| name | create-research-study-document |
| description | Describes how to create documents (document resources) that track an in-progress research study: motivation, question, objectives, materials, study labels, and a todo list of next steps. Use when starting a study, creating a study document, updating study todos/objectives, or when the user mentions a study document by name (e.g. study-$ID) or asks about study tracking.
|
Creating a Research Study Document
A research study document describes a research study to perform, or that is
already underway. It is stored as a document resource in the metastore.
For generic document create/query, see
resource-yaml-creation — Document.
For CLI syntax, see using-ado-cli.
Naming and metadata
metadata.name must be study-$ID where $ID is the short study slug
(e.g. study-cplex-mip).
Required metadata fields:
name — study-$ID
description — one-line summary of the study
todo — YAML list of current next-step strings
Leave relatedResources empty. Associate spaces and operations with the study
via labels only (avoids colliding with op/space report queries on
relatedResources).
Content sections
Use this reading order. Exact heading names matter for consistency.
Required
## Study question — what the study aims to learn (investigable questions)
## Study objective — current campaigns / methods to answer those questions
## Study labels — labels that must be applied to spaces and operations
in this study (include at least study: $ID)
Optional
## Motivation — why the study exists (stakeholder framing; keep stable).
Place before Study question when present.
## Materials — instances, formulations, input files, and related paths.
Place after Study objective and before Study labels when present.
Prefer a short summary here and link out for long formulations.
Do not fold motivation into Study question, or formulation/instance detail into
Study objective.
Template
metadata:
name: study-cplex-mip
description: Compare TPE vs LHS for CPLEX MIP gap on bab6
todo:
- Run TPE baseline (60m, 200 trials)
- Analyse mip-gap histograms
content: |
## Motivation
Establish a well-tuned classical baseline for this problem class.
Which sampler finds better MIP gaps faster on bab6?
Run comparable TPE and LHS campaigns; compare mip-gap
Omit Motivation and/or Materials when they add nothing.
uv run ado create document -f study-cplex-mip_document.yaml --dry-run
uv run ado create document -f study-cplex-mip_document.yaml
Apply the study labels to every space and operation created for the study
(see resource-yaml-creation metadata).
Query
uv run ado get document -q 'config.metadata.name=study-$ID' --details
uv run ado get document --details
uv run ado get spaces -l study=$ID --details
uv run ado get operations -l study=$ID --details
Fetch the study body with uv run ado describe document DOCUMENT_ID.
Update
- Use
ado edit document DOCUMENT_ID to update todo and description.
- Full body refresh: create a replacement document resource