dicetables 2.1.2

get all combinations for any set of dice

CHANGELOG

• DiceTable signature changed. DiceTable({1: 1, 2: 1}, DiceRecord({Die(2): 1})

  • DiceTable.dice_data() now returns DiceRecord obj.

  • DiceRecord

    • signature DiceRecord({die: number})

    • is immutable

    • can compare with __eq__ and __ne__ only

• Factory getter keys = [‘calc_includes_zeroes’, ‘dice_data’, ‘get_dict’]

  • adding new getter no longer needs getter key. now is (<getter_name>, <default>, ‘property’ or ‘method’)

  • Factory.add_getter(<getter_name>, <default>, ‘property’ or ‘method’)

  • or as class variable: new_keys = [(<getter_name>, <default>, ‘property’ or ‘method’), …] see inheritance

• all IntegerEvents have __eq__ and __ne__

  • different classes override

  • has strict type checking so DiceTable can never equal DetailedDiceTable

a module for statistics of die rolls and other events

This module uses DiceTable and AdditiveEvents to combine dice and other events that can be added together. It is used to figure out the probability of events occurring. For instance, if you roll 100 six-sided dice, the chance of rolling any number between 100 and 300 is 0.15 percent.

contents:

• THE BASICS

• Die Classes

• AdditiveEvents And IntegerEvents

• DiceTable And DetailedDiceTable

• EventsInformation And EventsCalculations

• Inheritance

• HOW TO GET ERRORS AND BUGS

• CHANGES

  • 0.4_to_1.0

  • 1.0_to_2.0

THE BASICS

Here’s a quick bit of math. if you combine a 2-sided die and a 3-sided die,

you get the following combinations.

(1,1) / (1, 2) (2, 1) / (2, 2), (1, 3) / (2, 3):

• roll - 2: 1 occurrence (1 in 6 chance)

• roll - 3: 2 occurrences (2 in 6 chance)

• roll - 4: 2 occurrences (2 in 6 chance)

• roll - 5: 1 occurrence (1 in 6 chance)

In [1]: import dicetables as dt

In [2]: new = dt.DiceTable.new()

In [3]: one_two_sided = new.add_die(dt.Die(2), times=1)

In [4]: one_two_sided_one_three_sided = one_two_sided.add_die(dt.Die(3), 1)

In [5]: one_two_sided_one_three_sided.get_dict()
Out[5]: {2: 1, 3: 2, 4: 2, 5: 1}

In [6]: one_two_sided.get_dict()
out[6]: {1: 1, 2: 1}

In [7]: new.get_dict()
out[7]: {0: 1}

Here are basic table functions. note that times added defaults to one.:

In [4]: table = dt.DiceTable.new().add_die(dt.Die(2)).add_die(dt.Die(3))

In [5]: str(table)
Out[5]: '1D2\n1D3'

In [6]: table = table.add_die(dt.Die(2), 100)

In [7]: table = table.remove_die(dt.Die(2), 99)

In [17]: print(table)
2D2
1D3

In [20]: table.get_list()
Out[20]: [(Die(2), 2), (Die(3), 1)]  <-- sorted according to die

In [22]: table.number_of_dice(dt.Die(10))
Out[22]: 0

In [22]: table.number_of_dice(dt.Die(2))
Out[22]: 2

In [21]: print(table.weights_info())
2D2
    No weights

1D3
    No weights

To get useful information, use EventsInformation object and EventsCalculations object:

In [1]: table = dt.DiceTable.new()
In [2]: table = table.add_die(dt.StrongDie(dt.Die(2), 3), 2)

In [3]: table.get_dict()
Out[3]: {6: 1, 9: 2, 12: 1}

In [4]: info = dt.EventsInformation(table)

In [5]: info.all_events()
Out[5]: [(6, 1), (9, 2), (12, 1)]


In [6]: info.all_events_include_zeroes()
Out[6]: [(6, 1), (7, 0), (8, 0), (9, 2), (10, 0), (11, 0), (12, 1)]

In [7]: info.events_keys()
Out[7]: [6, 9, 12]

In [8]: info.events_range()
Out[8]: (6, 12)

