Merge pull request #100 from PyconUK/imports

Create converter.py in order to avoid circular imports

Author: drvinceknight

docs/howto/obtain_mathematical_representation.rst

@@ -6,7 +6,7 @@ different format.
 Let us schedule a simple conference as described in :ref:`tutorial`::
 
     >>> from conference_scheduler.resources import Slot, Event
-    >>> from conference_scheduler import scheduler
+    >>> from conference_scheduler import scheduler, converter
 
     >>> slots  = [Slot(venue='Big', starts_at='15-Sep-2016 09:30', duration=30, session="A", capacity=200),
     ...           Slot(venue='Big', starts_at='15-Sep-2016 10:00', duration=30, session="A", capacity=200)]
@@ -26,15 +26,15 @@ If we want to recover the mathematical array form of our solution (as described
 in :ref:`mathematical-model`), we use the :code:`scheduler.schedule_to_array`
 function::
 
-    >>> array = scheduler.schedule_to_array(schedule, events=events, slots=slots)
+    >>> array = converter.schedule_to_array(schedule, events=events, slots=slots)
     >>> array
     array([[1, 0],
            [0, 1]], dtype=int8)
 
 We can also return from a mathematical array to the schedule using the
 :code:`scheduler.array_to_schedule` function::
 
-    >>> for item in scheduler.array_to_schedule(array, events=events, slots=slots):
+    >>> for item in converter.array_to_schedule(array, events=events, slots=slots):
     ...     print(f"{item.event.name} at {item.slot.starts_at} in {item.slot.venue}")
     Talk 1 at 15-Sep-2016 09:30 in Big
     Talk 2 at 15-Sep-2016 10:00 in Big

src/conference_scheduler/converter.py

@@ -0,0 +1,117 @@
+"""Convert a schedule between the three possible forms"""
+import numpy as np
+from conference_scheduler.resources import ScheduledItem
+
+
+def solution_to_array(solution, events, slots):
+    """Convert a schedule from solution to array form
+
+    Parameters
+    ----------
+    solution : list or tuple
+        of tuples of event index and slot index for each scheduled item
+    events : list or tuple
+        of :py:class:`resources.Event` instances
+    slots : list or tuple
+        of :py:class:`resources.Slot` instances
+
+    Returns
+    -------
+    np.array
+        An E by S array (X) where E is the number of events and S the
+        number of slots. Xij is 1 if event i is scheduled in slot j and
+        zero otherwise
+
+    Example
+    -------
+    For For 3 events, 7 slots and the solution::
+
+        [(0, 1), (1, 4), (2, 5)]
+
+    The resulting array would be::
+
+        [[0, 1, 0, 0, 0, 0, 0],
+         [0, 0, 0, 0, 1, 0, 0],
+         [0, 0, 0, 0, 0, 1, 0]]
+    """
+    array = np.zeros((len(events), len(slots)), dtype=np.int8)
+    for item in solution:
+        array[item[0], item[1]] = 1
+    return array
+
+
+def solution_to_schedule(solution, events, slots):
+    """Convert a schedule from solution to schedule form
+
+    Parameters
+    ----------
+    solution : list or tuple
+        of tuples of event index and slot index for each scheduled item
+    events : list or tuple
+        of :py:class:`resources.Event` instances
+    slots : list or tuple
+        of :py:class:`resources.Slot` instances
+
+    Returns
+    -------
+    list
+        A list of instances of :py:class:`resources.ScheduledItem`
+    """
+    return [
+        ScheduledItem(
+            event=events[item[0]],
+            slot=slots[item[1]]
+        )
+        for item in solution
+    ]
+
+
+def schedule_to_array(schedule, events, slots):
+    """Convert a schedule from schedule to array form
+
+    Parameters
+    ----------
+    schedule : list or tuple
+        of instances of :py:class:`resources.ScheduledItem`
+    events : list or tuple
+        of :py:class:`resources.Event` instances
+    slots : list or tuple
+        of :py:class:`resources.Slot` instances
+
+    Returns
+    -------
+    np.array
+        An E by S array (X) where E is the number of events and S the
+        number of slots. Xij is 1 if event i is scheduled in slot j and
+        zero otherwise
+    """
+    array = np.zeros((len(events), len(slots)), dtype=np.int8)
+    for item in schedule:
+        array[events.index(item.event), slots.index(item.slot)] = 1
+    return array
+
+
+def array_to_schedule(array, events, slots):
+    """Convert a schedule from array to schedule form
+
+    Parameters
+    ----------
+    array : np.array
+        An E by S array (X) where E is the number of events and S the
+        number of slots. Xij is 1 if event i is scheduled in slot j and
+        zero otherwise
+    events : list or tuple
+        of :py:class:`resources.Event` instances
+    slots : list or tuple
+        of :py:class:`resources.Slot` instances
+
+    Returns
+    -------
+    list
+        A list of instances of :py:class:`resources.ScheduledItem`
+    """
+    scheduled = np.transpose(np.nonzero(array))
+    return [
+        ScheduledItem(event=events[item[0]], slot=slots[item[1]])
+        for item in scheduled
+    ]

