gpp-client 25.9.0

Gemini Program Platform client.

GPP Client

---

Pain-free python/CLI communication with the Gemini Program Platform (GPP).

Documentation: https://gpp-client.readthedocs.io/en/latest/

Source Code: https://github.com/gemini-hlsw/gpp-client

---

Python client and CLI for the GPP. Key features:

• Type‑safe GraphQL
  • Pydantic models from the GPP GraphQL schema, so every query, mutation, and input type is validated.
• Resource Managers
  • High‑level Manager classes (e.g. GPPClient.program, GPPClient.observation) with convenient get_by_id, get_all, create, update_by_id, delete_by_id, restore_by_id and more methods, no need to write raw GraphQL.
• Flexible payloads
  • Create or update via in‑memory Pydantic inputs or from_json files.
• gpp CLI
  • Full CRUD surface on the command line: gpp <resource> list|get|create|update|delete, with rich table output and JSON export options.

Requirements

• python>=3.10
• toml
• typer
• ariadne-codegen

Development Status

🚧 Alpha: the library is under heavy development. The public API and CLI flags may change between releases.

Installation

pip install gpp-client

Quickstart

from gpp_client import GPPClient

# Initialize with your GraphQL endpoint and credentials.
client = GPPClient(url="YOUR_URL", token="YOUR_TOKEN")

# List the first 5 program notes.
notes = await client.program_note.get_all(limit=5)
for note in notes["matches"]:
    print(f"{note['id']}: {note['title']}")

# Create a new note from a JSON file.
new_note = await client.program_note.create(
    from_json="path/to/program_note_payload.json",
    program_id="p-123"
)
print("Created:", new_note)

# Or create a note from the pydantic model.
from gpp_client.api.enums import Existence
from gpp_client.api.input_types import ProgramNotePropertiesInput

properties = ProgramNotePropertiesInput(
    title="Example",
    text="This is an example.",
    is_private=False,
    existence=Existence.PRESENT
)
another_note = await client.program_note.create(properties=properties, program_id="p-123")

print("Created another:", another_note)

As a CLI

# Get help.
gpp --help

# Get observation help.
gpp obs --help

# List observations.
gpp obs list --limit 3

# Get details for one.
gpp obs get o-123

# Create via JSON.
gpp obs create --from-json new_obs.json --program-id p-123

# Update by ID via JSON.
gpp obs update --observation-id o-123 --from-json updated_obs.json

Reporting Bugs and Feature Requests

Jira: https://noirlab.atlassian.net/jira/software/projects/GPC/boards/162

NOIRLab Slack channel: #gpp-client

While in heavy development, please file requests or report bugs via our Jira board or Slack channel.

Developer Notes

To update the GPP GraphQL schema and generate client code, run the scripts from the project’s top-level directory located in scripts/.

This project uses uv to manage dependencies and execute scripts.

When using uv run, a temporary virtual environment is created with only the dependencies required to run the script, based on the groups defined in pyproject.toml.

There’s no need to manually create or activate a virtual environment, uv handles everything.

Set Up pre-commit

To install pre-commit using uv, run:

uv tool install pre-commit --with pre-commit-uv

You may be prompted to add .local/bin to your PATH, uv installs tools there by default.

Next, install the hooks defined in .pre-commit-config.yaml:

pre-commit install

Once installed, pre-commit will automatically run the configured hooks each time you make a commit. This helps catch formatting issues, docstring violations, and other problems before code is committed.

To manually run all pre-commit hooks on the entire codebase:

pre-commit run --all-files

Download the Schema

This script downloads the latest GPP GraphQL schema to schema.graphql. You must have GPP_URL and GPP_TOKEN env variables set for downloading to work.

uv run --group schema python scripts/download_schema.py

Run Codegen

This script regenerates the client code based on the updated schema.

uv run --group codegen python scripts/run_codegen.py

Creating and Deploying a Release

Releases are managed using GitHub Actions. When you’re ready to publish a new version of the package, use the Create Release workflow.

1. Go to the Actions tab on GitHub.
2. Select the Create Release workflow from the sidebar.
3. Click Run workflow.
4. Enter the release version (e.g., 25.6.0). Note: Do not include a leading v or unnecessary zero padding.
5. Click Run workflow to trigger the release.

This workflow performs the following steps:

• Updates the version field in pyproject.toml.
• Updates the uv.lock file to reflect the new version.
• Commits and pushes the changes to the repository.
• Creates a Git tag.
• Drafts a GitHub release.

Finalizing the Release

After the workflow completes:

1. Go to the Releases section on GitHub.
2. Locate the newly created draft release.
3. Click Publish release.

Once published, the package will be automatically uploaded to PyPI. It may take a few minutes for the release to appear.