ptychography-parameter-preparation-and-submission

id	ptychography
name	Ptychography Parameter Preparation And Submission
description	Prepare ptychography reconstruction parameters, append experimental parameters, save them, and submit reconstruction to the server.
version	1.1.0

Agents Involved

• agent_ptychography

Required Tools

• agent_ptychography.prepare_parameters with arguments {"total_thickness": <number|null>, "num_slices": <integer|null>, "convergence_angle": <number|null>, "accela_voltage": <number|null>, "fov_nm": <number|null>, "PLRotation_deg": <number|string|null>, "rez_pixel_size_mrad": <number|null>, "rez_pixel_size_A": <number|null>, "aberrations": <array>, "data_pad": <integer|null>, "bright_threshold": <number>, "scan_step_A": <number|null>, "transform_axis_matrix": <array>}.
• agent_ptychography.append_exp_params with arguments {}.
• agent_ptychography.save_params with arguments {}.
• agent_ptychography.start_server_ptychography with arguments {"param_path": "<path returned by save_params or current ptychography UI path_params>"}.
• Optional support tool: agent_ptychography.load_preset with arguments {} when the user wants to start from a preset.
• Optional support tool: agent_ptychography.search_RAG_for_parameters with arguments {"parameter_name_list": ["<parameter names>"]} when parameter meanings/defaults are unclear.
• Optional support tool: agent_ptychography.change_pypty_parameters with arguments {"user_query": "<requested parameter change>"} when the user asks to modify current pypty or experimental parameters.

Steps

1. Prepare essential and experimental parameters: Call agent_ptychography.prepare_parameters. Use values from user input, current ptychography UI state, SEM/acquisition metadata, or JSON null for unknown optional numeric values. Use current UI/default values for non-null array/number fields such as aberrations, bright_threshold, and transform_axis_matrix.
2. Append experimental parameters after review: Call agent_ptychography.append_exp_params, but put confirm_prepared_ptychography_parameters in this step's required_inputs so execution stops and asks the user to review the prepared parameters before calling the tool.
3. Save parameters: Call agent_ptychography.save_params. The output should include or imply the parameter file path.
4. Submit reconstruction after saved-parameter review: Call agent_ptychography.start_server_ptychography with param_path from save_params output or the current ptychography UI experimental_params.path_params, but put confirm_saved_ptychography_parameters_before_server in this step's required_inputs so execution stops and asks the user before server submission.

Planning Contract

• The planner must use the exact tool names above. Do not use old names such as check_parameters, append_parameter, save_parameters, or send_to_server.
• Do not collapse prepare_parameters and append_exp_params; they are separate MCP tools and separate workflow steps.
• Do not add separate no-op confirmation steps. Use required_inputs on append_exp_params for confirming prepared parameters, and on start_server_ptychography for confirming saved parameters.
• Optional numeric arguments must be JSON numbers when known and JSON null or omitted when unknown. Non-null fields (aberrations, bright_threshold, transform_axis_matrix) must be populated from UI/defaults. Do not use the string "None".
• If the user asks to explain or modify a ptychography parameter, use search_RAG_for_parameters or change_pypty_parameters as a support step only when needed.

Success Criteria

• Essential and experimental ptychography parameters are prepared for the selected 4D-STEM dataset.
• Experimental parameters are appended into the pypty parameter dictionary.
• Parameters are saved to disk.
• Reconstruction job is submitted to the server.

Failure Policy

• If required reconstruction parameters are missing, put them in the relevant step's required_inputs list and use on_failure: ask_user; do not block the whole plan with PlanReview missing_inputs.
• If parameter preparation, appending, saving, or server submission fails, stop the workflow and report the exact failing tool/error.