fmdtools 2.5.4__tar.gz → 2.5.5__tar.gz

{fmdtools-2.5.4 → fmdtools-2.5.5}/LICENSE

@@ -1,4 +1,4 @@
-fmdtools version 2.5.4
+fmdtools version 2.5.5
 ---------------------------
                                  Apache License
                            Version 2.0, January 2004

{fmdtools-2.5.4/fmdtools.egg-info → fmdtools-2.5.5}/PKG-INFO

@@ -1,10 +1,10 @@
 Metadata-Version: 2.4
 Name: fmdtools
-Version: 2.5.4
+Version: 2.5.5
 Summary: System resilience modelling, simulation, and assessment in Python
 Author: fmdtools Developers
 Maintainer-email: Daniel Hulse <daniel.e.hulse@nasa.gov>
-License: fmdtools version 2.5.4
+License: fmdtools version 2.5.5
         ---------------------------
                                          Apache License
                                    Version 2.0, January 2004
@@ -394,7 +394,7 @@ More guidance on development installations is provided in [How to Contribute](ht
 Note that this version reflects the latest version on the ``main`` git branch along with any local changes. To use a particular version of ``fmdtools`` in the context of a development install you can checkout the tag corresponding to the version you want, e.g.:
 
 ```
-   git checkout v2.5.4
+   git checkout v2.5.5
 ```
 
 
@@ -500,9 +500,9 @@ Decent (but **very imperfect**) results can be achieved using LLMs for model dev
 
 ```prompt
 You are an experienced software engineer learning a new Python library for the first time.
-Make an fmdtools simulation of a <enter your system here> using the latest syntax from `https://github.com/nasa/fmdtools/tree/v2.5.1` for all modeling (`fmdtools.define`), simulation (`fmdtools.sim`), and analysis (`fmdtools.analyze`) classes.
+Make an fmdtools simulation of a <enter your system here> using the latest syntax from `https://github.com/nasa/fmdtools/tree/v2.5.5` for all modeling (`fmdtools.define`), simulation (`fmdtools.sim`), and analysis (`fmdtools.analyze`) classes.
 The output should be a bare bones but well-commented .py file with simple behaviors for each class and a small script at the bottom leveraging the built-in capabilities in the `fmdtools.analyze` package instead of directly using external libraries like matplotlib.
-As you are going, explain the structure of each class, the usage of each function call and how they each leverage the correct fmdtools v2.5.1 syntax.
+As you are going, explain the structure of each class, the usage of each function call and how they each leverage the correct fmdtools v2.5.5 syntax.
 ```
 
 The important aspects of this prompt are specifying the URL and version--otherwise the LLM is likely to hallucinate usage or recall an earlier version of fmdtools. Do not expect perfect output from this--you may have to spend some time identifying the problems so they can be fixed!
@@ -548,8 +548,8 @@ To cite a particular version of the fmdtools, you may use:
   author = {{NASA}},
   title = {fmdtools},
   url = {https://github.com/nasa/fmdtools},
-  version = {2.5.4},
-  date = {2026-10-5},
+  version = {2.5.5},
+  date = {2026-10-08},
 }
 ```
 

{fmdtools-2.5.4 → fmdtools-2.5.5}/README.md

@@ -123,7 +123,7 @@ More guidance on development installations is provided in [How to Contribute](ht
 Note that this version reflects the latest version on the ``main`` git branch along with any local changes. To use a particular version of ``fmdtools`` in the context of a development install you can checkout the tag corresponding to the version you want, e.g.:
 
 ```
-   git checkout v2.5.4
+   git checkout v2.5.5
 ```
 
 
@@ -229,9 +229,9 @@ Decent (but **very imperfect**) results can be achieved using LLMs for model dev
 
 ```prompt
 You are an experienced software engineer learning a new Python library for the first time.
-Make an fmdtools simulation of a <enter your system here> using the latest syntax from `https://github.com/nasa/fmdtools/tree/v2.5.1` for all modeling (`fmdtools.define`), simulation (`fmdtools.sim`), and analysis (`fmdtools.analyze`) classes.
