Mesa 2.2.3__tar.gz → 2.2.4__tar.gz

{mesa-2.2.3 → mesa-2.2.4}/HISTORY.md

@@ -2,6 +2,22 @@
 title: Release History
 ---
 
+# 2.2.4 (2024-01-26)
+
+## Highlights
+Mesa v2.2.4 is a small but important bugfix release for the 2.2 release series. It fixes an essential bug in where agents weren't shuffled in the `RandomActivation` scheduler (effectively making it sequential activation)([#2007](https://github.com/projectmesa/mesa/pull/2007)). It also fixes a small behaviour change in `RandomActivationByType.agents_by_type()` ([#1996](https://github.com/projectmesa/mesa/pull/1996)). Furthermore, this release adds an internal clock to the `Model`, which allows to use a Mesa model without a scheduler (using the `AgentSet` API)([#1942](https://github.com/projectmesa/mesa/pull/1942)).
+
+Updating from previous 2.2 releases is highly recommended, especially when using the `RandomActivation` scheduler.
+
+## What's Changed
+### 🛠 Enhancements made
+* refactor: Remove dependence on model.schedule, add clock to Model by @rht in https://github.com/projectmesa/mesa/pull/1942
+### 🐛 Bugs fixed
+* Fix AgentSet inplace shuffle (and thus RandomActivation), add tests by @EwoutH in https://github.com/projectmesa/mesa/pull/2007
+* fix: Reverse dict key and value for agents_by_type by @rht in https://github.com/projectmesa/mesa/pull/1996
+
+**Full Changelog**: https://github.com/projectmesa/mesa/compare/v2.2.3...v2.2.4
+
 # 2.2.3 (2024-01-22)
 
 ## Highlights

{mesa-2.2.3 → mesa-2.2.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: Mesa
-Version: 2.2.3
+Version: 2.2.4
 Summary: Agent-based modeling (ABM) in Python
 Project-URL: homepage, https://github.com/projectmesa/mesa
 Project-URL: repository, https://github.com/projectmesa/mesa

{mesa-2.2.3 → mesa-2.2.4}/mesa/__init__.py

@@ -25,7 +25,7 @@ __all__ = [
 ]
 
 __title__ = "mesa"
-__version__ = "2.2.3"
+__version__ = "2.2.4"
 __license__ = "Apache 2.0"
 _this_year = datetime.datetime.now(tz=datetime.timezone.utc).date().year
 __copyright__ = f"Copyright {_this_year} Project Mesa Team"

{mesa-2.2.3 → mesa-2.2.4}/mesa/agent.py

@@ -54,6 +54,7 @@ class Agent:
         except AttributeError:
             # model super has not been called
             self.model.agents_ = defaultdict(dict)
+            self.model.agents_[type(self)][self] = None
             self.model.agentset_experimental_warning_given = False
 
             warnings.warn(

{mesa-2.2.3 → mesa-2.2.4}/mesa/datacollection.py

@@ -14,8 +14,7 @@ name.
 
 When the collect() method is called, each model-level function is called, with
 the model as the argument, and the results associated with the relevant
-variable. Then the agent-level functions are called on each agent in the model
-scheduler.
+variable. Then the agent-level functions are called on each agent.
 
 Additionally, other objects can write directly to tables by passing in an
 appropriate dictionary object for a table row.
@@ -30,8 +29,7 @@ The DataCollector then stores the data it collects in dictionaries:
 Finally, DataCollector can create a pandas DataFrame from each collection.
 
 The default DataCollector here makes several assumptions:
-    * The model has a schedule object called 'schedule'
-    * The schedule has an agent list called agents
+    * The model has an agent list called agents
     * For collecting agent-level variables, agents must have a unique_id
 """
 import contextlib
@@ -67,7 +65,7 @@ class DataCollector:
 
         Model reporters can take four types of arguments:
         1. Lambda function:
-           {"agent_count": lambda m: m.schedule.get_agent_count()}
+           {"agent_count": lambda m: len(m.agents)}
         2. Method of a class/instance:
            {"agent_count": self.get_agent_count} # self here is a class instance
            {"agent_count": Model.get_agent_count} # Model here is a class
@@ -180,11 +178,14 @@ class DataCollector:
         rep_funcs = self.agent_reporters.values()
 
         def get_reports(agent):
-            _prefix = (agent.model.schedule.steps, agent.unique_id)
+            _prefix = (agent.model._steps, agent.unique_id)
             reports = tuple(rep(agent) for rep in rep_funcs)
             return _prefix + reports
 
-        agent_records = map(get_reports, model.schedule.agents)
+        agent_records = map(
+            get_reports,
+            model.schedule.agents if hasattr(model, "schedule") else model.agents,
+        )
         return agent_records
 
     def collect(self, model):
@@ -207,7 +208,7 @@ class DataCollector:
 
         if self.agent_reporters:
             agent_records = self._record_agents(model)
-            self._agent_records[model.schedule.steps] = list(agent_records)
+            self._agent_records[model._steps] = list(agent_records)
 
     def add_table_row(self, table_name, row, ignore_missing=False):
         """Add a row dictionary to a specific table.

