default random.Random() rng, np.random.default_rng() elsewhere

Author: graceg571

src/votekit/ballot_generator/ballot_generator.py

@@ -1,9 +1,9 @@
 import math
+import random
 from abc import abstractmethod
 from typing import Optional, Tuple, Union
 
 import numpy as np
-from numpy.random import Generator
 
 from votekit.ballot import Ballot
 from votekit.pref_interval import PreferenceInterval
@@ -85,6 +85,7 @@ def from_params(
         bloc_voter_prop: dict,
         cohesion_parameters: dict,
         alphas: dict,
+        *,
         random_seed: Optional[int] = None,
         **data,
     ):
@@ -126,7 +127,7 @@ def from_params(
                 interval = PreferenceInterval.from_dirichlet(
                     candidates=slate_to_candidates[b],
                     alpha=alphas[current_bloc][b],
-                    rng=rng,
+                    numpy_rng=rng,
                 )
                 intervals[b] = interval
 
@@ -163,7 +164,7 @@ def generate_profile(
         pass
 
     @staticmethod
-    def _round_num(num: float, *, rng: Optional[Generator] = None) -> int:
+    def _round_num(num: float, *, rng: Optional[random.Random] = None) -> int:
         """
 
         Rounds up or down a float randomly.
@@ -176,7 +177,8 @@ def _round_num(num: float, *, rng: Optional[Generator] = None) -> int:
         Returns:
             int: A whole number.
         """
-        rand = np.random.default_rng(rng).random()
+        rng = random.Random() if rng is None else rng
+        rand = rng.random()
         return math.ceil(num) if rand > 0.5 else math.floor(num)
 
     @staticmethod

src/votekit/ballot_generator/bloc_slate_generator/cambridge.py

@@ -276,7 +276,7 @@ def _inner_cambridge_sampler(
         )
 
         pref_profile_by_bloc[bloc] = _convert_slate_ballots_to_profile(
-            config, bloc, slate_ballots, rng=rng
+            config, bloc, slate_ballots, numpy_rng=rng
         )
 
     return pref_profile_by_bloc

src/votekit/ballot_generator/bloc_slate_generator/config/core.py

@@ -11,6 +11,7 @@
 
 import numpy as np
 import pandas as pd
+from numpy.random import Generator
 
 from votekit.ballot_generator.bloc_slate_generator.config.collections import (
     BlocProportions,
@@ -1407,7 +1408,10 @@ def __keycheck_dirichlet_alphas(
                     )
 
     def set_dirichlet_alphas(
-        self, alphas: Union[Mapping[str, Mapping[str, Union[float, int]]], pd.DataFrame]
+        self,
+        alphas: Union[Mapping[str, Mapping[str, Union[float, int]]], pd.DataFrame],
+        *,
+        random_seed: Optional[int] = None,
     ) -> None:
         """
         Set the Dirichlet alphas for the configuration and resample the preference intervals.
@@ -1417,6 +1421,8 @@ def set_dirichlet_alphas(
                 of bloc names to mappings of slate names to their Dirichlet alpha values. Each bloc
                 must have a mapping for every slate defined in slate_to_candidates. All alpha
                 values must be positive finite reals.
+            random_seed (int | None): seed for RNG, allows for reproducible results given the same
+                inputs. Seed set to None by default, different results will be generated each time.
 
         Raises:
             ConfigurationWarning: If preference intervals have already been set without
@@ -1438,7 +1444,9 @@ def set_dirichlet_alphas(
         else:
             self.__alphas = alphas.copy().astype(float)
         self.__clear_alpha_bool = False
-        self.resample_preference_intervals_from_dirichlet_alphas()
+        self.resample_preference_intervals_from_dirichlet_alphas(
+            numpy_rng=np.random.default_rng(seed=random_seed)
+        )
 
     def clear_dirichlet_alphas(self) -> None:
         """Remove the Dirichlet alphas from the configuration."""
@@ -1457,6 +1465,8 @@ def read_dirichlet_alphas(self) -> Optional[pd.DataFrame]:
 
     def resample_preference_intervals_from_dirichlet_alphas(
         self,
+        *,
+        numpy_rng: Optional[Generator] = None,
     ) -> None:
         """
         Resample the preference intervals for each bloc from the current Dirichlet alphas.
@@ -1475,6 +1485,7 @@ def resample_preference_intervals_from_dirichlet_alphas(
                     candidates=list(self.slate_to_candidates[slate]),
                     alpha=float(self.__alphas[slate].loc[bloc]),
                     allow_zero_support=self.allow_zero_support_candidates,
+                    numpy_rng=numpy_rng,
                 )
             preference_dict[bloc] = slate_intervals
         preference_df = convert_preference_map_to_preference_df(preference_dict)