+Make an fmdtools simulation of a <enter your system here> using the latest syntax from `https://github.com/nasa/fmdtools/tree/v2.5.5` for all modeling (`fmdtools.define`), simulation (`fmdtools.sim`), and analysis (`fmdtools.analyze`) classes.
 The output should be a bare bones but well-commented .py file with simple behaviors for each class and a small script at the bottom leveraging the built-in capabilities in the `fmdtools.analyze` package instead of directly using external libraries like matplotlib.
-As you are going, explain the structure of each class, the usage of each function call and how they each leverage the correct fmdtools v2.5.1 syntax.
+As you are going, explain the structure of each class, the usage of each function call and how they each leverage the correct fmdtools v2.5.5 syntax.
 ```
 
 The important aspects of this prompt are specifying the URL and version--otherwise the LLM is likely to hallucinate usage or recall an earlier version of fmdtools. Do not expect perfect output from this--you may have to spend some time identifying the problems so they can be fixed!
@@ -277,8 +277,8 @@ To cite a particular version of the fmdtools, you may use:
   author = {{NASA}},
   title = {fmdtools},
   url = {https://github.com/nasa/fmdtools},
-  version = {2.5.4},
-  date = {2026-10-5},
+  version = {2.5.5},
+  date = {2026-10-08},
 }
 ```
 

{fmdtools-2.5.4 → fmdtools-2.5.5/fmdtools.egg-info}/PKG-INFO

@@ -1,10 +1,10 @@
 Metadata-Version: 2.4
 Name: fmdtools
-Version: 2.5.4
+Version: 2.5.5
 Summary: System resilience modelling, simulation, and assessment in Python
 Author: fmdtools Developers
 Maintainer-email: Daniel Hulse <daniel.e.hulse@nasa.gov>
-License: fmdtools version 2.5.4
+License: fmdtools version 2.5.5
         ---------------------------
                                          Apache License
                                    Version 2.0, January 2004
@@ -394,7 +394,7 @@ More guidance on development installations is provided in [How to Contribute](ht
 Note that this version reflects the latest version on the ``main`` git branch along with any local changes. To use a particular version of ``fmdtools`` in the context of a development install you can checkout the tag corresponding to the version you want, e.g.:
 
 ```
-   git checkout v2.5.4
+   git checkout v2.5.5
 ```
 
 
@@ -500,9 +500,9 @@ Decent (but **very imperfect**) results can be achieved using LLMs for model dev
 
 ```prompt
 You are an experienced software engineer learning a new Python library for the first time.
-Make an fmdtools simulation of a <enter your system here> using the latest syntax from `https://github.com/nasa/fmdtools/tree/v2.5.1` for all modeling (`fmdtools.define`), simulation (`fmdtools.sim`), and analysis (`fmdtools.analyze`) classes.
+Make an fmdtools simulation of a <enter your system here> using the latest syntax from `https://github.com/nasa/fmdtools/tree/v2.5.5` for all modeling (`fmdtools.define`), simulation (`fmdtools.sim`), and analysis (`fmdtools.analyze`) classes.
 The output should be a bare bones but well-commented .py file with simple behaviors for each class and a small script at the bottom leveraging the built-in capabilities in the `fmdtools.analyze` package instead of directly using external libraries like matplotlib.
-As you are going, explain the structure of each class, the usage of each function call and how they each leverage the correct fmdtools v2.5.1 syntax.
+As you are going, explain the structure of each class, the usage of each function call and how they each leverage the correct fmdtools v2.5.5 syntax.
 ```
 
 The important aspects of this prompt are specifying the URL and version--otherwise the LLM is likely to hallucinate usage or recall an earlier version of fmdtools. Do not expect perfect output from this--you may have to spend some time identifying the problems so they can be fixed!
