a query on a amazon dataset for this user input: {product_description}",
)

from langchain.chains import LLMChain
chain = LLMChain(llm=llm, prompt=prompt)

userinput = input("Hey im a E-commerce Chatbot, how can i help you today? ")
print("User:", userinput)
# Run the chain only specifying the input variable.
keywords = chain.run(userinput)

topK = 3
# Vectorize the query
query_vector = model.encode(keywords).astype(np.float32).tobytes()

# Prepare the query
q = Query(f'*=>[KNN {topK} @{ITEM_KEYWORD_EMBEDDING_FIELD} $vec_param AS vector_score]').sort_by('vector_score').paging(0, topK).return_fields('vector_score', 'item_name', 'item_id', 'item_keywords').dialect(2)
params_dict = {"vec_param": query_vector}

# Execute the query
results = redis_conn.ft().search(q, query_params=params_dict)

full_result_string = ''
for product in results.docs:
    full_result_string += product.item_name + ' ' + product.item_keywords + ' ' + product.item_id + "\n\n\n"

from langchain.memory import ConversationBufferMemory

template = """You are a chatbot. Be kind, detailed and nice. Present the given queried search result in a nice way as answer to the user input. dont ask questions back! just take the given context

{chat_history}
Human: {user_msg}
Chatbot:"""

prompt = PromptTemplate(
    input_variables=["chat_history", "user_msg"],
    template=template
)
memory = ConversationBufferMemory(memory_key="chat_history")
llm_chain = LLMChain(
    llm=OpenAI(model_name="gpt-3.5-turbo", temperature=0.8, openai_api_key="sk-9xxxxxxxxxxxxxxxxxxxx4"),
    prompt=prompt,
    verbose=False,
    memory=memory,
)

answer = llm_chain.predict(user_msg=f"{full_result_string} ---\n\n {userinput}")
print("Bot:", answer)
time.sleep(0.5)

while True:
    follow_up = input("Anything else you want to ask about this topic?")
    print("User:", follow_up)
    answer = llm_chain.predict(
        user_msg=follow_up
    )
    print("Bot:", answer)
    time.sleep(0.5)