Output Format Reference

Use this reference to determine which BEACH files are generated, which columns or receipts to read, and which files are required for restart. A receipt is a resolved setting or runtime state written to summary.txt as key=value, not merely a copy of the input value.

For a first inspection, follow Inspect Output Files before using this page. The machine-readable source of truth for file-generation conditions is schemas/beach.output-manifest.json.

File-generation conditions

Each row keeps a file’s generation condition, format, purpose, and reader or restart role together. For a complex file, the linked section repeats this information before expanding its columns and decision rules.

File	Generation condition	Format and purpose	Python reader / restart
summary.txt	output.write_files=true	key=value receipts. Completion and major keys	Fields of FortranRunResult. Required for restart
charges.csv	output.write_files=true	elem_idx, charge_C. Final signed charge per element	result.charges. Required for restart
mesh_triangles.csv	output.write_files=true	elem_idx, v0x, v0y, v0z, v1x, v1y, v1z, v2x, v2y, v2z, charge_C, mesh_id	result.triangles, result.mesh_ids. Not used for restart
mesh_sources.csv	OBJ or template mesh with output.write_files=true	mesh_id, source_kind, template_kind, surface_model, epsilon_r, elem_count	result.mesh_sources. Not used for restart
mesh_potential.csv	output.write_files=true and output.write_mesh_potential=true	elem_idx, potential_V. Potential composition	result.mesh_potential_v. Not used for restart
charge_history.csv	output.write_files=true and output.history_stride>0	batch, processed_particles, rel_change, elem_idx, charge_C	result.history. Not used for restart
potential_history.csv	output.write_files=true, output.write_potential_history=true, and output.history_stride>0	batch, elem_idx, potential_V. Join to the reference	No dedicated FortranRunResult attribute; read the CSV directly
top_reference_history.csv	output.write_files=true, output.write_potential_history=true, output.history_stride>0, and a [domain] box	batch, simulated_time_s, z_high_m, sample_n, potential_mean_V, potential_std_V, potential_min_V, potential_max_V	No dedicated FortranRunResult attribute; read the CSV directly
matching_plane_history.csv	output.write_files=true, output.history_stride>0, and surface_current_model.model=matching_plane_quasistatic	Seventeen-column accepted state	result.matching_plane_history. Not used for restart
charge_ledger.csv	output.write_files=true and charge-ledger state is present	Twenty-five columns per species	result.charge_ledger. Required for restart when summary has ledger metadata
rng_state.txt / rng_state_rankNNNNN.txt	output.write_files=true; the former for serial, the latter per MPI rank	Internal RNG state	No FortranRunResult attribute. Required for restart
macro_residuals.csv	output.write_files=true and macro-particle remainder state is allocated	species_idx, face, residual	No FortranRunResult attribute. Schema v8+: required when declared by the manifest
checkpoint_complete.txt	output.write_files=true for final output and each periodic slot	key=value completion manifest. Meaning and selection	No FortranRunResult attribute. Required for schema v8+ restart
checkpoint_latest.txt	output.write_files=true and after a matching accepted-batch commit with output.checkpoint_stride>0	Periodic-slot index	Candidate-selection hint, not evidence of completeness
performance_profile.csv	BEACH_PROFILE=1 and output.write_files=true	After comment lines: region, calls_sum, calls_mean, rank_min_s, rank_mean_s, rank_max_s, imbalance_ratio	beach-plot-performance-profile <output_dir>. Not used for restart

The Python reader requires summary.txt and charges.csv. Other files may be absent according to configuration or runtime state.

Locate configuration-specific values

summary.txt is a key=value file generated with output.write_files=true.

load_fortran_result(...) converts major values to FortranRunResult attributes; use beach.summary.load_summary_file(...) for untyped receipts.

Completion checks

Run completion and checkpoint completeness are separate conditions.

Check	Condition	This condition does not establish
Requested batches completed	batches == sim.batch_count using summary.txt and the input configuration	Numerical convergence or physical validity
Schema v8+ checkpoint is complete	checkpoint_complete.txt has state=complete, and the manifest agrees with every required file	Completion through sim.batch_count
Python reader succeeds	summary.txt and charges.csv satisfy the reader contract	Checkpoint completeness or physical validity

checkpoint_complete.txt marks completion of a write transaction. An intermediate periodic checkpoint can also have state=complete, so do not use it as evidence that the run reached its requested target. See Validating Simulation Results for numerical and physical acceptance. For a current-schema normal-completion output, require both of the first two conditions in the table.

