Create Graph from Triples¶

Create a WhyHow graph from triples using LangChain LLMGraphTransformer

In [ ]:

import itertools
import os
import pickle

from langchain.text_splitter import RecursiveCharacterTextSplitter
from langchain_community.document_loaders import PyPDFLoader
from langchain_experimental.graph_transformers import LLMGraphTransformer
from langchain_openai import ChatOpenAI

from whyhow import WhyHow, Node, Relation, Triple

import itertools import os import pickle from langchain.text_splitter import RecursiveCharacterTextSplitter from langchain_community.document_loaders import PyPDFLoader from langchain_experimental.graph_transformers import LLMGraphTransformer from langchain_openai import ChatOpenAI from whyhow import WhyHow, Node, Relation, Triple

In [ ]:

from dotenv import find_dotenv, load_dotenv
load_dotenv(find_dotenv())

from dotenv import find_dotenv, load_dotenv load_dotenv(find_dotenv())

In [ ]:

llm = ChatOpenAI(model="gpt-4o")
llm_transformer = LLMGraphTransformer(llm=llm)

llm = ChatOpenAI(model="gpt-4o") llm_transformer = LLMGraphTransformer(llm=llm)

Load Text from Selected File¶

In [ ]:

filepath = "{YOUR FILEPATH}"

filepath = "{YOUR FILEPATH}"

In [ ]:

loader = PyPDFLoader(filepath)
docs = loader.load()

loader = PyPDFLoader(filepath) docs = loader.load()

Process Document¶

In [ ]:

text_splitter = RecursiveCharacterTextSplitter(chunk_size=2000, chunk_overlap=0)
split_docs = text_splitter.split_documents(docs)

text_splitter = RecursiveCharacterTextSplitter(chunk_size=2000, chunk_overlap=0) split_docs = text_splitter.split_documents(docs)

Convert Processed Text to Triples¶

In [ ]:

# Select the entity types and realtions you want for your triples
allowed_nodes=["Company", "Risk Factor", "Legal Proceeding", "Business Segment"]
allowed_relationships=["AFFECTS", "INVOLVED_IN", "WORKED_AT", "POSES_RISK"]

# Select the entity types and realtions you want for your triples allowed_nodes=["Company", "Risk Factor", "Legal Proceeding", "Business Segment"] allowed_relationships=["AFFECTS", "INVOLVED_IN", "WORKED_AT", "POSES_RISK"]

In [ ]:

llm_transformer_props = LLMGraphTransformer(
    llm=llm,
    allowed_nodes=updated_allowed_nodes,
    allowed_relationships=updated_allowed_relationships
)

llm_transformer_props = LLMGraphTransformer( llm=llm, allowed_nodes=updated_allowed_nodes, allowed_relationships=updated_allowed_relationships )

In [ ]:

graph_documents_props = await llm_transformer_props.aconvert_to_graph_documents(split_docs)

graph_documents_props = await llm_transformer_props.aconvert_to_graph_documents(split_docs)

In [ ]:

print(f"Nodes:{graph_documents_props[0].nodes}")
print(f"Relationships:{graph_documents_props[0].relationships}")

print(f"Nodes:{graph_documents_props[0].nodes}") print(f"Relationships:{graph_documents_props[0].relationships}")

In [ ]:

for i in range(10):
    print(graph_documents_props[i].relationships)

for i in range(10): print(graph_documents_props[i].relationships)

In [ ]:

triples = [chunk.relationships for chunk in graph_documents_props]
flat_triples = list(itertools.chain(*triples))

triples = [chunk.relationships for chunk in graph_documents_props] flat_triples = list(itertools.chain(*triples))

In [ ]:

flat_triples[0]

flat_triples[0]

[Optional] Store the Triples¶

In [ ]:

# Serialize the list and write it to the file
with open('langchain_triples_improved.pkl', 'wb') as file:
    pickle.dump(flat_triples, file)

# Serialize the list and write it to the file with open('langchain_triples_improved.pkl', 'wb') as file: pickle.dump(flat_triples, file)

WhyHow Integration¶

In [346]:

# Initialise the client with your WhyHow API key
client = WhyHow(api_key=os.environ["WHYHOW_API_KEY"], base_url="https://api.whyhow.ai")

# Initialise the client with your WhyHow API key client = WhyHow(api_key=os.environ["WHYHOW_API_KEY"], base_url="https://api.whyhow.ai")

Initialise the Workspace¶

In [331]:

workspace = client.workspaces.create(name="Amazon 10-K Testing")
# or, if you already have a workspace
# workspace = client.workspaces.get(workspace_id="<workspace_id>")

workspace = client.workspaces.create(name="Amazon 10-K Testing") # or, if you already have a workspace # workspace = client.workspaces.get(workspace_id="")

[Optional] Load the Triples¶

In [ ]:

with open('langchain_triples.pkl', 'rb') as file:
    flat_triples = pickle.load(file)

with open('langchain_triples.pkl', 'rb') as file: flat_triples = pickle.load(file)

In [ ]:

flat_triples[0]

flat_triples[0]

Preprocess the Triples¶

In [ ]:

def format_triple(triple):
    """
    Format the LangChain triple into the desired structure.

