File Management

gllm-inference | Tutorial: File Management | API Reference

Supported by: GoogleLMInvoker

What is file management?

File management is a feature that allows the language model to manage uploaded files in their server side. These files can then be used as inputs during invocations.

File management is only available for certain LM invokers. This feature can be accessed via the file attribute of the LM invoker. As an example, let's try to upload a file using GoogleLMInvoker and then use it as an input during invocation.

Init an LM Invoker

First of all, let's create a GoogleLMInvoker that we will use to manage the data store:

from dotenv import load_dotenv
load_dotenv()

from gllm_inference.lm_invoker import GoogleLMInvoker

lm_invoker = GoogleLMInvoker("gemini-2.5-flash-lite")

Upload a File

Next, let's upload a file. The upload() method will output an UploadedAttachment object to be used in later operations.

from gllm_inference.schema import Attachment

file = Attachment.from_path('path/to/file.pdf')
uploaded_file = await lm_invoker.file.upload(attachment)

List the Files

We can verify that the file has been successfully uploaded to the server side by using the list() method.

files = await lm_invoker.file.list()

if not files:
    print("No files found.")

for file in files:
    print(f" - {file}")

Invoke LM invoker with Uploaded File

Then, we can use the uploaded file as inputs during invocation.

inputs = [uploaded_file, "Explain this file in a single sentence"]
output = await lm_invoker.invoke(inputs)
print(f"output:\n{output}")

Delete the File

Finally, if the file is no longer used, it can be deleted via the delete() method.

await lm_invoker.file.delete(uploaded_file)

Troubleshooting

If you encounter errors, refer to the Troubleshooting Guide for detailed explanations of common errors and how to resolve them.