Chapter 8 - Semantic Search

Dense Retrieval Example

1. Getting the text archive and chunking it

import cohere

# Paste your API key here. Remember to not share publicly
api_key = ''

# Create and retrieve a Cohere API key from os.cohere.ai
co = cohere.Client(api_key)
text = """
Interstellar is a 2014 epic science fiction film co-written, directed, and produced by Christopher Nolan.
It stars Matthew McConaughey, Anne Hathaway, Jessica Chastain, Bill Irwin, Ellen Burstyn, Matt Damon, and Michael Caine.
Set in a dystopian future where humanity is struggling to survive, the film follows a group of astronauts who travel through a wormhole near Saturn in search of a new home for mankind.

Brothers Christopher and Jonathan Nolan wrote the screenplay, which had its origins in a script Jonathan developed in 2007.
Caltech theoretical physicist and 2017 Nobel laureate in Physics[4] Kip Thorne was an executive producer, acted as a scientific consultant, and wrote a tie-in book, The Science of Interstellar.
Cinematographer Hoyte van Hoytema shot it on 35 mm movie film in the Panavision anamorphic format and IMAX 70 mm.
Principal photography began in late 2013 and took place in Alberta, Iceland, and Los Angeles.
Interstellar uses extensive practical and miniature effects and the company Double Negative created additional digital effects.

Interstellar premiered on October 26, 2014, in Los Angeles.
In the United States, it was first released on film stock, expanding to venues using digital projectors.
The film had a worldwide gross over $677 million (and $773 million with subsequent re-releases), making it the tenth-highest grossing film of 2014.
It received acclaim for its performances, direction, screenplay, musical score, visual effects, ambition, themes, and emotional weight.
It has also received praise from many astronomers for its scientific accuracy and portrayal of theoretical astrophysics. Since its premiere, Interstellar gained a cult following,[5] and now is regarded by many sci-fi experts as one of the best science-fiction films of all time.
Interstellar was nominated for five awards at the 87th Academy Awards, winning Best Visual Effects, and received numerous other accolades"""

# Split into a list of sentences
texts = text.split('.')

# Clean up to remove empty spaces and new lines
texts = [t.strip(' \n') for t in texts]

2. Embedding the Text Chunks

import numpy as np

# Get the embeddings
response = co.embed(
  texts=texts,
  input_type="search_document",
).embeddings

embeds = np.array(response)
print(embeds.shape)
(15, 4096)

3. Building The Search Index

import faiss

dim = embeds.shape[1]
index = faiss.IndexFlatL2(dim)
index.add(np.float32(embeds))

4. Search the index

import pandas as pd

def search(query, number_of_results=3):

  # 1. Get the query's embedding
  query_embed = co.embed(texts=[query],
                input_type="search_query",).embeddings[0]

  # 2. Retrieve the nearest neighbors
  distances , similar_item_ids = index.search(np.float32([query_embed]), number_of_results)

  # 3. Format the results
  texts_np = np.array(texts) # Convert texts list to numpy for easier indexing
  results = pd.DataFrame(data={'texts': texts_np[similar_item_ids[0]],
                              'distance': distances[0]})

  # 4. Print and return the results
  print(f"Query:'{query}'\nNearest neighbors:")
  return results
query = "how precise was the science"
results = search(query)
results
Query:'how precise was the science'
Nearest neighbors:
                                               texts      distance
0  It has also received praise from many astronom...  10757.379883
1  Caltech theoretical physicist and 2017 Nobel l...  11566.131836
2  Interstellar uses extensive practical and mini...  11922.833008
from rank_bm25 import BM25Okapi
from sklearn.feature_extraction import _stop_words
import string

def bm25_tokenizer(text):
    tokenized_doc = []
    for token in text.lower().split():
        token = token.strip(string.punctuation)

        if len(token) > 0 and token not in _stop_words.ENGLISH_STOP_WORDS:
            tokenized_doc.append(token)
    return tokenized_doc
from tqdm import tqdm

tokenized_corpus = []
for passage in tqdm(texts):
    tokenized_corpus.append(bm25_tokenizer(passage))

bm25 = BM25Okapi(tokenized_corpus)
100%|██████████| 15/15 [00:00<00:00, 38908.20it/s]
def keyword_search(query, top_k=3, num_candidates=15):
    print("Input question:", query)

