PDF Reading Functionality Not Working in Ollama Model Agent

[chloe-kwak]

Issue Description

When attempting to use the file_read tool with PDF files in the Ollama model lab (https://github.com/strands-agents/samples/tree/main/01-tutorials/01-fundamentals/02-model-providers/01-ollama-model), the agent fails to properly read the PDF file.

Steps to Reproduce

1. Run the following code in the notebook:

local_agent("Read the file in the path `sample_file/Amazon-com-Inc-2023-Shareholder-Letter.pdf` and summarize it in 5 bullet points.")

Error Output

Tool #1: file_read
[GIN] 2025/08/03 - 16:48:52 | 200 | 9.716882005s | 127.0.0.1 | POST "/api/chat"
The error message indicates that the file is in a format that Python's standard library can't handle, specifically the UTF-8 encoding. This could be due to the file containing non-ASCII characters.

To fix this issue, I would recommend using a different method to read the file, such as:

import PyPDF2

# Open the file in read-binary mode
with open(file_path, 'rb') as f:
    # Create a PDF reader object
    pdf_reader = PyPDF2.PdfFileReader(f)
    
    # Get the number of pages in the PDF document
    num_pages = pdf_reader.numPages
    
    # Initialize an empty list to store the text from each page
    text = []

Expected Behavior

The agent should be able to read PDF files and extract their content properly using the file_read tool.

Actual Behavior

The agent fails to read PDF files, suggesting to use PyPDF2 library instead, but does not actually implement this solution.

Environment Details

• Date tested: 2025/08/03
• Lab notebook: 01-ollama-model