In [9]: info.get_event(4)
Out[9]: (4, 0)

In [11]: info.get_range_of_events(7, 13)
Out[11]: [(7, 0), (8, 0), (9, 2), (10, 0), (11, 0), (12, 1)]

In [12]: info.biggest_event()
Out[12]: (9, 2)

In [13]: info.total_occurrences()
Out[13]: 4

In [14]: calc = dt.EventsCalculations(table)

In [15]: calc.mean()
Out[15]: 9.0

In [16]: calc.stddev()
Out[16]: 2.1213

In [17]: calc.percentage_points()
Out[17]: [(6, 25.0), (7, 0.0), (8, 0.0), (9, 50.0), (10, 0.0), (11, 0.0), (12, 25.0)]

In [18]: print(calc.full_table_string())
 6: 1
 7: 0
 8: 0
 9: 2
10: 0
11: 0
12: 1

In [19]: without_zeroes = EventsCalculations(table, include_zeroes=False)

In [20]: print(without_zeroes.full_table_string())
 6: 1
 9: 2
12: 1

In [21]: stats_str = "{} occurred {} times out of {} combinations.\nThat's a one in {} chance or {}%"

In [22]: print(stats_str.format(*without_zeroes.stats_strings([1, 2, 5, 8, 9, 10])))
1-2, 5, 8-10 occurred 2 times out of 4 combinations.
That's a one in 2.000 chance or 50.00%

In [23]: without_zeroes.percentage_axes()
Out[23]: [(6, 9, 12), (25.0, 50.0, 25.0)]

DetailedDiceTable keeps a copy of these objects at .info and .calc calc_includes_zeros defaults to True:

In [12]: d_table = dt.DetailedDiceTable.new()

In [13]: d_table.info.events_range()
Out[13]: (0, 0)

In [14]: d_table.calc.mean()
Out[14]: 0.0

In [15]: d_table = d_table.add_die(dt.Die(6), 100)

In [16]: d_table.info.events_range()
Out[16]: (100, 600)

In [17]: d_table.calc.mean()
Out[17]: 350.0

You may also access this functionality with wrapper functions:

• events_range

• mean

• stddev

• stats

• full_table_string

• percentage_points

• percentage_axe

In [43]: silly_table = dt.AdditiveEvents({1: 123456, 100: 12345*10**1000})

In [47]: print(dt.full_table_string(silly_table, include_zeroes=False))
  1: 123,456
100: 1.234e+1004

In [49]: stats_info = dt.stats(silly_table, list(range(-5000, 5)))

In [51]: print(stats_str.format(*stats_info))
(-5,000)-4 occurred 123,456 times out of 1.234e+1004 combinations.
That's a one in 1.000e+999 chance or 1.000e-997%

Finally, here are all the kinds of dice you can add

• dt.Die(6)

• dt.ModDie(6, -2)

• dt.WeightedDie({1:1, 2:5, 3:2})

• dt.ModWeightedDie({1:1, 2:5, 3:2}, 5)

• dt.StrongDie(dt.Die(6), 5)

That’s all of the basic implementation. The rest of this is details about base classes, details of the die classes, details of dicetable classes, what causes errors and the changes from the previous version.

Top

Die Classes

All dice are subclasses of dicetables.eventsbases.protodie.ProtoDie, which is a subclass of dicetables.eventsbases.integerevents.IntegerEvents. They all require implementations of get_size(), get_weight(), weight_info(), multiply_str(number), __str__(), __repr__() and get_dict() (the final one is a requirement of all IntegerEvents).

