StringForge

Shared database, model-loading, and vacua-vault infrastructure for string-compactification workflows.

StringForge provides the cross-package layer used to query Calabi-Yau geometry databases, construct JAXVacua models from database rows, persist vacuum solutions, and coordinate shared HuggingFace datasets. The physics engines live in sibling packages; this site documents the infrastructure interfaces that connect them.

StringForge is intentionally solver-light. It owns catalogue access, cache management, model handoff, vault layout, validation, and provenance. It does not replace JAXVacua, and it is not a public release of planned packages such as KahlerJAX or JAXiverse.

How to navigate

New to StringForge

Start with installation, the package boundary, and the shortest database-to-JAXVacua model-loading example.

Getting started

Running examples

Use the tutorial catalogue for quickstart, database access, vault, cluster, and ecosystem workflow notebooks.

Tutorials

Database and vault concepts

Read about HuggingFace layouts, catalogue versus mirror conventions, lazy caching, offline mode, and the shared vacua vault.

Calabi-Yau Geometry Database

Cluster production runs

Use Vulcan to stage vacuum batches on workers and batch-commit them to a shared HuggingFace dataset repo within the HF commit-rate cap.

Vulcan: production vacuum-forging for cluster workloads

API lookup

Go to the curated API reference when you already know which database, vault, or model-loading class you need.

API reference

Recommended first path

For a first pass through the documentation, read:

1. Getting started for installation and package orientation.

2. Calabi-Yau Geometry Database for data layout, cache behaviour, and convention boundaries.

3. Tutorials for executable notebooks.

4. API reference once you need precise signatures.

Users working specifically with the curated KKLT index should read KKLT Database after the general TDF/CICY database material.

Reference lookup

• Index

• Module Index