@@ -548,8 +548,8 @@ To cite a particular version of the fmdtools, you may use:
   author = {{NASA}},
   title = {fmdtools},
   url = {https://github.com/nasa/fmdtools},
-  version = {2.5.4},
-  date = {2026-10-5},
+  version = {2.5.5},
+  date = {2026-10-08},
 }
 ```
 

{fmdtools-2.5.4 → fmdtools-2.5.5}/fmdtools.egg-info/SOURCES.txt

@@ -131,10 +131,13 @@ tests/test_appended_history_snapshots.py
 tests/test_approach_joint_domains.py
 tests/test_approach_model_timestep.py
 tests/test_approach_phase_map_ownership.py
+tests/test_architecture_vector_inputs.py
 tests/test_array_differences.py
 tests/test_array_random_states.py
 tests/test_array_result_equality.py
 tests/test_array_state_limits.py
+tests/test_array_state_rounding.py
+tests/test_binomial_trial_counts.py
 tests/test_bootstrap_statistic_arguments.py
 tests/test_bounded_array_increments.py
 tests/test_calc_metric_ci_constant_outputs.py
@@ -142,6 +145,7 @@ tests/test_callable_parameter_domains.py
 tests/test_choice_draw_mass.py
 tests/test_choice_probability_precision.py
 tests/test_common.py
+tests/test_comparison_ci_metrics.py
 tests/test_comparison_groups.py
 tests/test_confidence_history_clipping.py
 tests/test_coordinate_cell_boundaries.py
@@ -149,14 +153,22 @@ tests/test_default_fault_sample_timestep.py
 tests/test_degradation_path_boundaries.py
 tests/test_dirichlet_batches.py
 tests/test_disconnected_graph_robustness.py
+tests/test_discrete_grid_values.py
 tests/test_discrete_sample_sizes.py
 tests/test_error_history_plots.py
+tests/test_even_sampling_mass.py
 tests/test_exact_nominal_expectation.py
+tests/test_expectation_without_nominal.py
+tests/test_explicit_history_times.py
 tests/test_explicit_result_file_formats.py
 tests/test_exponential_rayleigh_draw_sizes.py
 tests/test_factor_sort_priority.py
 tests/test_fault_history_paths.py
+tests/test_fault_object_overrides.py
 tests/test_fault_sample_without_nominal.py
+tests/test_fault_space_ownership.py
+tests/test_fault_space_sample_limits.py
+tests/test_fault_summary_paths.py
 tests/test_fault_time_without_traces.py
 tests/test_fault_weight_arrays.py
 tests/test_faulty_scenario_selection.py
@@ -166,17 +178,27 @@ tests/test_generated_history_phase_timestep.py
 tests/test_geom.py
 tests/test_graph.py
 tests/test_graph_edge_comparison.py
+tests/test_graph_zero_seed.py
 tests/test_hierarchical.py
+tests/test_history_alignment_ownership.py
 tests/test_history_phase_identifiers.py
+tests/test_history_sample_indices.py
 tests/test_history_standard_deviation.py
+tests/test_hypergeometric_count_conversion.py
+tests/test_immutable_fault_injection.py
+tests/test_implicit_sampling_start.py
 tests/test_independent_single_replicate_seeds.py
 tests/test_individual_result_paths.py
+tests/test_integer_bound_conversion.py
 tests/test_integer_draw_mass.py
+tests/test_integer_history_differences.py
+tests/test_integer_rate_metrics.py
 tests/test_join_phase_timestep.py
 tests/test_join_phasemaps.py
 tests/test_joint_fault_metadata_order.py
 tests/test_joint_fault_phase_map.py
 tests/test_joint_fault_weights.py
+tests/test_joint_scenario_quadrature.py
 tests/test_json_result_arrays.py
 tests/test_location_scale_draw_sizes.py
 tests/test_lognormal_defaults.py
@@ -188,14 +210,19 @@ tests/test_multivariate_hypergeometric_options.py
 tests/test_multivariate_normal_arguments.py
 tests/test_named_default_restoration.py
 tests/test_named_metric_rounding.py
+tests/test_named_parameter_maps.py
 tests/test_nearest_point_exclusion.py
+tests/test_nested_comparison_inputs.py
+tests/test_nested_comparison_parentage.py
 tests/test_nested_history_summary.py
 tests/test_nested_nominal_groups.py
