Patch sharrow sandag (#1009)

* descriptive error

* allow partial success in setting up skims

- if there is no time period we should still be able to get time-agnostic values

* comments for better context on why

* require sharrow 2.15

* update lock file

* blacken

* update docstring

* address whitespace

Author: jpn--

activitysim/abm/models/trip_destination.py

@@ -1496,23 +1496,42 @@ def run_trip_destination(
             for primary_purpose, trips_segment in nth_trips.groupby(
                 "primary_purpose", observed=True
             ):
-                choices, destination_sample = choose_trip_destination(
-                    state,
-                    primary_purpose,
-                    trips_segment,
-                    alternatives,
-                    tours_merged,
-                    model_settings,
-                    want_logsums,
-                    want_sample_table,
-                    size_term_matrix,
-                    skim_hotel,
-                    estimator,
-                    chunk_size,
-                    trace_label=tracing.extend_trace_label(
-                        nth_trace_label, primary_purpose
-                    ),
-                )
+                try:
+                    choices, destination_sample = choose_trip_destination(
+                        state,
+                        primary_purpose,
+                        trips_segment,
+                        alternatives,
+                        tours_merged,
+                        model_settings,
+                        want_logsums,
+                        want_sample_table,
+                        size_term_matrix,
+                        skim_hotel,
+                        estimator,
+                        chunk_size,
+                        trace_label=tracing.extend_trace_label(
+                            nth_trace_label, primary_purpose
+                        ),
+                    )
+                except KeyError as err:
+                    if err.args[0] == "purpose_index_num":
+                        logger.error(
+                            """
+
+                        When using the trip destination model with sharrow, it is necessary
+                        to set a value for `purpose_index_num` in the trip destination 
+                        annotate trips preprocessor.  This allows for an optimized compiled 
+                        lookup of the size term from the array of size terms.  The value of
+                        `purpose_index_num` should be the integer column position in the size 
+                        matrix, with usual zero-based numpy indexing semantics (i.e. the first 
+                        column is zero).  The preprocessor expression most likely needs to be
+                        "size_terms.get_cols(df.purpose)" unless some unusual transform of 
+                        size terms has been employed.
+
+                        """
+                        )
+                    raise
 
                 choices_list.append(choices)
                 if want_sample_table:

activitysim/core/simulate.py

@@ -989,7 +989,7 @@ def to_array(x):
 #     return utilities
 
 
-def set_skim_wrapper_targets(df, skims):
+def set_skim_wrapper_targets(df, skims, allow_partial_success: bool = True):
     """
     Add the dataframe to the SkimWrapper object so that it can be dereferenced
     using the parameters of the skims object.
@@ -1007,6 +1007,11 @@ def set_skim_wrapper_targets(df, skims):
         dataframe that comes back from interacting choosers with
         alternatives.  See the skims module for more documentation on how
         the skims object is intended to be used.
+    allow_partial_success : bool, optional
+        If True (default), failures to set skim targets for some skim objects
+        (for example due to missing required columns in `df`) will be collected
+        and logged as warnings but will not raise an exception. If False, any
+        such failure will be raised immediately, preventing partial success.
     """
 
     skims = (
@@ -1016,13 +1021,31 @@ def set_skim_wrapper_targets(df, skims):
         if isinstance(skims, dict)
         else [skims]
     )
+    problems = []
 
     # assume any object in skims can be treated as a skim
     for skim in skims:
         try:
             skim.set_df(df)
         except AttributeError:
+            # sometimes when passed as a dict, the skims have a few keys given as
+            # settings or constants, which are not actually "skim" objects and have
+            # no `set_df` attribute.  This is fine and we just let them pass.
             pass
+        except AssertionError as e:
+            # An assertion error will get triggered if the columns of `df` are
+            # missing one of the required keys needed to look up values in the
+            # skims.  This may not be a problem, if this particular set of skims
+            # is not actually used in this model component.  So we'll warn about
+            # it but usually not raise a showstopping error.
+            problems.append(e)
+            if not allow_partial_success:
+                raise
+
+    if problems:
+        # if problems were discovered, log them as warnings
+        for problem in problems:
+            logger.warning(str(problem))
 
 
 #

pyproject.toml

@@ -26,7 +26,7 @@ dependencies = [
     "requests >= 2.7",
     "scikit-learn >= 1.2",
     "setuptools>=80.9.0",
-    "sharrow >= 2.9.1",
+    "sharrow>=2.15",
     "sparse",
     "tables >= 3.9",  # pytables is tables in pypi
     "xarray >= 2024.05",