Common summary.txt receipts

The tables below list major receipts used to interpret results and decide restart compatibility. They do not claim that every internal diagnostic key in summary.txt is a fixed API. The canonical producer is bem_output_writer.f90; io.py owns the keys converted to typed Python attributes.

Receipt	Meaning
mesh_nelem, mesh_count, mpi_world_size	Mesh and MPI configuration
batches, processed_particles, absorbed, escaped, escaped_boundary, survived_max_step	Executed work and major particle outcomes
simulated_time_s, last_rel_change	Accepted simulated time and final relative change. last_rel_change is not an early-stop condition
checkpoint_schema_version, checkpoint_stride	Checkpoint format generation and resolved output interval
mesh_fingerprint	Ordered mesh identity used to map saved charges to elements
reservoir_inflow_map, particle_ordinary_open_model	Resolved reservoir and ordinary-open boundary models

The build receipt consists of these five keys.

Receipt	Meaning
build_info_schema_version	Build-receipt schema
build_version	BEACH version
build_version_mode	How the version was resolved
build_source_commit	Source commit
build_id	Build identifier

Diagnose multiple box-boundary events with these five keys.

Receipt	Meaning
multiple_box_events_retry_attempted	Number of particles for which retry was attempted
multiple_box_events_retry_resolved	Number of particles resolved by retry
multiple_box_events_soft_discarded	Number of soft-discarded particles
multiple_box_events_soft_discard_fraction	Soft-discard fraction of processed_particles
multiple_box_events_soft_discarded_abs_charge_C	Total absolute soft-discarded charge [C]

Field and boundary receipts

Field-reconstruction schema v2 provides the receipts needed to reconstruct the runtime electric field. For new output, use result.field_reconstruction instead of inferring settings from a nearby beach.toml.

Receipt	Meaning
field_reconstruction_schema_version	Field-reconstruction receipt schema; currently 2
field_reconstruction_resolved_field_solver	Actual direct, treecode, or fmm selection
field_reconstruction_fmm_expansion_order	FMM expansion order. Schema v2 requires this and the preceding key
field_reconstruction_field_bc_mode	Resolved free or periodic2 mode
field_reconstruction_e0_V_m	Resolved uniform external electric field E0
field_reconstruction_use_box, field_reconstruction_box_min_m, field_reconstruction_box_max_m	Box presence and bounds
field_reconstruction_boundary_low, field_reconstruction_boundary_high	Resolved states of the six faces
field_reconstruction_tree_theta, field_reconstruction_tree_leaf_max	Tree settings actually used

The periodic-field keys are listed below. See periodic2 Field Model for their physical meaning and solver selection.

Receipt	Meaning
field_reconstruction_periodic_image_layers	Number of near-image shells
field_reconstruction_periodic_far_correction	Far-correction model
field_reconstruction_periodic_nonzero_mode_backend	Nonzero-mode backend
field_reconstruction_periodic_zero_mode_policy	Zero-mode policy
field_reconstruction_periodic_lower_boundary_model	Lower-boundary model
field_reconstruction_periodic_reference_mode_layers	Image layers for the reference nonzero mode
field_reconstruction_periodic_panel_quadrature_order	Panel quadrature order
field_reconstruction_periodic_ewald_alpha	Ewald splitting parameter
field_reconstruction_periodic_ewald_layers	Number of Ewald layers
field_reconstruction_periodic_cache_dir	Operator-cache directory
field_reconstruction_periodic_generation_tolerance	Operator-generation tolerance

zhao_stationary

See Zhao stationary closure for the [surface_current_model] physical model, Type A/B/C, and validity limits. This section defines how to read its receipts.

Receipt	Interpretation
surface_current_model, surface_current_model_zhao_branch	Selected model and branch
surface_current_model_kinetic_contract	Velocity-space boundary-map contract paired with the Zhao currents
surface_current_model_photoelectron_active	Whether PE channels exist. PE receipts are zero when false
surface_current_model_reference_area_m2	Reference area for current targets
surface_current_model_phi0_V, surface_current_model_phi_m_V	Resolved surface and minimum potentials
surface_current_model_ambient_electron_density_m3	Resolved ambient electron density
surface_current_model_electron_current_density_A_m2, surface_current_model_ion_current_density_A_m2	Signed electron and ion current densities
surface_current_model_pe_emission_current_density_A_m2, surface_current_model_pe_escape_current_density_A_m2, surface_current_model_pe_return_current_density_A_m2	Signed PE current densities. Emission and escape are positive; return is negative
surface_current_model_pe_escape_particle_current_A	Negative because it is the signed current carried by outward PE particles
surface_current_model_net_current_density_A_m2	Net surface current density over all channels
surface_current_model_pe_budget_residual_current_density_A_m2	PE emission-return-escape continuity residual
surface_current_model_surface_budget_residual_current_density_A_m2	Stationary surface-current residual
surface_current_model_electron_inflow_reservoir_potential_V, surface_current_model_electron_inflow_access_potential_V, surface_current_model_electron_inflow_face	Electron-inflow reservoir, access potential, and face
surface_current_model_pe_outflow_barrier_potential_V, surface_current_model_pe_outflow_barrier_face	PE outflow-barrier potential and face; zero when PE is inactive

