pycsamt.agents.web#

Gradio web interface for the pycsamt agent system.

From Python:

from pycsamt.agents.web import launch
launch()                      # opens http://localhost:7860

From the CLI:

python -m pycsamt.agents web

Chat (Orchestrator): Type a free-text request. The WorkflowOrchestratorAgent classifies the workflow, builds the chain, and runs it.
Load & QC: Browse to an EDI directory. MTLoaderAgent + DataQCAgent display the per-station quality table and confidence section figure.
Static Shift: Choose correction method. StaticShiftAgent shows before/after comparison.
Phase Analysis: PhaseAnalysisAgent runs the full tensor suite and displays figures inline.
Forward Modelling: Enter a layered model (resistivities and thicknesses). ForwardModelAgent plots the synthetic response.
AI Inversion (1-D): Choose architecture and epochs. AIInversionAgent trains and predicts, displaying the model section.
2-D AI Inversion: Inv2DAgent (U-Net) inverts the full profile at once, capturing lateral continuity.
Ensemble Inversion: EnsembleAgent — multiple independent models produce mean + σ uncertainty bands with conformal calibration.
Joint Inversion: JointInversionAgent (DRCNN) fuses MT with a secondary modality (TEM, CSAMT, or synthesised low-frequency sub-band).
Model Zoo: ModelZooAgent — list and load pre-trained checkpoints from the earthai-tech model zoo; run zero-shot prediction on any EDI dataset.
Report: Consolidate all results into a downloadable markdown / HTML report.

Requires gradio >= 4.0. Install with:

pip install gradio

Functions

`AIInversionAgent_lazy`(**kwargs)
`build_app`()	Build and return the Gradio Blocks app.
`launch`(*[, share, server_port, server_name])	Launch the Gradio web app.

pycsamt.agents.web.build_app()[source]#: Build and return the Gradio Blocks app.

pycsamt.agents.web.launch(*, share=False, server_port=7860, server_name='0.0.0.0', **kwargs)[source]#

Launch the Gradio web app.

Parameters:

Return type:

None