What is the best vision model currently that performs as well as gpt4o? That is an LLM with vision modality.

does LI have a solution to managing the size of a page to send the oai image model?
to manage resolution vs. tokens?

InvalidRequestError: The model text-davinci-003 has been

Does title extractor work with version .8.43

I may be doing something wrong with respect to getting line numbers.

I am not getting the most accurate citations when it comes down to the line number. The page numbers seem spot on. Is there an art to fetching the line number that is significantly different than getting the page number?

Does llama index support agent that can read data from and write data to a SQL db (postgres , sqlite, etc.)?

I'm guessing that the they are now

from llama_index.readers.file import PyMuPDFReader
and this
from llama_index.core.node_parser import SentenceSplitter respectively

what happened to this:

from llama_index.vector_stores import PineconeVectorStore vector_store = PineconeVectorStore(pinecone_index=pinecone_index)

ModuleNotFoundError: No module named 'openai.openai_object'

any idea how I can extend this to PDFs instead of web pages?

@Logan M bro are LI agents in any way easier to use than langchain 🙂

I can't believe I'm asking this, but I have no idea now how to call the open ai api any longer

does anyone know the solution to this very quickly. This no longer works/ deprecated

response = openai.ChatCompletion.create(
model='gpt-4-1106-preview',
messages=messages
)

question - sometimes I see products featured on Llama Index thread on Linkedin. How can we showcase a prodcut?

What do you guys think of how Gemini will change things?

do you guys have any solution to sending a pdf to gpt vision api?

I'm writing an article about my experience with Llama index, amongst other technologies. I am going to post it tomorrow, most likely - will have a link to your site 🙂

While not in the millions of followers - I do have 10's of thousands so let's see what happens. Any advice about how to call you guys out (in addition to a link to your site?) of which I should be mindful?

I am looking for API docs for llama index in PDF form

for the sec-insights UI, can I just use my own RAG code, and leverage the UI as just a front-end?

I am able to get the knowledge graph page citations. but I can't get the document name to print:

line_number = 1
documents = []
for doc_idx, (chunk, page_number) in enumerate(chunks):
line_count_in_chunk = chunk.count('\n') + 1
metadata = {
"source_doc_idx": doc_idx,
"filename": "1.2.2.2 Customer Contract - Stockwood Dr - Woodstock - GA.pdf",
"page_number": page_number,
"document_title": "Customer Contract - Stockwood Dr - Woodstock - GA",
"line_count": line_count_in_chunk,
"starting_line_number": line_number
}
documents.append(Document(text=chunk, metadata=metadata))
line_number += line_count_in_chunk

for doc in documents:
print(f"Document Metadata: {doc.metadata}")

service_context = ServiceContext.from_defaults(
llm=OpenAI(model="gpt-3.5-turbo", temperature=0)
)

file_path = "./1.2.2.2 Customer Contract - Stockwood Dr - Woodstock - GA.pdf"
file_name = os.path.basename(file_path)

try:
index = VectorStoreIndex.from_documents(nodes, service_context=service_context)
except Exception as e:
print(f"Error: {e}")

query_engine = CitationQueryEngine.from_args(
index,
similarity_top_k=3,
citation_chunk_size=512,
)

response = query_engine.query("what is the purchase commitment?")
print("Query Response:", response)

G = nx.Graph()
for i, source_node in enumerate(response.source_nodes):
node_content = source_node.node.get_text()
citation_page = documents[i].metadata['page_number']
G.add_node((file_name, citation_page), content=node_content)

I then leverage llama-index, and get bogus page number