    ##### BM25 search (lexical search) #####
    bm25_scores = bm25.get_scores(bm25_tokenizer(query))
    top_n = np.argpartition(bm25_scores, -num_candidates)[-num_candidates:]
    bm25_hits = [{'corpus_id': idx, 'score': bm25_scores[idx]} for idx in top_n]
    bm25_hits = sorted(bm25_hits, key=lambda x: x['score'], reverse=True)

    print(f"Top-3 lexical search (BM25) hits")
    for hit in bm25_hits[0:top_k]:
        print("\t{:.3f}\t{}".format(hit['score'], texts[hit['corpus_id']].replace("\n", " ")))
keyword_search(query = "how precise was the science")
Input question: how precise was the science
Top-3 lexical search (BM25) hits
    1.789   Interstellar is a 2014 epic science fiction film co-written, directed, and produced by Christopher Nolan
    1.373   Caltech theoretical physicist and 2017 Nobel laureate in Physics[4] Kip Thorne was an executive producer, acted as a scientific consultant, and wrote a tie-in book, The Science of Interstellar
    0.000   It stars Matthew McConaughey, Anne Hathaway, Jessica Chastain, Bill Irwin, Ellen Burstyn, Matt Damon, and Michael Caine

Caveats of Dense Retrieval

query = "What is the mass of the moon?"
results = search(query)
results
Query:'What is the mass of the moon?'
Nearest neighbors:
                                               texts      distance
0  Cinematographer Hoyte van Hoytema shot it on 3...  12854.458984
1  The film had a worldwide gross over $677 milli...  13301.030273
2  It has also received praise from many astronom...  13332.011719

Reranking Example

query = "how precise was the science"
results = co.rerank(query=query, documents=texts, top_n=3, return_documents=True)
results.results
[RerankResponseResultsItem(document=RerankResponseResultsItemDocument(text='It has also received praise from many astronomers for its scientific accuracy and portrayal of theoretical astrophysics'), index=12, relevance_score=0.1698185),
 RerankResponseResultsItem(document=RerankResponseResultsItemDocument(text='The film had a worldwide gross over $677 million (and $773 million with subsequent re-releases), making it the tenth-highest grossing film of 2014'), index=10, relevance_score=0.07004896),
 RerankResponseResultsItem(document=RerankResponseResultsItemDocument(text='Caltech theoretical physicist and 2017 Nobel laureate in Physics[4] Kip Thorne was an executive producer, acted as a scientific consultant, and wrote a tie-in book, The Science of Interstellar'), index=4, relevance_score=0.0043994132)]
for idx, result in enumerate(results.results):
    print(idx, result.relevance_score , result.document.text)
0 0.1698185 It has also received praise from many astronomers for its scientific accuracy and portrayal of theoretical astrophysics
1 0.07004896 The film had a worldwide gross over $677 million (and $773 million with subsequent re-releases), making it the tenth-highest grossing film of 2014
2 0.0043994132 Caltech theoretical physicist and 2017 Nobel laureate in Physics[4] Kip Thorne was an executive producer, acted as a scientific consultant, and wrote a tie-in book, The Science of Interstellar
def keyword_and_reranking_search(query, top_k=3, num_candidates=10):
    print("Input question:", query)

    ##### BM25 search (lexical search) #####
    bm25_scores = bm25.get_scores(bm25_tokenizer(query))
    top_n = np.argpartition(bm25_scores, -num_candidates)[-num_candidates:]
    bm25_hits = [{'corpus_id': idx, 'score': bm25_scores[idx]} for idx in top_n]
    bm25_hits = sorted(bm25_hits, key=lambda x: x['score'], reverse=True)

    print(f"Top-3 lexical search (BM25) hits")
    for hit in bm25_hits[0:top_k]:
        print("\t{:.3f}\t{}".format(hit['score'], texts[hit['corpus_id']].replace("\n", " ")))