Face indices are 1..6 = x_low, x_high, y_low, y_high, z_low, z_high.

matching_plane_quasistatic

For this model, treat the accepted-batch matching-plane receipts listed below as state values instead of using static surface-current targets. See the matching-plane numerical and response-table reference for the canonical fixed-point equations, response CSV, and implicit_zero_mode contract.

Item	matching_plane_history.csv contract
Generation	output.write_files=true, output.history_stride>0, and surface_current_model.model=matching_plane_quasistatic
One row	Leading batch, simulated_time_s, followed by the 15 state columns below; only accepted states are written
Interpretation	Read the fluxes, barriers, and iteration receipt on one row as one batch’s fixed-point solution
Python	result.matching_plane_history; result.matching_plane_state holds the last accepted state
Restart	The CSV is not used. Schema v9 stores the last accepted state in summary.txt

Accepted state

summary.txt receipt	matching_plane_history.csv column	Meaning
matching_plane_displacement_C_m2	D_H_C_m2	Displacement charge density $D_H$ at the matching plane
matching_plane_phi_V	phi_H_V	Matching-plane potential $\Phi_H$
matching_plane_electron_inward_flux_m2_s	electron_inward_flux_m2_s	Inward electron flux
matching_plane_ion_inward_flux_m2_s	ion_inward_flux_m2_s	Inward ion flux
matching_plane_electron_access_potential_V	electron_access_potential_V	Electron access potential
matching_plane_ion_access_potential_V	ion_access_potential_V	Ion access potential
matching_plane_photoelectron_barrier_potential_V	photoelectron_barrier_potential_V	PE barrier potential
matching_plane_photoelectron_outward_flux_m2_s	photoelectron_outward_flux_m2_s	Outward PE flux fed back to the fixed point
matching_plane_photoelectron_mean_normal_energy_eV	photoelectron_mean_normal_energy_eV	PE mean normal energy fed back to the fixed point
matching_plane_electron_outward_flux_m2_s	electron_outward_flux_m2_s	Outward electron flux fed back to the fixed point
matching_plane_ion_outward_flux_m2_s	ion_outward_flux_m2_s	Outward ion flux fed back to the fixed point
matching_plane_photoelectron_return_flux_m2_s	photoelectron_return_flux_m2_s	PE return flux produced by the backend
matching_plane_photoelectron_escape_flux_m2_s	photoelectron_escape_flux_m2_s	PE escape flux produced by the backend
matching_plane_iterations	iterations	Fixed-point iteration count
matching_plane_residual	residual	Effective relative residual at acceptance

The first two history columns are batch and simulated_time_s, giving 17 columns in total. When matching_plane_state_valid is false, do not treat the state receipts listed above as an accepted state.

Check the PE balance within the same batch:

$$\mathtt{photoelectron\_outward\_flux\_m2\_s} =\mathtt{photoelectron\_return\_flux\_m2\_s} +\mathtt{photoelectron\_escape\_flux\_m2\_s}.$$

The $D_H$ and $\Phi_H$ in each record correspond to the pre-commit surface-charge state used to track that batch. In contrast, simulated_time_s is the time after the trial was accepted and advanced. Do not interpret the record as the post-commit field at the start of the next batch.

