canvas-langchain 0.9.3

A canvas langchain integration

Canvas langchain document loader

Features:

Indexes Canvas Modules, Pages, Announcements, Assignments and Files

The following file types are supported: md htm html docx xls xlsx pptx pdf rtf txt

(doc support would require libreoffice, so has not been implemented in this library)

Running locally (development)

You can build/run the provided Dockerfile, or install dependencies as described below

Docker

Edit canvas-test.py, fill in the correct api_url, api_key, and course_id.

Run (this also builds docker):

docker run -it $(docker build -q .)

Install Dependencies

pip install -r requirements.txt

Usage example:

from canvas_langchain.canvas import CanvasLoader

loader = CanvasLoader(
	api_url = "https://canvas.instructure.com",
	api_key = "API_KEY_GOES_HERE",
	course_id = 123456789
)

documents = loader.load()

print("\nDocuments:\n")
print(documents)

print("\nInvalid files:\n")
print(loader.invalid_files)
print("")

print("\nErrors:\n")
print(loader.errors)
print("")

print("\nIndexed:\n")
print(loader.indexed_items)
print("")

If errors are present, loader.errors will contain one list element per error. It will consist of an error message (key named message) and if the error pertains to a specific item within canvas, it will list the entity_type and the entity_id of the resource where the exception occurred.