Skip to content

themes.md

Latest commit

 

History

History
43 lines (37 loc) · 2.53 KB

themes.md

File metadata and controls

43 lines (37 loc) · 2.53 KB
Raw

Themes

The controlled vocabulary for a prompt's optional Themes: header — what the work is about, as opposed to Target:, which is only where the code lives. Themes describe related work across repositories and help with explicit task selection.

Usage, in a prompt's light header (see REFERENCE.md, "Prompt file format"):

Themes:
- mge
- jax-gradient

The first bullet is the primary theme. One to three keywords is the intended shape; unknown keywords are reported in dashboard Hygiene.

Vocabulary

  • mge: Multi-Gaussian Expansion light and mass profiles, and fitting with them.
  • point-source: Point-source lens modelling — image-plane solvers, positions, flux ratios.
  • pixelization: Pixelized source reconstruction — meshes, regularization, inversions.
  • interferometer: Visibility-space datasets — NUFFT transforms, interferometer fits.
  • cluster: Cluster- and group-scale lensing with many deflectors.
  • multi-band: Multi-wavelength or multi-dataset fitting across bands and instruments.
  • jax-compile: JAX tracing and jit compilation — compile time, retracing, the jit pipeline.
  • jax-gradient: JAX autodiff — gradient correctness, custom rules, gradient-based search.
  • samplers: Non-linear search machinery — nested sampling, MCMC, optimizers, search tiers.
  • numba-cpu: The numba CPU likelihood path — kernels, caching, CPU performance.
  • graphical-ep: Graphical models and expectation propagation — hierarchical fits.
  • profiling: Performance measurement campaigns — likelihood profiling, benchmarks, timings.
  • hpc-gpu: HPC and GPU execution — cluster submission, device placement, remote runs.
  • visualization: Figures and plotters — what the organism draws, and how it reads.
  • ci-smoke: CI workflows and workspace smoke tests — the automated green/red surface.
  • release: Release execution — versions, tags, PyPI publication, readiness gates.
  • dashboard: The Mind dashboard and its sibling one-tap boards.
  • mind-workflow: The task lifecycle itself — prompts, registries, skills, the dev workflow.
  • assistants: The domain science-assistant workspaces and their wikis.
  • docs-hub: Public documentation surfaces — RTD, the docs hub, the chat-surface routes.
  • notebooks: Notebook generation and the workspace script-to-notebook pipeline.
  • cti: Charge Transfer Inefficiency calibration and modelling.
  • reduce: Data reduction of raw imaging into modelling-ready datasets.
  • hygiene: Code-quality upkeep — dead code, slow tests, stale docs, dependency drift.