dump-research-info

---

This repo provides a setup to gather research information related to Center for Open Neuroscience (CON) as (meta)data and a tool to dump the gathered (meta)data to an instance of dump-things-server. The (meta)data gathered must conform to the data models defined in the demo-research-information-schema, which is documented at DataLad Concepts, and the dump-things-server instance receives the (meta)data in a collection that must conform to the data models defined in the demo-research-information-schema as well.

The setup is done in a way to facilitate AI coding agents, such as Claude Code, to gather the (meta)data from various sources, such as CON related websites and project repositories, and to validate the gathered (meta)data against the target data models through the REST API of a dump-things-server instance. Once any gathered (meta)data is validated, it will be added to this repository. The tool provided in this repository can be used for dumping all the gathered (meta)data in this repo to a dump-things-server instance.

The gathered (meta)data are stored in separate JSON files each named with the name of a model (or class) in the demo-research-information-schema, and the files are stored in a subdirectory named after the source of the (meta)data. All (meta)data records of a particular class from a particular source are stored in the file named after the class in the subdirectory named after the source. All subdirectories are stored in the data directory in this repository.

Table of Contents

• Server Configuration
• Commands and Tools
  • CLI
  • Hatch Commands
  • Claude Code Slash Commands
• License

Server Configuration

Tokens for accessing the dump-things-server instance for validating gathered metadata are defined in .dumpthings.yaml in the store/ directory. The interpretation of the token modes is documented at christian-monch/dump-things-server#67 (comment) and dump-things-server#118.

Commands and Tools

CLI

• dump-research-info — Main CLI entry point (also invocable via python -m dump_research_info). Used for dumping the gathered (meta)data in this repo to a dump-things-server instance.

Hatch Commands

• hatch run dump-server:start — Start a local dump-things-server instance configured to serve the store/ directory. The server listens on localhost:8111 and allows CORS requests from http://localhost:8000. Used for validating gathered metadata against the demo-research-information-schema.
• hatch run tools:serve-frontend — Serve the frontend web UI from _ext/pool.psychoinformatics.de-ui/dist/ on http://localhost:8000. Used for browsing and inspecting data in the dump-things-server instance.
• hatch run types:check — Run mypy type checking on the source and test packages.

Claude Code Slash Commands

• /gather_metadata <source> — Gather metadata from a specified source. This command guides Claude Code through fetching information from the source, constructing records conforming to the demo-research-information-schema, validating each record against a running dump-things-server instance, and storing valid records as JSON files in data/<source_name>/.
• /start_frontend — Launch the local frontend web UI for the dump-things-server. This command handles cloning and building the frontend repo (if needed) and serving it on http://localhost:8000.

License

dump-research-info is distributed under the terms of the MIT license.