 tests/test_nested_parameter_constraints.py
+tests/test_nested_sample_metrics.py
 tests/test_nested_scenario_parameters.py
 tests/test_nesting_custom_separators.py
 tests/test_not_equal_constraints.py
 tests/test_object_history_copy.py
+tests/test_offset_time_grids.py
 tests/test_packages.py
 tests/test_parameter_grid_bounds.py
 tests/test_parameter_sample_defaults.py
@@ -204,6 +231,7 @@ tests/test_parameter_subsampling_types.py
 tests/test_pareto_density.py
 tests/test_pathplan.py
 tests/test_permutation_axis_probabilities.py
+tests/test_permuted_nan_probabilities.py
 tests/test_permuted_slice_probabilities.py
 tests/test_phase_map_default_isolation.py
 tests/test_phase_timestep.py
@@ -211,12 +239,14 @@ tests/test_pruned_fault_times.py
 tests/test_ragged_confidence_histories.py
 tests/test_rand.py
 tests/test_random_combination_types.py
+tests/test_reference_risk_calculations.py
 tests/test_result.py
 tests/test_result_constructor_input.py
 tests/test_result_folder_selection.py
 tests/test_sample.py
 tests/test_scenario.py
 tests/test_scenario_defaults.py
+tests/test_scenario_metric_mappings.py
 tests/test_shape_distribution_sizes.py
 tests/test_simple_problem_vector_inputs.py
 tests/test_simulation_exposure.py
@@ -228,6 +258,7 @@ tests/test_time.py
 tests/test_time_reset.py
 tests/test_triangular_arraylike_parameters.py
 tests/test_tuple_table_factors.py
+tests/test_uniform_bound_dtypes.py
 tests/test_uniform_density.py
 tests/test_unit_random_density.py
 tests/test_vonmises_density.py

{fmdtools-2.5.4 → fmdtools-2.5.5}/pyproject.toml

@@ -18,7 +18,7 @@ classifiers = [
 description = "System resilience modelling, simulation, and assessment in Python"
 license = {file="LICENSE"}
 name = "fmdtools"
-version="2.5.4"
+version="2.5.5"
 readme = "README.md"
 requires-python = ">=3.12"
 
@@ -109,3 +109,81 @@ omit = [
     "*/Temp/*"
     ]
 
+[tool.bumpversion]
+current_version = "2.5.5"
+parse = """(?x)
+    (?P<major>0|[1-9]\\d*)\\.
+    (?P<minor>0|[1-9]\\d*)\\.
+    (?P<patch>0|[1-9]\\d*)
+    (?:
+        .                             # dot for pre-release section
+        (?P<stage>[a-zA-Z-]+)         # pre-release label
+        (?P<num>0|[1-9]\\d*)        # pre-release version number
+    )?                                # pre-release section is optional
+"""
+serialize = [
+    "{major}.{minor}.{patch}.{stage}{num}",
+    "{major}.{minor}.{patch}",
+]
+search = "{current_version}"
+replace = "{new_version}"
+regex = false
+ignore_missing_version = false
+tag = true
+sign_tags = true
+tag_name = "v{new_version}"
+tag_message = "Bump version: {current_version} → {new_version}"
+allow_dirty = false
+commit = true
+message = "Bump version: {current_version} → {new_version}"
+commit_args = ""
+
+[tool.bumpversion.parts.stage]
+values = ["dev", "rc", "final"]
+optional_value = "final"
+
+[[tool.bumpversion.files]]
+filename = "src/fmdtools/__init__.py"
+search = "__version__ = '{current_version}'"
+replace = "__version__ = '{new_version}'"
+
+[[tool.bumpversion.files]]
+filename = "conf.py"
+search = "release = '{current_version}'"
+replace = "release = '{new_version}'"
+
+[[tool.bumpversion.files]]
+filename = "LICENSE"
+search = "fmdtools version {current_version}"
+replace = "fmdtools version {new_version}"
+
+[[tool.bumpversion.files]]
+filename = "README.md"
+search = "git checkout v{current_version}"
+replace = "git checkout v{new_version}"
+
+[[tool.bumpversion.files]]
+filename = "README.md"
+search = "https://github.com/nasa/fmdtools/tree/v{current_version}"
+replace = "https://github.com/nasa/fmdtools/tree/v{new_version}"
+
+[[tool.bumpversion.files]]