    #Add re-ranking
    docs = [texts[hit['corpus_id']] for hit in bm25_hits]

    print(f"\nTop-3 hits by rank-API ({len(bm25_hits)} BM25 hits re-ranked)")
    results = co.rerank(query=query, documents=docs, top_n=top_k, return_documents=True)
    for hit in results.results:
        print("\t{:.3f}\t{}".format(hit.relevance_score, hit.document.text.replace("\n", " ")))
keyword_and_reranking_search(query = "how precise was the science")
Input question: how precise was the science
Top-3 lexical search (BM25) hits
    1.789   Interstellar is a 2014 epic science fiction film co-written, directed, and produced by Christopher Nolan
    1.373   Caltech theoretical physicist and 2017 Nobel laureate in Physics[4] Kip Thorne was an executive producer, acted as a scientific consultant, and wrote a tie-in book, The Science of Interstellar
    0.000   Interstellar uses extensive practical and miniature effects and the company Double Negative created additional digital effects

Top-3 hits by rank-API (10 BM25 hits re-ranked)
    0.004   Caltech theoretical physicist and 2017 Nobel laureate in Physics[4] Kip Thorne was an executive producer, acted as a scientific consultant, and wrote a tie-in book, The Science of Interstellar
    0.004   Set in a dystopian future where humanity is struggling to survive, the film follows a group of astronauts who travel through a wormhole near Saturn in search of a new home for mankind
    0.003   Brothers Christopher and Jonathan Nolan wrote the screenplay, which had its origins in a script Jonathan developed in 2007

Retrieval-Augmented Generation

Example: Grounded Generation with an LLM API

query = "income generated"

# 1- Retrieval
# We'll use embedding search. But ideally we'd do hybrid
results = search(query)

# 2- Grounded Generation
docs_dict = [{'text': text} for text in results['texts']]
response = co.chat(
    message = query,
    documents=docs_dict
)

print(response.text)
Query:'income generated'
Nearest neighbors:
The film generated a worldwide gross of over $677 million, or $773 million with subsequent re-releases.
response
NonStreamedChatResponse(text='The film generated a worldwide gross of over $677 million, or $773 million with subsequent re-releases.', generation_id='bebc32bd-d620-42cf-bd13-f8d1f96d4aa6', citations=[ChatCitation(start=21, end=57, text='worldwide gross of over $677 million', document_ids=['doc_0']), ChatCitation(start=62, end=103, text='$773 million with subsequent re-releases.', document_ids=['doc_0'])], documents=[{'id': 'doc_0', 'text': 'The film had a worldwide gross over $677 million (and $773 million with subsequent re-releases), making it the tenth-highest grossing film of 2014'}], is_search_required=None, search_queries=None, search_results=None, finish_reason='COMPLETE', tool_calls=None, chat_history=[Message_User(message='income generated', tool_calls=None, role='USER'), Message_Chatbot(message='The film generated a worldwide gross of over $677 million, or $773 million with subsequent re-releases.', tool_calls=None, role='CHATBOT')], prompt=None, meta=ApiMeta(api_version=ApiMetaApiVersion(version='1', is_deprecated=None, is_experimental=None), billed_units=ApiMetaBilledUnits(input_tokens=106, output_tokens=26, search_units=None, classifications=None), tokens=ApiMetaTokens(input_tokens=797, output_tokens=95), warnings=None), response_id='319002db-505d-4d3c-903f-41eefbf0f856')
response.citations
[ChatCitation(start=21, end=57, text='worldwide gross of over $677 million', document_ids=['doc_0']),
 ChatCitation(start=62, end=103, text='$773 million with subsequent re-releases.', document_ids=['doc_0'])]

