LMFit Workflow¶
In this repository, Prefit is the lmfit-facing side of the SAXS workflow. It gives you a fast, editable model preview before you commit to a pyDREAM run.
That preview only becomes meaningful after the project has a template, experimental data, and a cluster-derived component set from the supporting applications.
Role of lmfit in SAXSShell¶
The template system defines an lmfit-compatible profile function. Prefit then uses that function together with the current parameter table to compute the preview model.
What the parameter table does¶
The parameter table in Prefit is the working parameter surface for the current template. Depending on the template, it can include:
- component weight parameters such as
w0,w1, ... - global scale, offset, or solvent parameters
- dynamic geometry parameters derived from cluster geometry metadata
Geometry-aware templates can now regenerate these parameter rows when the allowed shape per cluster changes.
Best Prefit¶
The current Prefit workflow supports a Best Prefit preset. This is useful when you want a stable parameter baseline for a specific template without overwriting every saved Prefit snapshot.
Save and restore¶
Use the save and restore actions to preserve:
- the current parameter table
- the cluster geometry table and active mode
- the current template runtime inputs
When lmfit is enough¶
Prefit is often enough when you want:
- a sanity check on component construction
- rough parameter exploration
- a baseline fit before setting up priors for DREAM
If you need posterior uncertainty, parameter correlations, or a more formal sampling workflow, move to the pyDREAM tab.