lima2mxh5master

lima2mxh5master creates, updates, and inspects NeXus/HDF5 master files for macromolecular crystallography workflows.

The project is currently in beta. Schema documents declare their specification version and whether they create or update a file; bundled transformation profiles also carry their own name and version.

Installation

Python 3.10 or newer is required.

python -m pip install lima2mxh5master

For a development checkout:

python -m pip install -e .

Create a file

from lima2mxh5master.api import dump_dict_to_nx

create_schema = {
    "spec_version": 1,
    "operation": "create",
    "registry": {"source_dir": [], "classes": {}},
    "schema": {
        "entry": {
            "@NX_class": "NXentry",
            "title": "Example scan",
            "data": [0, 1, 2, 3, 4, 5, 6, 7, 8, 9],
        }
    },
}

dump_dict_to_nx(create_schema, "master.h5")

Update a file

from lima2mxh5master.api import update_h5_from_dict

update_schema = {
    "spec_version": 1,
    "operation": "update",
    "registry": {"source_dir": [], "classes": {}},
    "schema": {
        "entry": {
            ">=sampled": "${h5:self:/entry/data}[::10]",
        }
    },
}

update_h5_from_dict(
    update_schema,
    path_h5_base_file="master.h5",
    path_h5_output_file="sampled_master.h5",
)

${h5:self:/entry/data} explicitly reads the original base file; ${h5:/entry/data} is the equivalent shorthand. Schemas can also declare named external inputs and reference them as ${h5:source_name:/entry/data} in create or update operations. Literal inputs use ${h5:<../raw/data.h5>:/entry/data}, and vds.concatenate([...], axis=0) combines sliced HDF5 references into a virtual dataset without reading frame data.

See the full documentation for schema syntax, in-place updates, deletions, HDF5 references, and bundled profiles.