src/votekit/ballot_generator/bloc_slate_generator/cumulative.py

@@ -13,6 +13,7 @@
 
 import apportionment.methods as apportion
 import numpy as np
+from numpy.random import Generator
 
 from votekit.ballot import ScoreBallot
 from votekit.ballot_generator.bloc_slate_generator.config import BlocSlateConfig
@@ -24,7 +25,7 @@
 
 
 def _inner_name_cumulative(
-    config: BlocSlateConfig, total_points: int, *, random_seed: Optional[int] = None
+    config: BlocSlateConfig, total_points: int, *, numpy_rng: Optional[Generator] = None
 ) -> dict[str, ScoreProfile]:
     """
     Inner function to generate cumulative profiles by bloc using the name-Cumulative model.
@@ -33,8 +34,9 @@ def _inner_name_cumulative(
         config (BlocSlateConfig): Configuration object containing all necessary parameters for
             working with a bloc-slate ballot generator.
         total_points (int): The total number of points to distribute among candidates.
-        random_seed (int | None): Seed for RNG, allows for reproducible results given the same
-            inputs. Seed set to None by default, different results will be generated each time.
+        numpy_rng (Generator | None): Random Number Generator, allows for reproducible results given
+            the same inputs. RNG set to None by default, different results will be generated each
+            time.
 
     Returns:
         dict[str, ScoreProfile]: A dictionary whose keys are bloc strings and values are
@@ -51,7 +53,7 @@ def _inner_name_cumulative(
     pp_by_bloc: dict[str, ScoreProfile] = {}
 
     pref_by_bloc = config.get_combined_preference_intervals_by_bloc()
-    rng = np.random.default_rng(seed=random_seed)
+    rng = np.random.default_rng() if numpy_rng is None else numpy_rng
 
     for bloc in bloc_lst:
         num_ballots = int(ballots_per_bloc.get(bloc, 0))
@@ -121,7 +123,8 @@ def name_cumulative_profile_generator(
     if total_points <= 0:
         raise ValueError("total_points must be a positive integer")
 
-    pp_by_bloc = _inner_name_cumulative(config, total_points=total_points, random_seed=random_seed)
+    rng = np.random.default_rng(seed=random_seed)
+    pp_by_bloc = _inner_name_cumulative(config, total_points=total_points, numpy_rng=rng)
 
     pp = ScoreProfile()
     for profile in pp_by_bloc.values():
@@ -169,7 +172,8 @@ def name_cumulative_ballot_generator_by_bloc(
     if total_points <= 0:
         raise ValueError("'total_points' must be a positive integer")
 
-    pp_by_bloc = _inner_name_cumulative(config, total_points=total_points, random_seed=random_seed)
+    rng = np.random.default_rng(seed=random_seed)
+    pp_by_bloc = _inner_name_cumulative(config, total_points=total_points, numpy_rng=rng)
 
     if group_ballots:
         for bloc in pp_by_bloc:

src/votekit/ballot_generator/bloc_slate_generator/name_bradley_terry.py

@@ -22,7 +22,6 @@
 import apportionment.methods as apportion
 import numpy as np
 import pandas as pd
-from numpy.random import Generator
 
 from votekit.ballot import RankBallot
 from votekit.ballot_generator.bloc_slate_generator.config import BlocSlateConfig
@@ -147,7 +146,7 @@ def _check_name_bt_memory(config: BlocSlateConfig) -> None:
 
 
 def _inner_name_bradley_terry(
-    config: BlocSlateConfig, random_seed: Optional[int] = None
+    config: BlocSlateConfig, *, rng: Optional[random.Random] = None
 ) -> dict[str, RankProfile]:
     """
     Sample from the BT distribution using direct sampling.
@@ -159,8 +158,9 @@ def _inner_name_bradley_terry(
     Args:
         config (BlocSlateConfig): Configuration object containing all necessary parameters for
             working with a bloc-slate ballot generator.
-        random_seed (int | None): Seed for RNG, allows for reproducible results given the same
-            inputs. Seed set to None by default, different results will be generated each time.
+        rng (random.Random | None): Random Number Generator, allows for reproducible results given
+            the same inputs. RNG set to None by default, different results will be generated each
+            time.
 
