A standalone workshop page for mapping classes from BER LinkML schemas onto high-level grouping classes.
Open index.html in any browser (file:// works). No server, no
build step, no JavaScript dependencies.
Source classes come from the BRIDGE research dumps (NMDC, LAMBDA-BER, KBase CDM, BASIN-3D, BERtron, MIAPPE, …). You map each source class onto a grouping class. The built-in grouping classes follow BASALT’s process flow, in the same order as the schema explorer:
- Sampling activity
- Samples
- Sample processing
- Processed sample
- Data generation
- Data processing
- Data products
plus site-specific metadata, sample metadata, and other / doesn’t fit.
- Pick a source schema in the left rail.
- Drag a class onto a grouping class, or click the class and then click the grouping class.
- Open BASALT classes inside a grouping class and drop onto a specific class for a tighter mapping.
On your own, mappings and grouping classes stay in this browser (localStorage).
Export JSON and Export CSV each contain both the mappings and the grouping
classes, from the mapper or from Abstractions. Import reads either file and
restores both.
Deep-link: index.html#s=nmdc opens with that source schema selected.
index.html#page=groups opens the grouping-class editor.
Keyboard: 1–9 and 0 place the selected class on the original ten grouping classes; / focuses search; Esc clears the selection.
Open Abstractions to rename a grouping class or create a new one. An added grouping class is its own column, marked Added, in the color and section you choose. It has no BASALT subclasses of its own.
Built-in ids stay fixed, so a renamed grouping class keeps the mappings already saved. The same Export and Import actions are on this page and on the mapper.
A shared session lets a group edit the same grouping classes and mappings. Each change is saved and pushed to everyone who has the session open.
python3 serve.pyOpen the address it prints (others on the same network can use the network address). Share stays disabled until that server is running. Then enter your name and Start a shared session. Copy the link, or give them the code to join.
Leave the session when you want a private copy again. The shared copy stays on
the machine running serve.py, under .data/rooms/, and is still there after
a restart. Opening index.html as a file keeps working for one person; sharing
needs the server.
Class lists are embedded in index.html so the page works offline. Refresh them
from the BRIDGE dumps with:
python3 build.pyDumps are cached under .cache/bridge-dumps/ (gitignored). Delete that folder
to force a re-download.
build.py uses the Python standard library only.