Provenance and convergence

Receipt	Meaning
surface_current_model_response_backend	table or zhao_online
surface_current_model_zhao_root_selection	Online-Zhao root policy: require_unique, minimum_energy, or continuation
surface_current_model_implicit_zero_mode	T advances the plane-mean $D_H$ with backward Euler; F holds its batch-start value fixed
surface_current_model_matching_plane_z_m	Matching-plane z coordinate
surface_current_model_electron_species, surface_current_model_ion_species, surface_current_model_photoelectron_species	Species assigned to each channel
surface_current_model_coupling_rtol	Relative tolerance for active components
surface_current_model_coupling_atol	Four componentwise absolute tolerances
surface_current_model_coupling_max_iterations	Maximum iteration count
surface_current_model_coupling_relaxation	Fixed-point relaxation factor
surface_current_model_dynamic_state_source	surface_current_model_dynamic_state_source=accepted_batch_fixed_point for committed matching trials. This compatibility string alone does not imply convergence

The four values of surface_current_model_coupling_atol are ordered as follows:

1. Outward PE flux [m^-2 s^-1]
2. PE mean normal energy [eV]
3. Outward electron flux [m^-2 s^-1]
4. Outward ion flux [m^-2 s^-1]

All default to zero, and inactive components must also remain zero. Each active component uses max(coupling_rtol * backend_scale, coupling_atol) as its threshold. An absolute-tolerance-dominated component is converted to an effective residual, so a converged state’s matching_plane_residual remains no greater than surface_current_model_coupling_rtol. A state committed with an iteration-limit warning exceeds that value and has matching_plane_iterations equal to the configured limit.

Backend-specific receipts

Backend	Receipt
table	surface_current_model_response_table_path. All ranks use the table loaded by root; no content fingerprint is written
zhao_online	surface_current_model_response_contract=matching_plane_zhao_online_v1
zhao_online	surface_current_model_zhao_branch
zhao_online	surface_current_model_zhao_root_selection
zhao_online	surface_current_model_outer_solver=charge_driven_finite_h_sagdeev
zhao_online	surface_current_model_photoelectron_closure=moment_matched_half_maxwellian
zhao_online	surface_current_model_ambient_outward_feedback=transparent
zhao_online with require_unique / minimum_energy	surface_current_model_outer_solver_state=stateless
zhao_online with continuation	surface_current_model_outer_solver_state=accepted_endpoint_continuation_v2

accepted_endpoint_continuation_v2 is a provenance receipt for the configured ownership rule: only an accepted endpoint seeds the next batch, and a unique nearest root from full multistart is used when local Newton cannot reacquire a nearby root. It does not report the number of full-multistart or step-subdivision evaluations or the root displacement. Use matching_plane_state_valid to determine whether an accepted state exists.

History

File	Generation	Columns and interpretation	Python
charge_history.csv	output.write_files=true and output.history_stride>0	batch, processed_particles, rel_change, elem_idx, charge_C; one snapshot contains all element charges	result.history
potential_history.csv	output.write_files=true, output.write_potential_history=true, and output.history_stride>0	batch, elem_idx, potential_V; one snapshot contains every element-centroid potential	No dedicated attribute; read the CSV directly
top_reference_history.csv	output.write_files=true, output.write_potential_history=true, output.history_stride>0, and a [domain] box	batch, simulated_time_s, z_high_m, sample_n, potential_mean_V, potential_std_V, potential_min_V, potential_max_V	No dedicated attribute; read the CSV directly
matching_plane_history.csv	Matching-plane with stride enabled	Seventeen-column accepted state	result.matching_plane_history

The reference in top_reference_history.csv is the mean over the box z-high face, not infinity or plasma potential. Join it to the same batch in potential_history.csv and compute potential_V - potential_mean_V for element-relative potential.

For matching-plane output, typed result.matching_plane_state and result.matching_plane_history receipts avoid manual column indexing.

Mesh potential

Item	mesh_potential.csv contract
Generation	output.write_files=true and output.write_mesh_potential=true
Columns	elem_idx, potential_V
Interpretation	Potential [V] at each element centroid
Python	result.mesh_potential_v
Restart	Not used

Field model	Terms included in the potential
Common	Analytic P0 triangle-panel self term
Finite periodic2	Configured image shell
cached_kneq0	Cached nonzero mode plus the configured physical zero mode

Charge ledger

Item	charge_ledger.csv contract
Generation	output.write_files=true and charge-ledger state is present
One row	Signed charges, corrections, weights, and counts for one species; all 25 columns appear below
Python	ChargeLedgerEntry objects in result.charge_ledger
Restart	Required when summary.txt contains ledger metadata

The column order is:

Column	Meaning
batch	Batch represented by the ledger state
species_idx	One-based species index
injected_from_remote_C	Signed charge injected from outside [C]
emitted_from_surface_C	Signed charge emitted from surfaces [C]
absorbed_on_surface_C	Signed charge absorbed by surfaces [C]
escaped_to_infinity_C	Raw trajectory escape charge [C]
discarded_unresolved_C	Signed charge removed as unresolved [C]
neutral_return_correction_C	Closed-PE neutral-return correction [C]
neutral_return_weight_scale	Weight scale used for neutral return
neutral_return_unresolved_fraction	Unresolved neutral-return fraction
fixed_absorbed_target_charge_C	Fixed-current absorption target [C]
fixed_absorbed_weight_scale	Weight scale applied to absorption samples
fixed_emission_target_charge_C	Fixed-current emission target [C]
fixed_emission_weight_scale	Weight scale applied to emission samples
fixed_current_correction_C	Charge added to the surface by the external closure [C]
fixed_absorbed_applied_charge_C	Compatibility alias equal to the absorption target
fixed_emission_applied_charge_C	Compatibility alias equal to the emission target
fixed_escape_target_charge_C	External escape target [C]
fixed_escape_applied_charge_C	Compatibility alias equal to the escape target
fixed_escape_correction_C	Difference between target and raw escape [C]
injected_count	Injected macro-particle count
emitted_count	Surface-emitted macro-particle count
absorbed_count	Surface-absorbed macro-particle count
escaped_count	Escaped macro-particle count
discarded_unresolved_count	Unresolved-discard macro-particle count

Companion receipts in summary.txt

Receipt	Meaning
charge_ledger_nspecies, charge_ledger_batch_count	Species count and batch represented by the ledger
charge_ledger_surface_charge_before_C, charge_ledger_surface_charge_after_C	Surface-charge stock before and after the update
charge_ledger_local_flight_charge_before_C, charge_ledger_local_flight_charge_after_C	Local-flight stock before and after the update
charge_ledger_unresolved_stock_before_C, charge_ledger_unresolved_stock_after_C	Unresolved stock before and after the update
charge_ledger_residual_C, charge_ledger_discarded_unresolved_abs_C	Conservation residual and absolute unresolved charge
charge_ledger_neutral_return_correction_C, charge_ledger_fixed_current_correction_C	Total closure corrections
charge_ledger_fixed_absorbed_applied_charge_C, charge_ledger_fixed_emission_applied_charge_C	Total surface fixed-current targets
charge_ledger_raw_escape_charge_C, charge_ledger_fixed_escape_target_charge_C, charge_ledger_fixed_escape_applied_charge_C, charge_ledger_fixed_escape_correction_C	Raw escape, target, compatibility alias, and correction
charge_ledger_fixed_applied_surface_net_charge_C	Net charge applied to the surface by fixed current
charge_ledger_fixed_pe_continuity_residual_C	Fixed-PE continuity residual when PE is active

Decision rules

Check	Read together	Caution
External escape	escaped_to_infinity_C, fixed_escape_target_charge_C, fixed_escape_correction_C	The raw trajectory value is preserved. The escape correction is not deposited onto surface elements
Conservation residual	charge_ledger_residual_C and the surface / local-flight / unresolved before / after pairs above	Includes stocks, external fluxes, neutral return, and fixed-current corrections
fixed_current sampling variance	absorbed_count, emitted_count, raw charge, fixed_absorbed_weight_scale, fixed_emission_weight_scale	A count of one localizes the whole target onto the element represented by one sample
Closed PE	Raw absorption and unresolved values, neutral_return_correction_C, neutral_return_unresolved_fraction	Check the unresolved fraction separately even when total surface charge closes

The output aliases for existing readers map to the equal in-memory targets as follows.

Output alias	In-memory target
fixed_absorbed_applied_charge_C	fixed_absorbed_target_charge_C
fixed_emission_applied_charge_C	fixed_emission_target_charge_C
fixed_escape_applied_charge_C	fixed_escape_target_charge_C

A small conservation residual does not establish elementwise sampling convergence. Test the charge distribution separately across particle or ray counts, batch widths, and RNG seeds. See Check the charge ledger for the basic procedure.

Adaptive-batch diagnostics

With periodic2.max_nonzero_mode_potential_step > 0, summary.txt records:

Receipt	Meaning
simulated_time_s	Simulated time through the accepted trial
periodic2_max_nonzero_mode_potential_step_V	Configured nonzero-mode potential-step limit
adaptive_nonzero_mode_rejected_trials	Cumulative rejected-trial count
adaptive_nonzero_mode_last_batch_duration_s	Duration of the last accepted batch
adaptive_nonzero_mode_last_potential_step_V	Nonzero-mode potential step of the last accepted batch
adaptive_nonzero_mode_omp_threads	OpenMP thread count fixed for adaptive restart

Rejected trials do not appear in history files or the charge ledger.