+filename = "README.md"
+search = "fmdtools v{current_version}"
+replace = "fmdtools v{new_version}"
+
+[[tool.bumpversion.files]]
+filename = "README.md"
+search = "version = {{{current_version}}}"
+replace = "version = {{{new_version}}}"
+
+[[tool.bumpversion.files]]
+filename = "README.md"
+search = "date = {{\\d{{4}}-\\d{{2}}-\\d{{2}}}}"
+replace = "date = {{{now:%Y-%m-%d}}}"
+regex = true
+
+[[tool.bumpversion.files]]
+filename = "docs-source/Intro_to_fmdtools.md"
+search = "fmdtools v{current_version}"
+replace = "fmdtools v{new_version}"

{fmdtools-2.5.4 → fmdtools-2.5.5}/src/fmdtools/__init__.py

@@ -22,6 +22,6 @@ import fmdtools.sim as sim
 import fmdtools.analyze as analyze
 
 __all__ = ['analyze', 'sim', 'define']
-__version__ = '2.5.4'
+__version__ = '2.5.5'
 __copyright__ = "Copyright 2024 United States Government"
 __license__ = "Apache-2.0"

{fmdtools-2.5.4 → fmdtools-2.5.5}/src/fmdtools/analyze/common.py

@@ -95,6 +95,23 @@ def to_include_keys(to_include):
         return tuple(keys)
 
 
+def _integer_difference(val1, val2, compact=False):
+    """Subtract integer values without overflowing their storage dtype."""
+    left, right = np.asarray(val1), np.asarray(val2)
+    dtype = (
+        object
+        if max(left.dtype.itemsize, right.dtype.itemsize) >= 8
+        else np.int64
+    )
+    delta = np.subtract(left, right, dtype=dtype)
+
+    if compact and np.asarray(delta).dtype == object:
+        limits = np.iinfo(np.int64)
+        if np.all((delta >= limits.min) & (delta <= limits.max)):
+            return np.asarray(delta, dtype=np.int64)[()]
+    return delta
+
+
 def diff(val1, val2, difftype='bool'):
     """
     Find inconsistent states between val1, val2.
@@ -102,6 +119,10 @@ def diff(val1, val2, difftype='bool'):
     The difftype option ('diff' (takes the difference), 'bool' (checks if the same),
                          and float (checks if under the provided tolerance))
 
+    Integer arithmetic is widened before subtraction and absolute value. Numeric
+    differences use int64 where possible, or exact Python integers for values
+    outside its range. Boolean comparisons and floating arithmetic are unchanged.
+
     Examples
     --------
     >>> diff([1, 2, 3], [2, 2, 3])
@@ -114,12 +135,25 @@ def diff(val1, val2, difftype='bool'):
             val1 = np.array(val1)
         if isinstance(val2, list):
             val2 = np.array(val2)
+
+        if difftype == 'bool':
+            return val1 != val2
+
+        left, right = np.asarray(val1), np.asarray(val2)
+        integer_values = left.dtype.kind in 'iu' and right.dtype.kind in 'iu'
+
         if difftype == 'diff':
+            if integer_values:
+                return _integer_difference(left, right, compact=True)
             return val1-val2
-        elif difftype == 'bool':
-            return val1 != val2
-        elif isinstance(difftype, float):
-            return abs(val1-val2) > difftype
+
+        if isinstance(difftype, float):
+            if integer_values:
+                return np.abs(_integer_difference(left, right)) > difftype
+            close = np.isclose(
+                left, right, atol=difftype, rtol=0.0, equal_nan=True
+            )
+            return ~close & ~np.isnan(left) & ~np.isnan(right)
     except ValueError as e:
         raise Exception("Unable to diff "+str(val1)+" and "+str(val2)) from e
 
@@ -214,10 +248,18 @@ def load_folder(folder, filetype):
 
 
 def metric_preamble(data, dtype=None, rates=None, r_dtype=None, r_norm=False):
-    """Process data for weighted metrics used by calc_metric and calc_metric_ci."""