src/conference_scheduler/lp_problem/objective_functions.py

@@ -1,4 +1,4 @@
-from conference_scheduler.scheduler import schedule_to_array
+from conference_scheduler.converter import schedule_to_array
 
 
 def capacity_demand_difference(slots, events, X, **kwargs):

src/conference_scheduler/scheduler.py

@@ -1,5 +1,5 @@
-"""Compute a schedule in one of three forms, convert a schedule between forms
-and compute the difference between .
+"""Compute a schedule in one of three forms and compute the difference between
+two schedules.
 
 A schedule can be represented in one of three forms:
 
@@ -11,18 +11,18 @@
 
 import pulp
 import numpy as np
+import conference_scheduler.converter as conv
 import conference_scheduler.lp_problem as lp
 import conference_scheduler.heuristics as heu
 import conference_scheduler.validator as val
 from conference_scheduler.resources import (
-    ScheduledItem, Shape, ChangedEventScheduledItem, ChangedSlotScheduledItem
+    Shape, ChangedEventScheduledItem, ChangedSlotScheduledItem
 )
 
 # __all__ is defined so that we can control the order in which the functions
 # are documented by sphinx.
 __all__ = [
-    'solution', 'array', 'schedule', 'solution_to_array',
-    'solution_to_schedule', 'schedule_to_array', 'array_to_schedule',
+    'solution', 'array', 'schedule',
     'event_schedule_difference', 'slot_schedule_difference']
 
 
@@ -185,7 +185,7 @@ def array(events, slots, objective_function=None, solver=None, **kwargs):
          [0, 0, 0, 0, 1, 0, 0],
          [0, 0, 0, 0, 0, 1, 0]]
     """
-    return solution_to_array(
+    return conv.solution_to_array(
         solution(events, slots, objective_function, solver=solver, **kwargs),
         events, slots
     )
@@ -212,128 +212,12 @@ def schedule(events, slots, objective_function=None, solver=None, **kwargs):
     list
         A list of instances of :py:class:`resources.ScheduledItem`
     """
-    return solution_to_schedule(
+    return conv.solution_to_schedule(
         solution(events, slots, objective_function, solver=solver, **kwargs),
         events, slots
     )
 
 
-# Functions to convert the schedule from one form to another
-
-def solution_to_array(solution, events, slots):
-    """Convert a schedule from solution to array form
-
-    Parameters
-    ----------
-    solution : list or tuple
-        of tuples of event index and slot index for each scheduled item
-    events : list or tuple
-        of :py:class:`resources.Event` instances
-    slots : list or tuple
-        of :py:class:`resources.Slot` instances
-
-    Returns
-    -------
-    np.array
-        An E by S array (X) where E is the number of events and S the
-        number of slots. Xij is 1 if event i is scheduled in slot j and
-        zero otherwise
-
-    Example
-    -------
-    For For 3 events, 7 slots and the solution::
-
-        [(0, 1), (1, 4), (2, 5)]
-
-    The resulting array would be::
-
-        [[0, 1, 0, 0, 0, 0, 0],
-         [0, 0, 0, 0, 1, 0, 0],
-         [0, 0, 0, 0, 0, 1, 0]]
-    """
-    array = np.zeros((len(events), len(slots)), dtype=np.int8)
-    for item in solution:
-        array[item[0], item[1]] = 1
-    return array
-
-
-def solution_to_schedule(solution, events, slots):
-    """Convert a schedule from solution to schedule form
-
-    Parameters
-    ----------
-    solution : list or tuple
-        of tuples of event index and slot index for each scheduled item
-    events : list or tuple
-        of :py:class:`resources.Event` instances
-    slots : list or tuple
-        of :py:class:`resources.Slot` instances
-
-    Returns
-    -------
-    list
-        A list of instances of :py:class:`resources.ScheduledItem`
-    """
-    return [
-        ScheduledItem(
-            event=events[item[0]],
-            slot=slots[item[1]]
-        )
-        for item in solution
-    ]
-
-
-def schedule_to_array(schedule, events, slots):
-    """Convert a schedule from schedule to array form
-
-    Parameters
-    ----------
-    schedule : list or tuple
-        of instances of :py:class:`resources.ScheduledItem`
-    events : list or tuple
-        of :py:class:`resources.Event` instances
-    slots : list or tuple
-        of :py:class:`resources.Slot` instances
-
-    Returns
-    -------
-    np.array
-        An E by S array (X) where E is the number of events and S the
-        number of slots. Xij is 1 if event i is scheduled in slot j and
-        zero otherwise
-    """
-    array = np.zeros((len(events), len(slots)), dtype=np.int8)
-    for item in schedule:
-        array[events.index(item.event), slots.index(item.slot)] = 1
-    return array
-
-
-def array_to_schedule(array, events, slots):
-    """Convert a schedule from array to schedule form
-
-    Parameters
-    ----------
-    array : np.array
-        An E by S array (X) where E is the number of events and S the
-        number of slots. Xij is 1 if event i is scheduled in slot j and
-        zero otherwise
-    events : list or tuple
-        of :py:class:`resources.Event` instances
-    slots : list or tuple
-        of :py:class:`resources.Slot` instances
-
-    Returns
-    -------
-    list
-        A list of instances of :py:class:`resources.ScheduledItem`
-    """
-    scheduled = np.transpose(np.nonzero(array))
-    return [
-        ScheduledItem(event=events[item[0]], slot=slots[item[1]])
-        for item in scheduled
-    ]
-
-
 # Functions to compute the difference between two schedules
 