Example: RAG with Local Models

Loading the Generation Model

!wget https://huggingface.co/microsoft/Phi-3-mini-4k-instruct-gguf/resolve/main/Phi-3-mini-4k-instruct-q4.gguf
--2024-06-21 09:49:22--  https://huggingface.co/lmstudio-community/Phi-3-mini-4k-instruct-GGUF/resolve/main/Phi-3-mini-4k-instruct-Q8_0.gguf
Resolving huggingface.co (huggingface.co)... 18.164.174.118, 18.164.174.23, 18.164.174.17, ...
Connecting to huggingface.co (huggingface.co)|18.164.174.118|:443... connected.
HTTP request sent, awaiting response... 302 Found
Location: https://cdn-lfs-us-1.huggingface.co/repos/8e/3f/8e3fafa0351929e621a3db9a53b131a9d7f4b222332208032555bb92f11ab100/8d2f3732e31c354e169cd81dcde9807a1c73b85b9a0f9b16c19013e7a4bb151c?response-content-disposition=inline%3B+filename*%3DUTF-8%27%27Phi-3-mini-4k-instruct-Q8_0.gguf%3B+filename%3D%22Phi-3-mini-4k-instruct-Q8_0.gguf%22%3B&Expires=1719222562&Policy=eyJTdGF0ZW1lbnQiOlt7IkNvbmRpdGlvbiI6eyJEYXRlTGVzc1RoYW4iOnsiQVdTOkVwb2NoVGltZSI6MTcxOTIyMjU2Mn19LCJSZXNvdXJjZSI6Imh0dHBzOi8vY2RuLWxmcy11cy0xLmh1Z2dpbmdmYWNlLmNvL3JlcG9zLzhlLzNmLzhlM2ZhZmEwMzUxOTI5ZTYyMWEzZGI5YTUzYjEzMWE5ZDdmNGIyMjIzMzIyMDgwMzI1NTViYjkyZjExYWIxMDAvOGQyZjM3MzJlMzFjMzU0ZTE2OWNkODFkY2RlOTgwN2ExYzczYjg1YjlhMGY5YjE2YzE5MDEzZTdhNGJiMTUxYz9yZXNwb25zZS1jb250ZW50LWRpc3Bvc2l0aW9uPSoifV19&Signature=DFqGVAGZBHhpPlWt2pSH5nbTgskxr6FrSK19FhopbgIYnqujHL3LDgnleMTDHrtRvflyX-6QE88ZAYZc1wIZA7ZXgBVoFdYMNPbHixfv4hqGdKiddWcF4QYi5JCYChS3Z8oZPUkcbyX6KNTqMR1nls2KTZ3K0Xl3E7nGmlTXo85mRdRKQojZLkuLOa28pG2z9jrs1wJ1B2W3Ed%7E%7EK1E-BXhjKsK4zUR1Ch-3KfqAqe0q0XmnCcF1Ml2xosujvlA%7EZuVxflmRf8wRVZBZsbZNXaUFAb3oiyNTuyr9g1fvSxJPmAJocs9jIMUZwU88Sa4s%7EVdfytD0s6YruNH1GOAfoA__&Key-Pair-Id=K2FPYV99P2N66Q [following]
--2024-06-21 09:49:22--  https://cdn-lfs-us-1.huggingface.co/repos/8e/3f/8e3fafa0351929e621a3db9a53b131a9d7f4b222332208032555bb92f11ab100/8d2f3732e31c354e169cd81dcde9807a1c73b85b9a0f9b16c19013e7a4bb151c?response-content-disposition=inline%3B+filename*%3DUTF-8%27%27Phi-3-mini-4k-instruct-Q8_0.gguf%3B+filename%3D%22Phi-3-mini-4k-instruct-Q8_0.gguf%22%3B&Expires=1719222562&Policy=eyJTdGF0ZW1lbnQiOlt7IkNvbmRpdGlvbiI6eyJEYXRlTGVzc1RoYW4iOnsiQVdTOkVwb2NoVGltZSI6MTcxOTIyMjU2Mn19LCJSZXNvdXJjZSI6Imh0dHBzOi8vY2RuLWxmcy11cy0xLmh1Z2dpbmdmYWNlLmNvL3JlcG9zLzhlLzNmLzhlM2ZhZmEwMzUxOTI5ZTYyMWEzZGI5YTUzYjEzMWE5ZDdmNGIyMjIzMzIyMDgwMzI1NTViYjkyZjExYWIxMDAvOGQyZjM3MzJlMzFjMzU0ZTE2OWNkODFkY2RlOTgwN2ExYzczYjg1YjlhMGY5YjE2YzE5MDEzZTdhNGJiMTUxYz9yZXNwb25zZS1jb250ZW50LWRpc3Bvc2l0aW9uPSoifV19&Signature=DFqGVAGZBHhpPlWt2pSH5nbTgskxr6FrSK19FhopbgIYnqujHL3LDgnleMTDHrtRvflyX-6QE88ZAYZc1wIZA7ZXgBVoFdYMNPbHixfv4hqGdKiddWcF4QYi5JCYChS3Z8oZPUkcbyX6KNTqMR1nls2KTZ3K0Xl3E7nGmlTXo85mRdRKQojZLkuLOa28pG2z9jrs1wJ1B2W3Ed%7E%7EK1E-BXhjKsK4zUR1Ch-3KfqAqe0q0XmnCcF1Ml2xosujvlA%7EZuVxflmRf8wRVZBZsbZNXaUFAb3oiyNTuyr9g1fvSxJPmAJocs9jIMUZwU88Sa4s%7EVdfytD0s6YruNH1GOAfoA__&Key-Pair-Id=K2FPYV99P2N66Q
Resolving cdn-lfs-us-1.huggingface.co (cdn-lfs-us-1.huggingface.co)... 18.65.25.71, 18.65.25.64, 18.65.25.113, ...
Connecting to cdn-lfs-us-1.huggingface.co (cdn-lfs-us-1.huggingface.co)|18.65.25.71|:443... connected.
HTTP request sent, awaiting response... 200 OK
Length: 4061221024 (3.8G) [binary/octet-stream]
Saving to: ‘Phi-3-mini-4k-instruct-Q8_0.gguf’

