|
|
import os |
|
|
from typing import Optional |
|
|
|
|
|
from smolagents import LiteLLMModel, CodeAgent, DuckDuckGoSearchTool, PythonInterpreterTool, VisitWebpageTool |
|
|
|
|
|
from agentcourse_unit4.tools.audio_transcriber import AudioTranscriberTool |
|
|
from agentcourse_unit4.tools.chess_board_recognizer import ChessBoardRecognizerTool |
|
|
from agentcourse_unit4.tools.chess_predictor import ChessPredictorTool |
|
|
from agentcourse_unit4.tools.csv_reader import CsvReaderTool |
|
|
from agentcourse_unit4.tools.excel_reader import ExcelReaderTool |
|
|
from agentcourse_unit4.tools.file_downloader import FileDownloaderTool |
|
|
from agentcourse_unit4.tools.image_describer import ImageDescriberTool |
|
|
from agentcourse_unit4.tools.image_text_extractor import ImageTextExtractorTool |
|
|
from agentcourse_unit4.tools.pdf_reader import PdfReaderTool |
|
|
from agentcourse_unit4.tools.py_code_interpreter import PyCodeInterpreterTool |
|
|
from agentcourse_unit4.tools.youtube_transcriber import YoutubeTranscriberTool |
|
|
|
|
|
|
|
|
class BasicAgent: |
|
|
def __init__(self): |
|
|
self.model = LiteLLMModel( |
|
|
|
|
|
model_id="gemini/gemini-2.5-flash-preview-04-17", |
|
|
api_key=os.getenv("GOOGLE_API_KEY"), |
|
|
max_tokens=8196, |
|
|
temperature=0.9 |
|
|
) |
|
|
self.basic_agent = CodeAgent( |
|
|
name="basic_agent", |
|
|
description="Basic code agent", |
|
|
tools=[ |
|
|
PythonInterpreterTool(), |
|
|
DuckDuckGoSearchTool(max_results=5), |
|
|
VisitWebpageTool(max_output_length=1_000_000), |
|
|
FileDownloaderTool(), |
|
|
ExcelReaderTool(), |
|
|
CsvReaderTool(), |
|
|
PdfReaderTool(), |
|
|
PyCodeInterpreterTool(), |
|
|
YoutubeTranscriberTool(), |
|
|
AudioTranscriberTool(), |
|
|
ChessBoardRecognizerTool(), |
|
|
ChessPredictorTool(), |
|
|
ImageDescriberTool(), |
|
|
ImageTextExtractorTool() |
|
|
], |
|
|
model=self.model, |
|
|
add_base_tools=False, |
|
|
additional_authorized_imports=["pandas", "numpy", "datetime", "json", "csv"], |
|
|
planning_interval=None, |
|
|
verbosity_level=1, |
|
|
max_steps=5, |
|
|
max_print_outputs_length=1_000_000 |
|
|
) |
|
|
print("==> Agent initialized.") |
|
|
|
|
|
def run(self, question: str, file_path: Optional[str] = None) -> str: |
|
|
""" |
|
|
Process the incoming question and then return the answer. |
|
|
|
|
|
Args: |
|
|
question: The question or task |
|
|
file_path: Optional path to a file associated with the question or task |
|
|
|
|
|
Returns: |
|
|
The final answer to the question |
|
|
""" |
|
|
|
|
|
associated_file_prompt = f"\nAssociated file at path: {file_path}\n\nReading file content by proper tool is mandatory." if file_path else '' |
|
|
|
|
|
prompt = f""" |
|
|
Question: |
|
|
\"{question}\" |
|
|
{associated_file_prompt} |
|
|
|
|
|
Remember that you answer to the question from GAIA benchmark. It requires short, exact and precise answer. |
|
|
Don't include: thinking, explanations, steps, reasoning, intermediate or additional text. |
|
|
|
|
|
Finish your answer with a number OR as few words as possible OR a comma separated list of numbers and/or strings. |
|
|
If you are asked for a number, don't use comma to write your number neither use units such as $ or percent sign unless specified otherwise. |
|
|
If you are asked for a string, don't use articles, neither abbreviations (e.g. for cities), and write the digits in plain text unless specified otherwise. |
|
|
If you are asked for a comma separated list, apply the above rules depending of whether the element to be put in the list is a number or a string. |
|
|
|
|
|
For instance, if question is "What is the capital of Spain?", respond with "Madrid". |
|
|
It is exact and expected answer. |
|
|
""" |
|
|
return self.basic_agent.run(prompt) |
|
|
|
