Metadata-Version: 2.1
Name: prompt-sender
Version: 0.2.1
Summary: Send prompts and model responses to the Argilla A/B testing platform for annotations
License: Apache 2.0 License
Author: Hannah Galbraith
Author-email: hannah.galbraith@bsc.es
Requires-Python: >=3.10,<3.13
Classifier: License :: Other/Proprietary License
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3.10
Classifier: Programming Language :: Python :: 3.11
Classifier: Programming Language :: Python :: 3.12
Requires-Dist: argilla (>=2.1.0,<3.0.0)
Requires-Dist: typer (>=0.12.5,<0.13.0)
Description-Content-Type: text/markdown

# prompt-sender

A tool for sending JSON files of prompts and model responses to the Argilla A/B testing platform
for human annotations.

## How to use

Simply install the package using pip:

```
pip install prompt-sender
```

Run `prompt-sender --help` for the required arguments:

```
$ prompt-sender --help
                                                                                                                                                                                  
 Usage: prompt-sender [OPTIONS]                                                                                                                                                   
                                                                                                                                                                                  
 Send prompts that need annotations to the A/B testing platform. Make sure WORKSPACE_NAME and DATASET_NAME already exist on the platform!                                         
                                                                                                                                                                                  
╭─ Options ──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────╮
│ *  --api-url                   TEXT  A/B testing platform URL [default: None] [required]                                                                                       │
│ *  --api-key                   TEXT  API key for testing platform [default: None] [required]                                                                                   │
│ *  --prompt-file               TEXT  The name of the file of prompts to send to the A/B testing platform [default: None] [required]                                            │
│ *  --workspace-name            TEXT  The name of the workspace in the Argilla A/B testing platform [default: None] [required]                                                  │
│ *  --dataset-name              TEXT  The dataset to add prompts to [default: None] [required]                                                                                  │
│    --install-completion              Install completion for the current shell.                                                                                                 │
│    --show-completion                 Show completion for the current shell, to copy it or customize the installation.                                                          │
│    --help                            Show this message and exit.                                                                                                               │
╰────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────╯
```

## How to develop

1. Clone the repo: `git clone git@github.com:langtech-bsc/prompt-sender.git`.

2. If `poetry` is not installed on your system, follow the instructions to install it:
    - Install `pipx` using the [official installation instructions](https://pipx.pypa.io/stable/installation/)
    - `pipx install poetry`

3. `cd prompt-sender/` and create a virtual environment, either by using something like `venv` or by using `poetry`'s
built-in virtual environment, which can be created and activated by running `poetry shell` (run `exit` to deactivate).

4. Run `poetry install` to install project dependencies.

5. Run `pre-commit install` to set up the git pre-commit hook scripts