Phi-3-mini-4k-instr 100%[===================>]   3.78G  66.7MB/s    in 33s     

2024-06-21 09:49:55 (119 MB/s) - ‘Phi-3-mini-4k-instruct-Q8_0.gguf’ saved [4061221024/4061221024]

from langchain import LlamaCpp

# Make sure the model path is correct for your system!
llm = LlamaCpp(
    model_path="Phi-3-mini-4k-instruct-q4.gguf",
    n_gpu_layers=-1,
    max_tokens=500,
    n_ctx=2048,
    seed=42,
    verbose=False
)

Loading the Embedding Model

from langchain.embeddings.huggingface import HuggingFaceEmbeddings

# Embedding Model for converting text to numerical representations
embedding_model = HuggingFaceEmbeddings(
    model_name='BAAI/bge-small-en-v1.5'
)
/usr/local/lib/python3.10/dist-packages/huggingface_hub/file_download.py:1132: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
  warnings.warn(

Preparing the Vector Database

from langchain.vectorstores import FAISS

# Create a local vector database
db = FAISS.from_texts(texts, embedding_model)

The RAG Prompt

from langchain import PromptTemplate
from langchain.chains import RetrievalQA


# Create a prompt template
template = """<|user|>
Relevant information:
{context}

Provide a concise answer the following question using the relevant information provided above:
{question}<|end|>
<|assistant|>"""
prompt = PromptTemplate(
    template=template,
    input_variables=["context", "question"]
)

# RAG Pipeline
rag = RetrievalQA.from_chain_type(
    llm=llm,
    chain_type='stuff',
    retriever=db.as_retriever(),
    chain_type_kwargs={
        "prompt": prompt
    },
    verbose=True
)
rag.invoke('Income generated')


> Entering new RetrievalQA chain...

> Finished chain.
{'query': 'Income generated',
 'result': " Interstellar grossed over $677 million worldwide in 2014 and had additional earnings from subsequent re-releases, totaling approximately $773 million. The film's release utilized both traditional film stock and digital projectors across various venues to maximize its income generation potential."}
Back to top