+    """Process data for weighted metrics used by calc_metric and calc_metric_ci.
+
+    Integer and boolean rate factors are promoted to float64 before normalization
+    or multiplication, avoiding fixed-width overflow in risk calculations.
+    Unweighted data and floating/object rate arithmetic are unchanged.
+    """
     vals = np.array(data, dtype=dtype)
     if rates is not None:
         rates = np.array(rates, dtype=r_dtype)
+        if rates.dtype.kind in 'biu':
+            # Weight in floating arithmetic before summing or multiplying counts.
+            rates = rates.astype(np.float64)
         if r_norm:
             rates = rates/np.sum(rates)
         if np.size(rates) > 1 and np.size(vals) > 1:

{fmdtools-2.5.4 → fmdtools-2.5.5}/src/fmdtools/analyze/graph/base.py

@@ -850,14 +850,14 @@ class Graph(object):
             number of times to run robustness coefficient algorithm
             (result is averaged over all trials)
         seed : int
-            optional seed to instantiate test with
+            Optional integer seed, including zero. False or None leaves trials unseeded.
 
         Returns
         -------
         RC : robustness coefficient
         """
         g = self.g.to_undirected()
-        if seed:
+        if seed is not False and seed is not None:
             rng = np.random.default_rng(seed=seed)
         else:
             rng = np.random.default_rng()

{fmdtools-2.5.4 → fmdtools-2.5.5}/src/fmdtools/analyze/history.py

@@ -424,12 +424,12 @@ class History(Result):
             return self
 
     def _prep_nom_faulty(self, nomhist={}, align=True):
-        """Create a nominal history of states from the current history."""
+        """Prepare local flattened histories so alignment leaves inputs unchanged."""
         if not nomhist:
             nomhist = self.nominal.flatten()
         else:
             nomhist = nomhist.flatten()
-        faulthist = self._prep_faulty()
+        faulthist = self._prep_faulty().flatten()
         if align:
             faulthist._align(nomhist)
         return nomhist, faulthist
@@ -466,6 +466,9 @@ class History(Result):
         """
         Get history of times when the attributes *attrs deviate from nominal values.
 
+        Align the comparison on local flattened views without truncating either
+        source history. Returned values use the common time range.
+
         Parameters
         ----------
         *attrs : names of attributes
@@ -535,6 +538,9 @@ class History(Result):
         """
         Get fault names associated with the given attributes.
 
+        Match complete fault-container paths in flat or nested histories, including
+        root-level sub-fault flags.
+
         Parameters
         ----------
         *attrs : strs
@@ -545,15 +551,14 @@ class History(Result):
         faults_hist : History
             History of the attrs and their corresponding faults
         """
-        faulthist = self._prep_faulty()
+        faulthist = self._prep_faulty().flatten()
         faults_hist = History()
         if not attrs:
             attrs = self.keys()
         for att in attrs:
             faults_hist[att] = History({k.split('.')[-1]: v for k, v in faulthist.items()
-                                        if ('.'+att+'.m.faults' in k) or
-                                        ('.'+att+'.m.sub_faults' in k) or
-                                        (att+'.m.faults' in k and k.startswith(att))})
+                                        if ('.'+att+'.m.faults.' in '.'+k+'.') or
+                                        ('.'+k).endswith('.'+att+'.m.sub_faults')})
         return faults_hist
 
     def get_faulty_hist(self, *attrs, withtime=True, withtotal=True, operator=np.any):

{fmdtools-2.5.4 → fmdtools-2.5.5}/src/fmdtools/analyze/result.py

@@ -584,29 +584,31 @@ class Result(UserDict):
         **groups : list
             Sets of scenarios to group (e.g. set_1=['scen1', 'scen2'...]).
