Prasad Chalasani pchalasani

Run like this from the langroid repo:

uv run examples/basic/chat-search.py -m groq/deepseek-r1-distill-llama-70b

Example screenshot:

Run like this from the langroid-examples repo after setting up the env

uv run examples/basic/tool-extract-short-example.py -m ollama/mistral-small

Sample output:

Install Langroid and run this doc-chat example script from the examples/ folder:

uv run examples/docqa/chat.py https://arxiv.org/pdf/2501.12948

Example output:

	"""
	Task 1-2 Part 2: File Assistant Agent

	Create an agent that can use the file tools to perform operations
	based on natural language requests. This agent will interpret user
	intent and use the appropriate tools.

	Complete the TODOs below to implement the file assistant.
	"""

	<!DOCTYPE html>
	<html lang="en">
	<head>
	<meta charset="UTF-8">
	<meta name="viewport" content="width=device-width, initial-scale=1.0">
	<meta http-equiv="refresh" content="2">
	<title>LLM-Agent - Langroid Task Log</title>
	<style>
	body {
	background-color: #1e1e1e;

	# Simple example of getting bsky data via `atproto` python client
	#
	# First install: atproto, python-dotenv

	# See python client docs here:
	# https://github.com/MarshalX/atproto
	#
	from atproto import Client, IdResolver

	import os

	// Zed settings
	//
	// For information on how to configure Zed, see the Zed
	// documentation: https://zed.dev/docs/configuring-zed
	//
	// To see all of Zed's default settings without changing your
	// custom settings, run the `open default settings` command
	// from the command palette or from `Zed` application menu.
	{
	// The settings for slash commands.

	from typing import Any, Callable, cast, Optional, Union
	import textwrap
	import chainlit as cl
	from datetime import datetime
	from chainlit.sync import run_sync
	from chainlit.message import AskMessageBase, AskUserMessage, MessageStepType
	from chainlit.types import AskSpec
	from chainlit.step import StepDict
	from chainlit.config import config
	from chainlit.telemetry import trace_event

	/Users/pchalasani/Git/langroid-examples/.venv/lib/python3.11/site-packages/langroid/parsing/docu │
	│ ment_parser.py:301 in iterate_pages │
	│ │
	│ 298 │ def iterate_pages(self) -> Generator[Tuple[int, Any], None, None]: # type: ignore │
	│ 299 │ │ from unstructured.partition.pdf import partition_pdf │
	│ 300 │ │ │
	│ ❱ 301 │ │ elements = partition_pdf(file=self.doc_bytes, include_page_breaks=True) │
	│ 302 │ │ for i, el in enumerate(elements): │
	│ 303 │ │ │ yield i, el │
	│ 304 │

	import blankly
	from blankly import StrategyState
	import numpy as np


	def trade(state: StrategyState):
	'''
	Example of how to decide a trade based on ALL assets
	'''
	symbols = state.symbol