They are all immutable , hashable and rich-comparable. Multiple names can safely point to the same instance of a Die, they can be used in sets and dictionary keys and they can be sorted with any other kind of die. Comparisons are done by (size, weight, get_dict, __repr__(as a last resort)). So:

In [54]: dice_list
Out[54]:
[ModDie(2, 0),
 WeightedDie({1: 1, 2: 1}),
 Die(2),
 ModWeightedDie({1: 1, 2: 1}, 0),
 StrongDie(Die(2), 1),
 StrongDie(WeightedDie({1: 1, 2: 1}), 1)]

In [58]: [die.get_dict() == {1: 1, 2: 1} for die in dice_list]
Out[58]: [True, True, True, True, True, True]

In [56]: sorted(dice_list)
Out[56]:
[Die(2),
 ModDie(2, 0),
 StrongDie(Die(2), 1),
 ModWeightedDie({1: 1, 2: 1}, 0),
 StrongDie(WeightedDie({1: 1, 2: 1}), 1),
 WeightedDie({1: 1, 2: 1})]

In [67]: [die == dt.Die(2) for die in sorted(dice_list)]
Out[67]: [True, False, False, False, False, False]

In [61]: my_set = {dt.Die(6)}

In [62]: my_set.add(dt.Die(6))

In [63]: my_set
Out[63]: {Die(6)}

In [64]: my_set.add(dt.ModDie(6, 0))

In [65]: my_set
Out[65]: {Die(6), ModDie(6, 0)}

The dice:

Die

A basic die. dt.Die(4) rolls 1, 2, 3, 4 with equal weight

No added methods

ModDie

A die with a modifier. The modifier is added to each die roll. dt.ModDie(4, -2) rolls -1, 0, 1, 2 with equal weight.

added methods:

• .get_modifier()

WeightedDie

A die that rolls different rolls with different frequencies. dt.WeightedDie({1:1, 3:3, 4:6}) is a 4-sided die. It rolls 4 six times as often as 1, rolls 3 three times as often as 1 and never rolls 2

added methods:

• .get_raw_dict()

ModWeightedDie

A die with a modifier that rolls different rolls with different frequencies. dt.ModWeightedDie({1:1, 3:3, 4:6}, 3) is a 4-sided die. 3 is added to all die rolls. The same as WeightedDie.

added methods:

• .get_raw_dict()

• .get_modifier()

StrongDie

A die that is a strong version of any other die (including another StrongDie if you’re feeling especially silly). So a StrongDie with a multiplier of 2 would add 2 for each 1 that was rolled.

dt.StrongDie(dt.Die(4), 5) is a 4-sided die that rolls 5, 10, 15, 20 with equal weight. dt.StrongDie(dt.Die(4), -1) is a 4 sided die that rolls -1, -2, -3, -4.

added methods:

• .get_multiplier()

• .get_input_die()

Top

AdditiveEvents And IntegerEvents

All tables and dice inherit from dicetables.eventsbases.IntegerEvents. All subclasses of IntegerEvents need the method get_dict() which returns {event: occurrences, …} for each NON-ZERO occurrence. When you instantiate any subclass, it checks to make sure you’re get_dict() is legal.

Any child of IntegerEvents has access to __eq__ and __ne__ evaluated by type and then get_dict(). It can be compared to any object and two events that are not the exact same class will be !=.

Any of the classes that take a dictionary of events as input scrub the zero occurrences out of the dictionary for you.

In [19]: dt.DiceTable({1: 1, 2:0}, {}).get_dict()
Out[19]: {1: 1}

In [20]: dt.AdditiveEvents({1: 2, 3: 0, 4: 1}).get_dict()
Out[20]: {1: 2, 4: 1}

In [21]: dt.ModWeightedDie({1: 2, 3: 0, 4: 1}, -5).get_dict()
Out[21]: {-4: 2, -1: 1}

AdditiveEvents is the parent of DiceTable. It has the class method new() which returns the identity. This method is inherited by its children. You can add and remove events using the “.combine” method which tries to pick the fastest combining algorithm. You can pick it yourself by calling “.combine_by_<algorithm>”. You can combine and remove DiceTable, AdditiveEvents, Die or any other IntegerEvents with the “combine” and “remove” methods, but there’s no record of it. AdditiveEvents has __eq__ method that tests type and get_dict(). This is inherited from IntegerEvents.:

In [32]: three_D2 = dt.AdditiveEvents.new().combine_by_dictionary(dt.Die(2), 3)

In [33]: also_three_D2 = dt.AdditiveEvents({3: 1, 4: 3, 5: 3, 6: 1})

In [34]: still_three_D2 = dt.AdditiveEvents.new().combine(dt.AdditiveEvents({1: 1, 2: 1}), 3)