 

src/conference_scheduler/validator.py

@@ -1,4 +1,4 @@
-from conference_scheduler import scheduler
+from conference_scheduler import converter
 from conference_scheduler.lp_problem import constraints
 
 
@@ -75,7 +75,7 @@ def is_valid_solution(solution, events, slots):
     """
     if len(solution) == 0:
         return False
-    array = scheduler.solution_to_array(solution, events, slots)
+    array = converter.solution_to_array(solution, events, slots)
     return is_valid_array(array, events, slots)
 
 
@@ -97,7 +97,7 @@ def solution_violations(solution, events, slots):
             of a list of strings indicating the nature of the violated
             constraints
     """
-    array = scheduler.solution_to_array(solution, events, slots)
+    array = converter.solution_to_array(solution, events, slots)
     return array_violations(array, events, slots)
 
 
@@ -121,7 +121,7 @@ def is_valid_schedule(schedule, events, slots):
     """
     if len(schedule) == 0:
         return False
-    array = scheduler.schedule_to_array(schedule, events, slots)
+    array = converter.schedule_to_array(schedule, events, slots)
     return is_valid_array(array, events, slots)
 
 
@@ -143,5 +143,5 @@ def schedule_violations(schedule, events, slots):
             of a list of strings indicating the nature of the violated
             constraints
     """
-    array = scheduler.schedule_to_array(schedule, events, slots)
+    array = converter.schedule_to_array(schedule, events, slots)
     return array_violations(array, events, slots)

tests/lp_problem/test_objective_functions.py

@@ -1,6 +1,6 @@
 import numpy as np
 from conference_scheduler.lp_problem import objective_functions as of
-from conference_scheduler.scheduler import array_to_schedule
+from conference_scheduler.converter import array_to_schedule
 
 
 def test_capacity_demand_difference(slots, events, X):

tests/test_converter.py

@@ -0,0 +1,28 @@
+import numpy as np
+from conference_scheduler import converter
+
+
+def test_solution_to_array(valid_solution, valid_array, events, slots):
+    array = converter.solution_to_array(valid_solution, events, slots)
+    assert np.array_equal(array, valid_array)
+    assert all([isinstance(x, np.int8) for x in array.flat])
+
+
+def test_solution_to_schedule(valid_solution, valid_schedule, events, slots):
+    schedule = converter.solution_to_schedule(valid_solution, events, slots)
+    assert type(schedule) is list
+    assert list(schedule) == valid_schedule
+
+
+def test_schedule_to_array(valid_schedule, valid_array, events, slots):
+    array = converter.schedule_to_array(valid_schedule, events, slots)
+    assert np.array_equal(array, valid_array)
+    assert all([isinstance(x, np.int8) for x in array.flat])
+
+
+def test_array_to_schedule(valid_schedule, valid_array, events, slots):
+    schedule = list(
+        converter.array_to_schedule(valid_array, events, slots)
+    )
+    assert type(schedule) is list
+    assert schedule == valid_schedule