             Scenario names match at dot-separated path boundaries; selecting a
-            parent scenario also includes its nested scenarios.
+            parent scenario also includes its nested scenarios. Flattened, nested
+            and mixed result storage produce the same groups of leaf paths.
 
         Returns
         -------
         group_hist : History
             Single-level history with structure {group:{scenname.valuename}}
         """
+        flat = self.flatten()
         if not groups:
-            groups = self.get_default_comp_groups()
+            groups = flat.get_default_comp_groups()
         if time not in values:
             values = values + (time, )
         group_hist = self.__class__()
         for group, scens in groups.items():
             if scens == 'default':
-                scens = {k.split('.')[0] for k in self.keys()}
+                scens = {k.split('.')[0] for k in flat.keys()}
             elif isinstance(scens, str):
                 scens = [scens]
-            k_vs = [k for k in self.keys() for scen in scens for v in values
+            k_vs = [k for k in flat.keys() for scen in scens for v in values
                     if k.startswith(scen+'.') and k.endswith(v) and '.t.' not in k]
             if len(k_vs) > 0 and (group not in group_hist):
                 group_hist[group] = self.__class__()
             for k in k_vs:
-                group_hist[group][k] = self[k]
+                group_hist[group][k] = flat[k]
         # Sort into comparison groups
         if not group_hist:
             raise Exception("Invalid comp_groups: " + str(groups) +
@@ -837,6 +839,10 @@ class Result(UserDict):
         """
         Take the expectation of numeric metrics in the result over given scenarios.
 
+        Without a nominal scenario, use the first scenario's numeric fields.
+        A nominal reference is required only for differences from nominal.
+        An empty result returns an empty result of the same type.
+
         Parameters
         ----------
         app : SampleApproach, optional
@@ -859,7 +865,12 @@ class Result(UserDict):
         """
         mh = self.nest(levels=1)
 
-        nomhist = {k: v for k, v in mh.nominal.items() if is_numeric(v)}
+        if difference_from_nominal and 'nominal' not in mh:
+            raise ValueError("A nominal scenario is required for nominal differences.")
+        if not mh:
+            return self.__class__()
+        reference = mh['nominal'] if 'nominal' in mh else next(iter(mh.values()))
+        nomhist = {k: v for k, v in reference.items() if is_numeric(v)}
         newhists = {k: hist for k, hist in mh.items()
                     if with_nominal or k != 'nominal'}
         if app:

{fmdtools-2.5.4 → fmdtools-2.5.5}/src/fmdtools/analyze/tabulate.py

@@ -475,10 +475,10 @@ class FMEA(BaseTab):
 
     >>> FMEA(res, exfs, average_metric=["rate"], sum_metric=["cost"], expected_metric=["cost"], rates="rate").as_table()
                                       average_rate  sum_cost  expected_cost
-    exfxnarch.fxns.ex_fxn2 short               1.5        13             20
-    exfxnarch.fxns.ex_fxn  short               1.5         9             14
-    exfxnarch.fxns.ex_fxn2 no_charge           1.5         5              8
-    exfxnarch.fxns.ex_fxn  no_charge           1.5         1              2
+    exfxnarch.fxns.ex_fxn2 short               1.5        13           20.0
+    exfxnarch.fxns.ex_fxn  short               1.5         9           14.0
+    exfxnarch.fxns.ex_fxn2 no_charge           1.5         5            8.0
+    exfxnarch.fxns.ex_fxn  no_charge           1.5         1            2.0
     >>> FMEA(res, exfs, sum_metric=["rate"], average_metric=["cost"]).as_table()
                                       sum_rate  average_cost
     exfxnarch.fxns.ex_fxn2 short             3           6.5
@@ -543,6 +543,8 @@ class BaseComparison(BaseTab):
         e.g. {'cost': 'average'}. Default is {}
     ci_metrics : list
         Metrics to calculate a confidence interval for (using bootstrap_ci).