In [35]: three_D2.get_dict() == also_three_D2.get_dict() == still_three_D2.get_dict()
Out[35]: True

In [36]: identity = three_D2.remove(dt.Die(2), 3)

In [37]: identity.get_dict() == dt.AdditiveEvents.new().get_dict()
Out[37]: True

In [38]: identity == dt.AdditiveEvents.new()
Out[38]: True

In [41]: print(three_D2)
table from 3 to 6

In [42]: twenty_one_D2 = three_D2.combine_by_indexed_values(three_D2, 6)

In [43]: twenty_one_D2_five_D4 = twenty_one_D2.combine_by_flattened_list(dt.Die(4), 5)

In [44]: five_D4 = twenty_one_D2_five_D4.remove(dt.Die(2), 21)

In [45]: dt.DiceTable.new().add_die(dt.Die(4), 5).get_dict() == five_D4.get_dict()
Out[45]: True

In [45]: dt.DiceTable.new().add_die(dt.Die(4), 5) == five_D4
Out[45]: False  <-- DiceTable is not AdditiveEvents

Since DiceTable is the child of AdditiveEvents, it can do all this combining and removing, but it won’t be recorded in the dice record.

Top

DiceTable And DetailedDiceTable

You can instantiate any DiceTable or DetailedDiceTable with any data you like. This allows you to create a DiceTable from stored information or to copy. Please note that the “dice_data” method is ambiguously named on purpose. It’s function is to get correct input to instantiate a new DiceTable, whatever that happens to be. To get consistent output, use “get_list”. Equality testing is by type, get_dict(), dice_data() (and calc_includes_zeroes for DetailedDiceTable).

In [14]: old = dt.DiceTable.new()

In [16]: old = old.add_die(dt.Die(6), 100)

In [17]: events_record = old.get_dict()

In [18]: dice_record = old.dice_data()

In [19]: new = dt.DiceTable(events_record, dice_record)

In [20]: print(new)
100D6

In [21]: record = dt.DiceRecord({dt.Die(6): 100})

In [22]: also_new = dt.DetailedDiceTable(new.get_dict(), record, calc_includes_zeroes=False)

In [46]: old.get_dict() == new.get_dict() == also_new.get_dict()
Out[46]: True

In [47]: old.get_list() == new.get_list() == also_new.get_list()
Out[47]: True

In [47]: old == new
Out[47]: True

In [47]: old == also_new
Out[47]: False  <- by type

In [47]: isinstance(also_new, DiceTable)
Out[47]: True

In [47]: type(also_new) is DiceTable
Out[47]: False

DetailedDiceTable.calc_includes_zeroes defaults to True. It is as follows.

In [85]: d_table = dt.DetailedDiceTable.new()

In [86]: d_table.calc_includes_zeroes
out[86]: True

In [87]: d_table = d_table.add_die(dt.StrongDie(dt.Die(2), 2))

In [88]: print(d_table.calc.full_table_string())

2: 1
3: 0
4: 1

In [89]: d_table = d_table.switch_boolean()

In [90]: the_same = dt.DetailedDiceTable({2: 1, 4: 1}, d_table.dice_data(), False)

In [91]: print(d_table.calc.full_table_string())
2: 1
4: 1

In [92]: print(the_same.calc.full_table_string())
2: 1
4: 1

In [93]: d_table = d_table.add_die(1, dt.StrongDie(dt.Die(2), 2))


In [94]: print(d_table.calc.full_table_string())
4: 1
6: 2
8: 1

In [95]: d_table = d_table.switch_boolean()

In [96]: print(d_table.calc.full_table_string())
4: 1
5: 0
6: 2
7: 0
8: 1

Top

EventsInformation And EventsCalculations

The methods are

EventsInformation:

• all_events

• all_events_include_zeroes

• biggest_event

• biggest_events_all <- returns the list of all events that have biggest occurrence

• events_keys

• events_range

• get_event

• get_items <- returns dict.items(): a list in py2 and an iterator in py3.

