Skip to content

QUICK_REFERENCE.md

Latest commit

 

History

History
122 lines (102 loc) · 3.41 KB

QUICK_REFERENCE.md

File metadata and controls

122 lines (102 loc) · 3.41 KB
Raw

Quick Reference - APEXA Commands

Calibration (CalibrationAgent)

APEXA> calibrate the CeO2 image in test1
APEXA> calibrate using LaB6_650mm.tif with 40 iterations
APEXA> convert 61.332 keV to wavelength
APEXA> what's the d-spacing for CeO2 (111)?
APEXA> list common calibrants

Integration (AnalysisAgent)

APEXA> integrate CeO2_000001.tif using refined params
APEXA> integrate with dark file subtraction
APEXA> batch integrate all .tif files in /data with 8 CPUs

FF-HEDM (AnalysisAgent)

APEXA> run FF-HEDM workflow on /data/experiment
APEXA> run FF-HEDM with 32 CPUs

NF-HEDM (AnalysisAgent)

APEXA> run NF-HEDM reconstruction with FF seed orientations
APEXA> reconstruct microstructure with 10 CPUs

Visualization (VisualizationAgent)

APEXA> show me the lineout for CeO2 in test1
APEXA> plot the caked output from test1/integration
APEXA> show calibration results for test1
APEXA> view the 2D diffraction image
APEXA> compare lineouts with ideal ring positions

Files

APEXA> list files in /data
APEXA> read the Parameters.txt file
APEXA> what's in the integration folder?

Knowledge (KnowledgeAgent)

APEXA> what is FF-HEDM?
APEXA> explain Bragg's law
APEXA> what are typical parameters for steel?
APEXA> get material properties for Ti-6Al-4V

X-ray Calculations

APEXA> convert 61.332 keV to wavelength
APEXA> calculate d-spacing for 2-theta 10.5 degrees at 0.2066 angstroms
APEXA> calculate strain for measured d=3.155 and reference d=3.124

CLI Commands (not sent to AI)

Command Description
models Show available AI models
model <name> Switch model (gpt4o, gpt54, claudesonnet45, gemini25pro)
timing Toggle API response time display on/off
tools List all analysis tools
servers Show connected servers
ls <path> List directory
clear Clear conversation history
help Show help
quit Exit

System Overview

User (natural language)
         |
    Argo Gateway (GPT-4o / Claude / Gemini)
         |
    OrchestratorAgent
         |
    +----------------+----------+-----------+----------------+
    | Calibration    | Analysis | Knowledge | Visualization  |
    +----------------+----------+-----------+----------------+
                         |
              +----------+----------+
              |   core (9 tools)    |
              |   midas (21 tools)  |
              +---------------------+

Servers: core (beamline_core_server.py) + midas (midas_comprehensive_server.py)

See USER_MANUAL.md for complete documentation.

Autonomous refinement (gsas2 server)

For any powder pattern that is not MIDAS caked output. For .zarr.zip, use run_gsas_refinement instead.

Ask for Tool
refine one pattern refine_pattern
find a structure from a pattern + a name propose_structures
refine an in-situ stack refine_series_submit → refinement_status
is an old refinement still usable assess_refinement 
APEXA> refine scan_042.xye with LiCoO2.cif
APEXA> find candidate structures for this pattern, it should be goethite
APEXA> refine the whole temperature series and report phase fractions

Every result carries a trust verdict naming any reason to inspect it before acting. Low M20 from propose_structures means widen the search rather than refine harder. See .agents/skills/gsas2-agentic/SKILL.md.