    Args:
        triple: An object containing source, target, and type attributes.

    Returns:
        Triple: A Triple object with formatted head, relation, and tail.
    """
    # Extract source and target from the triple
    source = triple.source
    target = triple.target
    
    # Create and return a formatted Triple object
    return Triple(
        head=Node(name=source.id, label=source.type),  # Head node with source id and type
        relation=Relation(name=triple.type),           # Relation with triple type
        tail=Node(name=target.id, label=target.type)   # Tail node with target id and type
    )

def format_triple(triple): """ Format the LangChain triple into the desired structure. Args: triple: An object containing source, target, and type attributes. Returns: Triple: A Triple object with formatted head, relation, and tail. """ # Extract source and target from the triple source = triple.source target = triple.target # Create and return a formatted Triple object return Triple( head=Node(name=source.id, label=source.type), # Head node with source id and type relation=Relation(name=triple.type), # Relation with triple type tail=Node(name=target.id, label=target.type) # Tail node with target id and type )

In [ ]:

# Generate a list of formatted triples with indices
formatted_triples = [format_triple(triple) for triple in flat_triples]

# Generate a list of formatted triples with indices formatted_triples = [format_triple(triple) for triple in flat_triples]

In [ ]:

# View the first 3 triples
formatted_triples[:1]

# View the first 3 triples formatted_triples[:1]

Create the graph¶

In [333]:

graph = client.graphs.create_graph_from_triples(
    workspace_id=workspace.workspace_id,
    triples=formatted_triples,
    name="Amazon 10-K Graph (triple id testing 5)"  
)

graph = client.graphs.create_graph_from_triples( workspace_id=workspace.workspace_id, triples=formatted_triples, name="Amazon 10-K Graph (triple id testing 5)" )

Query the graph¶

In [337]:

# Query graph for Amazon's business segments
question = "What are Amazons main business segments?"

query_response = client.graphs.query_unstructured(
    graph_id='66cbd0f8781bfda4a5a4e2ef',
    query=question,
)

# Query graph for Amazon's business segments question = "What are Amazons main business segments?" query_response = client.graphs.query_unstructured( graph_id='66cbd0f8781bfda4a5a4e2ef', query=question, )

In [338]:

print(f"LLM Response: {query_response.answer}")
print(f"Returned Triples: {query_response.triples}")

print(f"LLM Response: {query_response.answer}") print(f"Returned Triples: {query_response.triples}")

LLM Response: North America, Amazon Web Services (AWS), International.
Returned Triples: [Triple(triple_id='66cbd0fc781bfda4a5a4e751', head=Node(node_id='66cbd0f8781bfda4a5a4e2fc', label='Company', name='Amazon.Com, Inc.', chunk_ids=[], properties={}, created_at=None, updated_at=None), tail=Node(node_id='66cbd0f8781bfda4a5a4e300', label='Business segment', name='North America', chunk_ids=[], properties={}, created_at=None, updated_at=None), relation=Relation(name='INVOLVED_IN', properties={}), chunk_ids=[], created_at=None, updated_at=None), Triple(triple_id='66cbd0fc781bfda4a5a4e752', head=Node(node_id='66cbd0f8781bfda4a5a4e2fc', label='Company', name='Amazon.Com, Inc.', chunk_ids=[], properties={}, created_at=None, updated_at=None), tail=Node(node_id='66cbd0f8781bfda4a5a4e301', label='Business segment', name='International', chunk_ids=[], properties={}, created_at=None, updated_at=None), relation=Relation(name='INVOLVED_IN', properties={}), chunk_ids=[], created_at=None, updated_at=None), Triple(triple_id='66cbd0fc781bfda4a5a4e753', head=Node(node_id='66cbd0f8781bfda4a5a4e2fc', label='Company', name='Amazon.Com, Inc.', chunk_ids=[], properties={}, created_at=None, updated_at=None), tail=Node(node_id='66cbd0f8781bfda4a5a4e302', label='Business segment', name='Amazon Web Services (Aws)', chunk_ids=[], properties={}, created_at=None, updated_at=None), relation=Relation(name='INVOLVED_IN', properties={}), chunk_ids=[], created_at=None, updated_at=None)]