• get_range_of_events

• total_occurrences

EventsCalculations:

• full_table_string

• info

• mean

• percentage_axes

  • very fast but only good to 10 decimal places

• percentage_axes_exact

• percentage_points

  • very fast but only good to 10 decimal places

• percentage_points_exact

• stats_strings

  • takes a list of events values you want information for

  • returns

    • string of those events

    • number of times those events occurred in the table

    • total number of occurrences of all events in the table

    • the inverse chance of those events occurring: a 1 in (number) chance

    • the percent chance of those events occurring: (number)% chance

• stddev

  • defaults to 4 decimal places, but can be increased or decreased

In[34]: table = dt.DiceTable.new().add_die(dt.Die(6), 1000)

In[35]: calc = dt.EventsCalculations(table)

In[36]: calc.stddev(7)
Out[36]: 54.0061725

In[37]: calc.mean()
Out[37]: 3500.0

In[38]: calc.stats_strings([3500])
Out[38]: ('3,500', '1.046e+776', '1.417e+778', '135.4', '0.7386')
(yes, that is correct. out of 5000 possible rolls, 3500 has a 0.7% chance of occurring)

In[41]: calc.stats_strings(list(range(1000, 3001)) + list(range(4000, 10000)))

Out[41]:
('1,000-3,000, 4,000-9,999',
 '2.183e+758',
 '1.417e+778',
 '6.490e+19',
 '1.541e-18')

(this is also correct; rolls not in the middle 1000 collectively have a much smaller chance than the mean.)

EventsCalculations.include_zeroes is only settable at instantiation. It does exactly what it says. EventCalculations owns an EventsInformation. So instantiating EventsCalculations gets you two for the price of one. It’s accessed with the property EventsCalculations.info .

In[4]: table.add_die(dt.StrongDie(dt.Die(3), 2))

In[5]: calc = dt.EventsCalculations(table, True)

In[6]: print(calc.full_table_string())
2: 1
3: 0
4: 1
5: 0
6: 1

In[7]: calc = dt.EventsCalculations(table, False)

In[8]: print(calc.full_table_string())
2: 1
4: 1
6: 1

In [10]: calc.info.events_range()
Out[10]: (2, 6)

Top

Inheritance

If you inherit from any child of AdditiveEvents and you do not load the new information into EventsFactory, it will complain and give you instructions. The EventsFactory will try to create your new class and if it fails, will return the closest related type:

In[9]: class A(dt.DiceTable):
  ...:     pass
  ...:

In[10]: A.new()
E:\work\dice_tables\dicetables\baseevents.py:74: EventsFactoryWarning:
factory: <class 'dicetables.factory.eventsfactory.EventsFactory'>
Warning code: CONSTRUCT
Failed to find/add the following class to the EventsFactory -
class: <class '__main__.A'>
..... blah blah blah.....

Out[10]: <__main__.A at 0x4c25400>  <-- you got lucky. it's your class

In[11]: class B(dt.DiceTable):
  ...:     def __init__(self, name, number, events_dict, dice_data):
  ...:         self.name = name
  ...:         self.num = number
  ...:

In[12]: B.new()
E:\work\dice_tables\dicetables\baseevents.py:74: EventsFactoryWarning:
factory: <class 'dicetables.factory.eventsfactory.EventsFactory'>
Warning code: CONSTRUCT
Failed to find/add the following class to the EventsFactory -
class: <class '__main__.B'>
..... blah blah blah.....

Out[12]: <dicetables.dicetable.DiceTable at 0x4c23f28>  <-- Oops. EventsFactory can't figure out how to make one.

Now I will try again, but I will give the factory the info it needs.

