dokuments

The python client for dokuments, the Arkitekt document service: files uploaded to the deployment's S3 datalayer, and documents made of pages, each page an image with its OCR result (text lines and page rotation).

Installation

pip install dokuments

Usage

Every dokuments operation is a method of the Dokuments client, in a blocking and an a-prefixed async flavour: from_file_like (upload a file), get_file, search_files, watch_files, create_document, create_page, request_file_access, ….

In an arkitekt app

The client is injected by annotation. Add the service to your app and ask for dokuments: Dokuments:

from arkitekt import App, run
from dokuments import Dokuments, dokuments_service
from dokuments.api.schema import File

app = App("scanner", "0.1.0", services=[dokuments_service])


@app.action
def upload_scan(path: str, dokuments: Dokuments) -> File:
    """Upload Scan"""
    return dokuments.from_file_like(file=path)


if __name__ == "__main__":
    run(app)

Files travel between actions by id (@dokuments/file), so an action can take and return them directly.

From a script

from arkitekt import easy
from dokuments import dokuments_service

with easy("my-script", dokuments_service) as dokuments:
    file = dokuments.from_file_like(file="scan.pdf")  # a path, a Path or a binary file object
    document = dokuments.create_document(title="My scan", file=file.id)
    print(document.id, file.store.presigned_url)

Uploads go straight to the datalayer: the client asks the server for upload credentials and writes the file to S3 before the mutation is sent.

Development

The generated API (dokuments/api/schema.py) comes from the checked-in schema.graphql and the documents in graphql/; see graphql.config.yaml.