titus

Filter

Is there a way to instantiate metadata filtering as part of the hybrid search? This cookbook (https://docs.llamaindex.ai/en/stable/examples/vector_stores/Qdrant_metadata_filter/) requires specifying the filter ahead of time before instantiating the retriever so in production you're going to keep instantiating the query engine with different filters with each query. This tutorial on the other hand allows for instantiation of richer metadata (https://docs.llamaindex.ai/en/stable/examples/metadata_extraction/MetadataExtraction_LLMSurvey/) and I'm currently using the LLMQuestionGenerator -> SubQuestionQueryEngine strategy to try and get it to look at the metadata:

Plain Text

from llama_index.core.question_gen import LLMQuestionGenerator
from llama_index.core.question_gen.prompts import (
    DEFAULT_SUB_QUESTION_PROMPT_TMPL,
)


question_gen = LLMQuestionGenerator.from_defaults(
    llm=llm,
    prompt_template_str="""
        Follow the example, but instead of giving a question, always prefix the question 
        with: 'By first identifying and quoting the most relevant sources, '. 
        """
    + DEFAULT_SUB_QUESTION_PROMPT_TMPL,
)

But I was thinking if it was possible to have a triple hybrid search - dense, sparse embedding search and metadata search?

3 comments

ttitus

Pydantic Validation Error with Redis Document Store

I treid following this guide (https://docs.llamaindex.ai/en/stable/examples/metadata_extraction/DocumentContextExtractor/) using a Redis Document Store and got a pydantic validation error

Plain Text

ValidationError: 1 validation error for DocumentContextExtractor
docstore
  Input should be an instance of SimpleDocumentStore [type=is_instance_of, input_value=<llama_index.storage.docs...bject at 0x72deaa8a9850>, input_type=RedisDocumentStore]
    For further information visit https://errors.pydantic.dev/2.10/v/is_instance_of

Is there any reason why only SimpleDocumentStore is supported?

PS: I spun up my Redis using Docker

3 comments

ttitus

Llama Deploy Cookbook for Human-in-the-Loop Agents

Does anyone know whether there is a LlamaIndex Workflow + Llama Deploy cookbook on human in the loop agents?

1 comment

ttitus

Draw All Flows Utility Fails With InputRequiredEvent

Hey! The draw_all_flows utility fails when there's HITL (basically whenever there is an InputRequiredEvent)

Plain Text

```
File /opt/anaconda3/envs/llamaindex/lib/python3.12/site-packages/llama_index/utils/workflow/draw.py:62, in draw_all_possible_flows(workflow, filename, notebook)
     60 for return_type in step_config.return_types:
     61     if return_type != type(None):
---> 62         net.add_edge(step_name, return_type.__name__)
     64 for event_type in step_config.accepted_events:
     65     net.add_edge(event_type.__name__, step_name)

File /opt/anaconda3/envs/llamaindex/lib/python3.12/site-packages/pyvis/network.py:372, in Network.add_edge(self, source, to, **options)
    368 # verify nodes exists
    369 assert source in self.get_nodes(), \
    370     "non existent node '" + str(source) + "'"
--> 372 assert to in self.get_nodes(), \
    373     "non existent node '" + str(to) + "'"
    375 # we only check existing edge for undirected graphs
    376 if not self.directed:

AssertionError: non existent node 'InputRequiredEvent'

6 comments

ttitus

Add longllmlingua2 integration by titusl...

Hello! I took a stab at sprucing up LlamaIndex's integration with llmlingua (to include integration with llmlingua2) but the linting action failed at the MakeFile xD

Not sure how to clear it - anyone got an idea?

https://github.com/run-llama/llama_index/pull/17531

15 comments

ttitus

upgrade dependencies of llama-extract, llama-deploy and llama-index-multi-modal-llms-ollama

Hey @Logan M! Remember to upgrade the dependencies of llama-extract, llama-deploy and llama-index-multi-modal-llms-ollama! They're still pointing to llama-index==0.11

2 comments

ttitus

[Bug]: Neo4j Property Graph Index doesn'...

Has anyone faced any issues with neo4j lately?

https://github.com/run-llama/llama_index/issues/16352

14 comments

ttitus

Flows

Has anyone tried the new LlamaIndex Workflow abstraction?

I find it quite interesting because using the Workflow abstraction requires for developers to be quite familiar with LlamaIndex's lower level abstractions (for e.g. llm.get_tool_calls_from_response(), manually load in chat conversations into memory, etc). Most of us will probably be aware of the RAG abstractions but not the agent ones because it's always just agent.chat(query). The ReAct agent example underscores the biggest difference: ReActAgent.from_tools() is just one line vs writing an entire ReAct agent workflow. If I were to write an agent using multiple RAG tools, would I have to write nested workflows?

I've not had the need to build agents from low level abstractions yet. Does the LlamaIndex team have a use case in mind when designing the "Workflow" abstraction?

5 comments

ttitus

For now I think only LLMs can return

For now I think only LLMs can return structured outputs. If I wanted to get an agent to return a structured output, my only way of doing so would be to use tool = FunctionTool.from_defaults(fn={fn}, return_direct=True) right?

Except that I need to be mindful to ensure that my tool output is a string so that I won't mess up the AgentChatResponse output class. So if I specify a pydantic base model as my desired return, I would have to store the key value pairs of the response model as a dictionary then json dumps that dictionary as my return, and then add postprocessing code to handle the return.

Of course the other way would be to write a custom agent using Workflow and handle the ToolOutput there.

1 comment

ttitus

QueryEngineTool returns "empty response"...

Is anyone facing this: https://github.com/run-llama/llama_deploy/issues/250

I'm not sure why but the moment I deployed my workflow I started having issues with my rag query engine tool

10 comments

ttitus