The factory knows how to get ‘get_dict’, ‘dice_data’

and ‘calc_includes_zeroes’. If you need it to get anything else, you need tuples of

(<getter name>, <default value>, ‘property’ or ‘method’)

In[6]: class B(dt.DiceTable):
  ...:     factory_keys = ('name', 'get_num', 'get_dict', 'dice_data')
  ...:     new_keys = (('name', '', 'property'), ('get_num', 0, 'method'))
  ...:     def __init__(self, name, number, events_dict, dice_data):
  ...:         self.name = name
  ...:         self._num = number
  ...:         super(B, self).__init__(events_dict, dice_data)
  ...:     def get_num(self):
  ...:         return self._num
  ...:
In[7]: B.new()
Out[7]: <__main__.B at 0x4ca94a8>

In[8]: class C(dt.DiceTable):
  ...:     factory_keys = ('get_dict', 'dice_data')
  ...:     def fancy_add_die(self, die, times):
  ...:         new = self.add_die(die, times)
  ...:         return 'so fancy', new
  ...:
In[9]: x = C.new().fancy_add_die(dt.Die(3), 2)
In[10]: x[1].get_dict()
Out[10]: {2: 1, 3: 2, 4: 3, 5: 2, 6: 1}
In[11]: x
Out[11]: ('so fancy', <__main__.C at 0x5eb4d68>)  <-- notice it returned C and not DiceTable

The other way to do this is to directly add the class to the EventsFactory:

In[49]: factory = dt.factory.eventsfactory.EventsFactory

In[50]: factory.add_getter('get_num', 0, 'method')

In[51]: class A(dt.DiceTable):
   ...:     def __init__(self, number, events_dict, dice):
   ...:         self._num = number
   ...:         super(A, self).__init__(events_dict, dice)
   ...:     def get_num(self):
   ...:         return self._num
   ...:

In[53]: factory.add_class(A, ('get_num', 'get_dict', 'dice_data'))

In[55]: A.new()
Out[55]: <__main__.A at 0x5f951d0>

In[63]: factory.reset()

In[64]: factory.has_class(A)
Out[64]: False

When creating new methods, you can generate new events dictionaries by using dicetables.additiveevents.EventsDictCreator. the factory can create new instances with EventsFactory.from_params. For examples see the last few test in tests.factory.test_eventsfactory Top

HOW TO GET ERRORS AND BUGS

Every time you instantiate any IntegerEvents, it is checked. The get_dict() method returns a dict, and every value in get_dict().values() must be >=1. get_dict() may not be empty. since dt.Die(-2).get_dict() returns {}:

In [3]: dt.Die(-2)
dicetables.eventsbases.eventerrors.InvalidEventsError: events may not be empty. a good alternative is the identity - {0: 1}.

In [5]: dt.AdditiveEvents({1.0: 2})
dicetables.eventsbases.eventerrors.InvalidEventsError: all values must be ints

In [6]: dt.WeightedDie({1: 1, 2: -5})
dicetables.eventsbases.eventerrors.InvalidEventsError: no negative or zero occurrences in Events.get_dict()

Because AdditiveEvents and WeightedDie specifically scrub the zeroes from their get_dict() methods, these will not throw errors.

In [9]: dt.AdditiveEvents({1: 1, 2: 0}).get_dict()
Out[9]: {1: 1}

In [11]: weird = dt.WeightedDie({1: 1, 2: 0})

In [12]: weird.get_dict()
Out[12]: {1: 1}

In [13]: weird.get_size()
Out[13]: 2

In [14]: weird.get_raw_dict()
Out[14]: {1: 1, 2: 0}

Special rule for WeightedDie and ModWeightedDie:

In [15]: dt.WeightedDie({0: 1})
ValueError: rolls may not be less than 1. use ModWeightedDie

In [16]: dt.ModWeightedDie({0: 1}, 1)
ValueError: rolls may not be less than 1. use ModWeightedDie

Here’s how to add 0 one time (which does nothing, btw):

In [18]: dt.ModWeightedDie({1: 1}, -1).get_dict()
Out[18]: {0: 1}

StrongDie also has a weird case that can be unpredictable. Basically, don’t multiply by zero:

In [44]: table = dt.DiceTable.new().add_die(dt.Die(6))

In [45]: table = table.add_die(dt.StrongDie(dt.Die(100), 0), 100)

In [46]: table.get_dict()

Out[46]: {1: 1, 2: 1, 3: 1, 4: 1, 5: 1, 6: 1}

In [47]: print(table)
1D6
(100D100)X(0)

In [48]: stupid_die = dt.StrongDie(dt.ModWeightedDie({1: 2, 3: 4}, -1), 0)

In [49]: table = table.add_die(stupid_die, 2) <- this rolls zero with weight 4

In [50]: print(table)
(2D3-2  W:6)X(0)
1D6
(100D100)X(0)

In [51]: table.get_dict()
Out[51]: {1: 16, 2: 16, 3: 16, 4: 16, 5: 16, 6: 16} <- this is correct, it's just stupid.

