Metadata-Version: 2.1
Name: canvas_langchain
Version: 0.4
Summary: A canvas langchain integration
Home-page: https://github.com/umich-its-ai/langchain-doc-canvas
Author: University of Michigan
Author-email: noreply@umich.edu
Classifier: Development Status :: 3 - Alpha
Classifier: License :: OSI Approved :: GNU General Public License (GPL)
Requires-Python: >=3.8.1
Description-Content-Type: text/markdown
License-File: LICENSE.txt

# 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):

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

### Install Dependencies

```bash
pip install -r requirements.txt
```

## Usage example:

```python
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.