     Returns:
         dict[str, RankProfile]: Generated preference profiles by bloc.
@@ -183,7 +183,7 @@ def _inner_name_bradley_terry(
     pp_by_bloc = {b: RankProfile() for b in bloc_lst}
 
     pref_interval_by_bloc_dict = config.get_combined_preference_intervals_by_bloc()
-    rng = np.random.default_rng(seed=random_seed)
+    rng = random.Random() if rng is None else rng
 
     for bloc in config.bloc_proportions.keys():
         n_ballots = ballots_per_bloc[bloc]
@@ -197,10 +197,10 @@ def _inner_name_bradley_terry(
 
         # The return of this will be a numpy array, so we don't need to make it into a list
         sampled_indices = np.array(
-            rng.choice(
-                a=len(rankings),
-                size=n_ballots,
-                p=probs,
+            rng.choices(
+                range(len(rankings)),
+                k=n_ballots,
+                weights=probs,
             ),
             ndmin=1,
         )
@@ -241,7 +241,7 @@ def _bradley_terry_mcmc(
     burn_in_time: int = 0,
     chain_length: Optional[int] = None,
     *,
-    rng: Optional[Generator] = None,
+    rng: Optional[random.Random] = None,
 ):
     """
     Sample from BT distribution for a given preference interval using MCMC. Defaults
@@ -258,8 +258,9 @@ def _bradley_terry_mcmc(
             chain_length//n_ballots steps from the chain until the desired number of ballots is
             reached. Defaults to None which sets the chain_length to the number of ballots in
             the config.
-        rng (Generator | None): Random Number Generator seeded with a known value for reproducible
-            results. By default, seeded with None to generate different results each time.
+        rng (random.Random | None): Random Number Generator seeded with a known value for
+            reproducible results. By default, seeded with None to generate different results each
+            time.
     """
 
     if chain_length is None:
@@ -280,9 +281,9 @@ def _bradley_terry_mcmc(
     burn_in_time = burn_in_time  # int(10e5)
     if verbose:
         print(f"Burn in time: {burn_in_time}")
-    rng = np.random.default_rng(rng)
+    rng = random.Random() if rng is None else rng
     swap_indices = [
-        (j1, j1 + 1) for j1 in rng.choice(n_candidates - 1, size=n_ballots + burn_in_time)
+        (j1, j1 + 1) for j1 in rng.choices(range(n_candidates - 1), k=n_ballots + burn_in_time)
     ]
 
     for i in range(burn_in_time):
@@ -356,7 +357,7 @@ def _inner_name_bradley_terry_mcmc(
     verbose: bool = False,
     burn_in_time: int = 0,
     chain_length: Optional[int] = None,
-    random_seed: Optional[int] = None,
+    rng: Optional[random.Random] = None,
 ) -> dict[str, RankProfile]:
     """
     Sample from the BT distribution using Markov Chain Monte Carlo.
@@ -372,8 +373,9 @@ def _inner_name_bradley_terry_mcmc(
             chain_length//n_ballots steps from the chain until the desired number of ballots is
             reached. Defaults to None which sets the chain_length to the number of ballots in
             the config.
-        random_seed (int | None): Seed for RNG, allows for reproducible results given the same
-            inputs. Seed set to None by default, different results will be generated each time.
+        rng (Generator | None): Seed for Random Number Generator, allows for reproducible results
+            given the same inputs. RNG set to None by default, different results will be generated
+            each time.
 
     Returns:
         Union[RankProfile, Tuple]
@@ -393,7 +395,7 @@ def _inner_name_bradley_terry_mcmc(
 
     pp_by_bloc = {b: RankProfile() for b in bloc_lst}
     pref_interval_by_bloc_dict = config.get_combined_preference_intervals_by_bloc()
-    rng = np.random.default_rng(seed=random_seed)
+    rng = random.Random() if rng is None else rng
 
     for bloc in bloc_lst:
         n_ballots = ballots_per_bloc[bloc]
@@ -448,7 +450,8 @@ def name_bt_profiles_by_bloc_generator(
     """
     _check_name_bt_memory(config)
     config.is_valid(raise_errors=True)
-    pp_by_bloc = _inner_name_bradley_terry(config, random_seed=random_seed)
+    rng = random.Random(random_seed)
+    pp_by_bloc = _inner_name_bradley_terry(config, rng=rng)
     if group_ballots:
         for bloc in pp_by_bloc:
             pp_by_bloc[bloc] = pp_by_bloc[bloc].group_ballots()
@@ -480,7 +483,8 @@ def name_bt_profile_generator(
     """
     _check_name_bt_memory(config)
     config.is_valid(raise_errors=True)
-    pp_by_bloc = _inner_name_bradley_terry(config, random_seed=random_seed)
+    rng = random.Random(random_seed)
+    pp_by_bloc = _inner_name_bradley_terry(config, rng=rng)
 