+        Each has a point estimate and lower/upper bounds, whether or not it is
+        also listed in metrics. Each distinct metric is calculated once.
         Default is [].
     ci_kwargs : dict
         kwargs to bootstrap_ci
@@ -551,13 +553,14 @@ class BaseComparison(BaseTab):
     def __init__(self, res, scen_groups, metrics=['cost'],
                  default_stat="expected", stats={}, ci_metrics=[], ci_kwargs={}):
 
-        met_dict = {met: {} for met in metrics}
+        all_metrics = list(dict.fromkeys([*metrics, *ci_metrics]))
+        met_dict = {met: {} for met in all_metrics}
         met_dict.update({met+"_lb": {} for met in ci_metrics})
         met_dict.update({met+"_ub": {} for met in ci_metrics})
 
         for fact_tup, scens in scen_groups.items():
             sub_res = res.get_scens(*scens)
-            for met in metrics+ci_metrics:
+            for met in all_metrics:
                 if met in stats:
                     stat = stats[met]
                 else:
@@ -702,7 +705,9 @@ class NestedComparison(BaseComparison):
         Factors (Scenario properties e.g., 'name', 'time', 'var') in samp to take
         statistic over. Default is ['time']
     samps : dict
-        Sample objects used to generate the scenarios. {'name': samp}
+        Sample objects keyed by their parent scenario name, {'name': samp}.
+        Each inner sample contributes only to that parent's factor group;
+        matching factor groups accumulate scenarios across parents.
     samps_factors : list
         Factors (Scenario properties e.g., 'name', 'time', 'var') in samp to take
         statistic over in the apps. Default is ['time']
@@ -713,13 +718,15 @@ class NestedComparison(BaseComparison):
     def __init__(self, res, samp, samp_factors, samps, samps_factors, **kwargs):
         overall_scen_groups = {}
         scen_groups = samp.get_scen_groups(*samp_factors)
-        for n_samp in samps.values():
+        for parent, n_samp in samps.items():
             n_scen_groups = n_samp.get_scen_groups(*samps_factors)
             for scen_group, scens in scen_groups.items():
+                if parent not in scens:
+                    continue
                 for n_scen_group, n_scens in n_scen_groups.items():
                     k = tuple(list(scen_group)+list(n_scen_group))
-                    v = [s+"."+ns for s in scens for ns in n_scens]
-                    overall_scen_groups[k] = v
+                    v = [parent+"."+ns for ns in n_scens]
+                    overall_scen_groups.setdefault(k, []).extend(v)
 
         self.factors = samp_factors + samps_factors
         super().__init__(res, overall_scen_groups, **kwargs)

{fmdtools-2.5.4 → fmdtools-2.5.5}/src/fmdtools/define/base.py

@@ -441,9 +441,14 @@ def nan_to_x(metric, x=0.0):
 
 
 def gen_timerange(start_time, end_time, dt=1.0, min_r=7):
-    """Generate the times in a given interval given the timestep dt."""
-    return round_float(np.arange(start_time, end_time + np.power(10.0, -min_r), dt),
-                       res=dt, min_r=min_r)
+    """Generate a timestep grid anchored at start_time, not at zero.
+
+    Round elapsed steps before restoring the origin so non-grid-aligned starts
+    do not shift or duplicate the generated times. Decimal cleanup and the
+    inclusive endpoint tolerance retain their min_r precision.
+    """
+    elapsed = np.arange(0.0, end_time - start_time + np.power(10.0, -min_r), dt)
+    return np.round(start_time + round_float(elapsed, res=dt, min_r=min_r), min_r)
 
 
 def get_code_attrs(obj):