{mesa-2.2.3 → mesa-2.2.4}/mesa/experimental/jupyter_viz.py

@@ -178,7 +178,7 @@ def ModelController(model, play_interval, current_step, reset_counter):
     def do_step():
         model.step()
         previous_step.value = current_step.value
-        current_step.value += 1
+        current_step.value = model._steps
 
     def do_play():
         model.running = True

{mesa-2.2.3 → mesa-2.2.4}/mesa/model.py

@@ -13,11 +13,13 @@ import warnings
 from collections import defaultdict
 
 # mypy
-from typing import Any
+from typing import Any, Union
 
 from mesa.agent import Agent, AgentSet
 from mesa.datacollection import DataCollector
 
+TimeT = Union[float, int]
+
 
 class Model:
     """Base class for models in the Mesa ABM library.
@@ -68,6 +70,9 @@ class Model:
         self.current_id = 0
         self.agents_: defaultdict[type, dict] = defaultdict(dict)
 
+        self._steps: int = 0
+        self._time: TimeT = 0  # the model's clock
+
         # Warning flags for current experimental features. These make sure a warning is only printed once per model.
         self.agentset_experimental_warning_given = False
 
@@ -112,6 +117,11 @@ class Model:
     def step(self) -> None:
         """A single step. Fill in here."""
 
+    def _advance_time(self, deltat: TimeT = 1):
+        """Increment the model's steps counter and clock."""
+        self._steps += 1
+        self._time += deltat
+
     def next_id(self) -> int:
         """Return the next unique ID for agents, increment current_id"""
         self.current_id += 1

{mesa-2.2.3 → mesa-2.2.4}/mesa/time.py

@@ -73,6 +73,8 @@ class BaseScheduler:
         self.model = model
         self.steps = 0
         self.time: TimeT = 0
+        self._original_step = self.step
+        self.step = self._wrapped_step
 
         if agents is None:
             agents = []
@@ -115,6 +117,11 @@ class BaseScheduler:
         self.steps += 1
         self.time += 1
 
+    def _wrapped_step(self):
+        """Wrapper for the step method to include time and step updating."""
+        self._original_step()
+        self.model._advance_time()
+
     def get_agent_count(self) -> int:
         """Returns the current number of agents in the queue."""
         return len(self._agents)
@@ -143,8 +150,8 @@ class BaseScheduler:
 
     def do_each(self, method, shuffle=False):
         if shuffle:
-            self.agents.shuffle(inplace=True)
-        self.agents.do(method)
+            self._agents.shuffle(inplace=True)
+        self._agents.do(method)
 
 
 class RandomActivation(BaseScheduler):
@@ -299,7 +306,7 @@ class RandomActivationByType(BaseScheduler):
 
         agentsbytype = defaultdict(dict)
         for k, v in self._agents_by_type.items():
-            agentsbytype[k] = {agent: agent.unique_id for agent in v}
+            agentsbytype[k] = {agent.unique_id: agent for agent in v}
 
         return agentsbytype
 

{mesa-2.2.3 → mesa-2.2.4}/tests/test_agent.py

@@ -251,3 +251,16 @@ def test_agentset_select_by_type():
     # Test with no type specified (should select all agents)
     all_agents = agentset.select()
     assert len(all_agents) == len(mixed_agents)
+
+
+def test_agentset_shuffle():
+    model = Model()
+    test_agents = [TestAgent(model.next_id(), model) for _ in range(12)]
+
+    agentset = AgentSet(test_agents, model=model)
+    agentset = agentset.shuffle()
+    assert not all(a1 == a2 for a1, a2 in zip(test_agents, agentset))
+
+    agentset = AgentSet(test_agents, model=model)
+    agentset.shuffle(inplace=True)
+    assert not all(a1 == a2 for a1, a2 in zip(test_agents, agentset))

{mesa-2.2.3 → mesa-2.2.4}/tests/test_time.py

@@ -224,6 +224,28 @@ class TestRandomActivation(TestCase):
         agent_ids = {agent.unique_id for agent in model.agents}
         assert all(entry in agent_ids for entry in keys)
 
+    def test_not_sequential(self):
+        model = MockModel(activation=RANDOM)
+        # Create 10 agents
+        for _ in range(10):
+            model.schedule.add(MockAgent(model.next_id(), model))
+        # Run 3 steps
+        for _ in range(3):
+            model.step()
+        # Filter out non-integer elements from the log
+        filtered_log = [item for item in model.log if isinstance(item, int)]
+
+        # Check that there are no 18 consecutive agents id's in the filtered log
+        total_agents = 10
+        assert not any(
+            all(
+                (filtered_log[(i + j) % total_agents] - filtered_log[i]) % total_agents
+                == j % total_agents
+                for j in range(18)
+            )
+            for i in range(len(filtered_log))
+        ), f"Agents are activated sequentially:\n{filtered_log}"
+
 
 class TestSimultaneousActivation(TestCase):
     """