     # combine the profiles
     pp = RankProfile()
@@ -525,12 +529,13 @@ def name_bt_profile_generator_using_mcmc(
         RankProfile: Generated preference profile.
     """
     config.is_valid(raise_errors=True)
+    rng = random.Random(random_seed)
     pp_by_bloc = _inner_name_bradley_terry_mcmc(
         config,
         verbose=verbose,
         burn_in_time=burn_in_time,
         chain_length=chain_length,
-        random_seed=random_seed,
+        rng=rng,
     )
     # combine the profiles
     pp = RankProfile()
@@ -576,12 +581,13 @@ def name_bt_profiles_by_bloc_generator_using_mcmc(
         dict[str, RankProfile]: Generated preference profiles by bloc.
     """
     config.is_valid(raise_errors=True)
+    rng = random.Random(random_seed)
     pp_by_bloc = _inner_name_bradley_terry_mcmc(
         config,
         verbose=verbose,
         burn_in_time=burn_in_time,
         chain_length=chain_length,
-        random_seed=random_seed,
+        rng=rng,
     )
 
     if group_ballots:

src/votekit/ballot_generator/bloc_slate_generator/name_plackett_luce.py

@@ -14,6 +14,7 @@
 import apportionment.methods as apportion
 import numpy as np
 import pandas as pd
+from numpy.random import Generator
 
 from votekit.ballot_generator.bloc_slate_generator.config import BlocSlateConfig
 from votekit.pref_profile import RankProfile
@@ -27,7 +28,7 @@ def _inner_name_plackett_luce(
     config: BlocSlateConfig,
     *,
     ballot_length: Optional[int] = None,
-    random_seed: Optional[int] = None,
+    numpy_rng: Optional[Generator] = None,
 ) -> dict[str, RankProfile]:
     """
     Inner function to generate preference profiles by bloc using the name-Plackett-Luce model.
@@ -40,8 +41,9 @@ def _inner_name_plackett_luce(
             working with a bloc-slate ballot generator.
         ballot_length (Optional[int]): Number of ranking positions allowed per ballot. If None,
             this is set to the total number of candidates in the configuration. Defaults to None.
-        random_seed (int | None): seed for RNG, allows for reproducible results given the same
-            inputs. Seed set to None by default, different results will be generated each time.
+        numpy_rng (Generator) | None): Random Number Generator, allows for reproducible results
+            given the same inputs. RNG set to None by default, different results will be generated
+            each time.
 
     Returns:
         dict[str, RankProfile]: Dictionary whose keys are bloc strings and values are
@@ -66,7 +68,7 @@ def _inner_name_plackett_luce(
 
     pp_by_bloc = {b: RankProfile() for b in bloc_lst}
     pref_interval_by_bloc_dict = config.get_combined_preference_intervals_by_bloc()
-    rng = np.random.default_rng(seed=random_seed)
+    rng = np.random.default_rng() if numpy_rng is None else numpy_rng
 
     for bloc in config.blocs:
         n_ballots = ballots_per_bloc[bloc]
@@ -141,7 +143,8 @@ def name_pl_profile_generator(
     config.is_valid(raise_errors=True)
 
     # pp_by_bloc = _inner_name_plackett_luce(config, ballot_length=ballot_length)
-    pp_by_bloc = _inner_name_plackett_luce(config, random_seed=random_seed)
+    rng = np.random.default_rng(seed=random_seed)
+    pp_by_bloc = _inner_name_plackett_luce(config, numpy_rng=rng)
 
     pp = RankProfile(ballots=tuple())
     for profile in pp_by_bloc.values():
@@ -182,7 +185,8 @@ def name_pl_profiles_by_bloc_generator(
     config.is_valid(raise_errors=True)
 
     # pp_by_bloc = _inner_name_plackett_luce(config, ballot_length=ballot_length)
-    pp_by_bloc = _inner_name_plackett_luce(config, random_seed=random_seed)
+    rng = np.random.default_rng(seed=random_seed)
+    pp_by_bloc = _inner_name_plackett_luce(config, numpy_rng=rng)
 
     if group_ballots:
         for bloc, profile in pp_by_bloc.items():