Flai SDK

The Flai SDK is a command-line toolset for interacting with the Flai web app. It is also used by our QGIS plugin.

Install

From PyPI

pip install flai-sdk

From source

1. Clone the repo and enter the directory:

   git clone https://github.com:flai-ai/flai-sdk.git
   cd flai-sdk
   

2. (Optional) Create and activate a virtual environment.

   With Conda

   conda create --name flai-sdk
   conda activate flai-sdk
   

   With venv

   python3(.VERSION) -m venv ~/.python-virtual-env/flai-sdk
   source ~/.python-virtual-env/flai-sdk/bin/activate
   

3. Install in “editable” mode in it's directory:

   pip install -e .

Developer setup

After cloning, enable the repo's git hooks (enforces semver on tags/releases):

git config core.hooksPath .githooks

First-time setup

To use our package you will need to call login command and pass at least two arguments for authorization (see Tips below).

flai-sdk login

This command will create a json file under path $HOME/.flai (Linux / macOS / Windows - Powershell) or %USERPROFILE%/.flai (Windows - cmd).

Tips:

• flai_access_token
  • can be found on Flai web app or
  • by going to our Web App at flai.ai > Sign in > Click your icon > Select Settings from menu > Click Access tokens tab > Click Add new personal access token
• flai_host
  • set it to https://api.flai.ai/

Examples

Upload a dataset to a project

Upload all .las/.laz files in the current folder to a project.

# simple example
flai-sdk upload-dataset --project_id PROJECT_NAME --dataset_name DATASET_NAME FILES

# working example
flai-sdk upload-dataset --project_id "Staring project" --dataset_name "Testing upload" "*.la?"

# compact version
flai-sdk upload-dataset -p "Staring project" -n "Testing upload" "*.la?"

The flag --project_id / -p can be omitted. In this case, the data will not be assigned to a project and will be shown in the catalog on your instance of our web app.

Download dataset

flai-sdk download-dataset -d "59d66bfc-c2a4-4e91-91f3-2f469078297b" "downloaded files.zip"

Explore commands

To see what arguments our SDK expects you can start from initial command and then add --help to see more hints.

# to see all available options
flai-sdk --help

# let's say we want to know what option 'download-flainet-model' expects
flai-sdk download-flainet-model --help

# and that is it :)

TODO

Cortex

Rewrite typed classes to pydantic model to match rest of flai-cortex. https://gitlab.com/flai-ai/dev/flai-cortex/-/work_items/35