Skip to main content

Terminal-first workflows for Hephaes datasets

Project description

Hephaes CLI

hephaes is the official command-line interface for Hephaes — a version-controlled dataset platform for robotics and machine learning teams. Use it to upload, version, and download datasets without leaving your terminal.

Install

pip install hephaes-cli

Requires Python 3.10+.

Quick Start

# 1. Log in
hephaes login

# 2. Create a repo (or select an existing one)
hephaes repo create "My Dataset" --format parquet
hephaes repo use my-dataset

# 3. Stage files, commit, and push
hephaes add data/train.parquet data/val.parquet
hephaes commit -m "Initial upload"
hephaes push

# 4. Download on another machine
hephaes repo use my-dataset
hephaes pull --target ./dataset

Commands

Authentication

hephaes login                   # Authenticate (prompts for email + password)
hephaes logout                  # Remove the local session
hephaes whoami                  # Show the currently authenticated user

Credentials are stored securely in the OS keychain (macOS Keychain, Windows Credential Manager, etc).

Repositories

hephaes repo list                          # List all repos you have access to
hephaes repo create "Name" --format parquet  # Create a new repo
hephaes repo use <slug>                    # Set the active repo for subsequent commands
hephaes repo current                       # Show the active repo
hephaes repo info [<slug>]                 # Show details for a repo

Staging & Versioning

Hephaes uses a git-like workflow: stage files, create a local commit manifest, then push to the remote.

hephaes add <file> [<file> ...]   # Stage files for the next commit
hephaes status                    # Show staged files
hephaes reset [<file> ...]        # Unstage files (or clear everything)
hephaes commit -m "message"       # Create a local commit manifest
hephaes push                      # Upload staged files and publish the commit

Inspecting Commits

hephaes log                       # List commits for the active repo
hephaes ls                        # List files in the latest commit
hephaes metrics                   # Show dataset-level metrics for the latest commit
hephaes episodes                  # List episodes in the latest commit
hephaes episode-metrics           # Show per-episode metrics
hephaes jobs                      # Show processing jobs for the active repo
hephaes activity                  # Show recent activity for the active repo
hephaes activity --all            # Show activity across all repos

Pass --commit <id> to any of these to inspect a specific commit rather than HEAD.

Downloading

hephaes pull                          # Download all files from the latest commit
hephaes pull --target ./data          # Download into a specific directory
hephaes pull --commit <id>            # Download a specific commit
hephaes pull --path train.parquet     # Download a single file

Collaborators

hephaes repo members list                                      # List collaborators
hephaes repo members add --email user@example.com --role editor  # Invite by email
hephaes repo members add --user-id <uuid> --role viewer          # Invite by user ID
hephaes repo members update <user-id> --role viewer              # Change role
hephaes repo members remove <user-id>                            # Remove collaborator

Roles: owner (full control), editor (upload/commit, no member management), viewer (read-only).

Account

hephaes account usage    # Show your plan, storage usage, and repo limits

Options

Most commands accept --repo <slug> to target a specific repo without switching the active context.

All commands support --output json (or -o json) for machine-readable output, with non-zero exit codes and structured error payloads on failure:

hephaes --output json repo list
hephaes --output json log --repo my-dataset

CI / Headless Use

For non-interactive environments (GitHub Actions, Docker, etc.), inject a token instead of running login:

export HEPHAES_ACCESS_TOKEN="<token>"
hephaes --output json whoami

Injected tokens are never written to disk or the keychain.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

hephaes_cli-0.1.0.tar.gz (36.3 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

hephaes_cli-0.1.0-py3-none-any.whl (29.3 kB view details)

Uploaded Python 3

File details

Details for the file hephaes_cli-0.1.0.tar.gz.

File metadata

  • Download URL: hephaes_cli-0.1.0.tar.gz
  • Upload date:
  • Size: 36.3 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.14.4

File hashes

Hashes for hephaes_cli-0.1.0.tar.gz
Algorithm Hash digest
SHA256 0aac36fe26ec2deca0f251702a4be473092ec1cd218403ccc19517a7d971f618
MD5 f5b8cc0a630905d917811487b0744efc
BLAKE2b-256 08a3a67e1213a6fa42a2f662cc476ee4fc705be496b78b8814ef841d22e55f2b

See more details on using hashes here.

File details

Details for the file hephaes_cli-0.1.0-py3-none-any.whl.

File metadata

  • Download URL: hephaes_cli-0.1.0-py3-none-any.whl
  • Upload date:
  • Size: 29.3 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.14.4

File hashes

Hashes for hephaes_cli-0.1.0-py3-none-any.whl
Algorithm Hash digest
SHA256 1802868979752518ea5d6bc5b6bc8806302a73772dde4ac7301e4cda66729705
MD5 6af89598ad614e6a4acd83ccf2f97e5c
BLAKE2b-256 4cf2a20c1051378fc2913bedc43df9e06ba43a3ecefe512c4a64e15ab97527c6

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page