Files used for resume

With output.checkpoint_stride > 0, BEACH updates this structure after an accepted-batch commit. Even with output.checkpoint_stride=0, a normal completion writes the final checkpoint.

Here, “complete” means that the transaction containing the restart files is complete. Separately establish completion of the requested run with batches == sim.batch_count using summary.txt and the input configuration.

<output.dir>/
├── checkpoint_complete.txt
├── checkpoint_latest.txt
└── checkpoints/
    ├── slot0/
    └── slot1/

Contents of each checkpoint

File	Required when	Format and role
summary.txt	Always	key=value statistics, schema, fingerprints, and ledger stocks
charges.csv	Always	Element charges as elem_idx, charge_C
rng_state.txt	Serial	RNG array length and state
rng_state_rankNNNNN.txt	MPI; one for every rank	Per-rank RNG array length and state
macro_residuals.csv	Declared by a schema v8+ manifest	species_idx, face, residual global macro-particle remainders
charge_ledger.csv	Summary / manifest contains ledger metadata	Twenty-five-column species ledger
checkpoint_complete.txt	Schema v8+	schema_version, state, batches, mpi_world_size, macro_residuals_present, charge_ledger_present

Update and automatic selection

Stage	Contract
Update	Finish the inactive slot before atomically switching checkpoint_latest.txt. At most two generations are retained
Search	Inspect the final output under output.restart_from and both slot0 / slot1 directories
Selection	Choose the loadable candidate with all required files and the largest batches value
Index failure	Recover a slot with a complete manifest even if checkpoint_latest.txt is missing, malformed, or stale
Mismatch	Stop when required files, the mesh fingerprint, saved-state array shapes, or MPI world size differ. Model and species configuration changes are not compared through hashes

output.restart_from changes only the read source; new output is written to output.dir.

Schema generations

Schema	Contract
v6	macro_residuals.csv uses species_idx,face,residual; face=0 is the legacy source and 1..6 are boundary faces. The older two-column species_idx,residual form remains readable
v8+	Publish state=in_progress before replacing state. Atomically publish state=complete only after closing summary, charges, every rank’s RNG, and manifest-declared residual / ledger files
v9	Store accepted matching-plane feedback, potential, return / escape fluxes, and iteration receipts in summary.txt
v10 (current)	Remove model, species, and response-content fingerprints. Continue reading v2–v9 and ignore their old identifiers. Preserve mesh identity and state-array formats

For schema v8 and later, checkpoint_complete.txt itself is required. If the final-output directory contains files from different generations, BEACH rejects it and falls back to a complete periodic slot.

Configuration changes are not compared through hashes. BEACH carries saved charges and statistics into the current configuration; compare the configuration files and explicit summary values when assessing changed conditions. A mesh fingerprint mismatch remains fatal because charges.csv rows cannot be mapped safely to different elements. The Python reader retains optional model_fingerprint and species_fingerprint attributes for historical outputs; both are None for current outputs.

Online implicit bracket and step-subdivision nodes are neither a persistent table nor checkpoint state. With continuation, BEACH tries to reconstruct the restart seed from the accepted response saved in summary.txt and falls back to the minimum-energy bootstrap if reconstruction fails. Every policy evaluates the needed response directly from the same Zhao contract, so no matching_query.csv or fingerprint of automatic search points is required.

See Execution and Resume for the actual procedure.

Reading from Python

The file-generation table gives each file’s attribute or explicitly states when no dedicated attribute exists. Read representative typed attributes as follows.

from beach import load_fortran_result

result = load_fortran_result("outputs/tutorial")
print(result.charges)
print(result.charge_ledger)
print(result.field_reconstruction)
print(result.matching_plane_state)

Output	Python attribute
charges.csv	result.charges
charge_ledger.csv	result.charge_ledger
Field-reconstruction receipt	result.field_reconstruction
Matching-plane summary	result.matching_plane_state
matching_plane_history.csv	result.matching_plane_history

summary.txt and charges.csv are required. Mesh, history, ledger, field-reconstruction, and matching-plane state are loaded only when the corresponding output is present.

When matching_plane_state_valid is false, the reader does not expose matching_plane_state or matching_plane_history as an accepted state.

For another case, replace outputs/tutorial with its actual output.dir. See the Python Post-processing API for class and attribute details, and Validating Simulation Results for physical and numerical acceptance.