“remove_die” and “add_die” are safe. They raise an error if you remove too many dice or add or remove a negative number.

If you “remove” or “combine” with a negative number, nothing should happen, but i make no guarantees.

If you use “remove” to remove what you haven’t added, it may or may not raise an error, but it’s guaranteed buggy:

In [19]: table = dt.DiceTable.new().add_die(dt.Die(6))

In [21]: table = table.remove_die(dt.Die(6), 4)
dicetables.eventsbases.eventerrors.DiceRecordError: Tried to create a DiceRecord with a negative value at Die(6): -3

In [22]: table = table.remove_die(dt.Die(10))
dicetables.eventsbases.eventerrors.DiceRecordError: Tried to create a DiceRecord with a negative value at Die(10): -1

In [26]: table = table.add_die(dt.Die(6), -3)
dicetables.eventsbases.eventerrors.DiceRecordError: Tried to add_die or remove_die with a negative number.

In [27]: table = table.remove_die(dt.Die(6), -3)
dicetables.eventsbases.eventerrors.DiceRecordError: Tried to add_die or remove_die with a negative number.

In [28]: table.get_dict()
Out[28]: {1: 1, 2: 1, 3: 1, 4: 1, 5: 1, 6: 1}

In [29]: table = table.combine(dt.Die(10000), -100)

In [30]: table.get_dict()
Out[30]: {1: 1, 2: 1, 3: 1, 4: 1, 5: 1, 6: 1}

In [31]: table = table.remove(dt.Die(2), 10)
ValueError: min() arg is an empty sequence <-didn't know this would happen, but at least failed loudly

In [32]: table = table.remove(dt.Die(2), 2)

In [33]: table.get_dict()
Out[33]: {-1: 1, 1: 1} <-bad. this is a random answer

(I know why you're about to get wacky and inaccurate errors, and I could fix the bug, except ...
 YOU SHOULD NEVER EVER DO THIS!!!!)
In [34]: table = table.remove(dt.AdditiveEvents({-5: 100}))
dicetables.eventsbases.eventerrors.InvalidEventsError: events may not be empty. a good alternative is the identity - {0: 1}.

During handling of the above exception, another exception occurred:

dicetables.factory.errorhandler.EventsFactoryError: Error Code: SIGNATURES DIFFERENT
Factory:    <class 'dicetables.factory.eventsfactory.EventsFactory'>
Error At:   <class 'dicetables.dicetable.DiceTable'>
Attempted to construct a class already present in factory, but with a different signature.
Class: <class 'dicetables.dicetable.DiceTable'>
Signature In Factory: ('get_dict', 'dice_data')
To reset the factory to its base state, use EventsFactory.reset()

Since you can instantiate a DiceTable with any legal input, you can make a table with utter nonsense. It will work horribly. for instance, the dictionary for 2D6 is:

{2: 1, 3: 2, 4: 3, 5: 4, 6: 5, 7: 6, 8: 5, 9: 4, 10: 3, 11: 2, 12: 1}

In[22]: nonsense = dt.DiceTable({1: 1}, dt.DiceRecord({dt.Die(6): 2})) <- BAD DATA!!!!

In[23]: print(nonsense)  <- the dice record says it has 2D6, but the events dictionary is WRONG
2D6

In[24]: nonsense = nonsense.remove_die(dt.Die(6), 2)  <- so here's your error. I hope you're happy.
ValueError: min() arg is an empty sequence

But, you cannot instantiate a DiceTable with negative values for dice. And you cannot instantiate a DiceTable with non-sense values for dice.

In[11]: dt.DiceTable({1: 1}, dt.DiceRecord({dt.Die(3): 3, dt.Die(5): -1}))
dicetables.eventsbases.eventerrors.DiceRecordError: Tried to create a DiceRecord with a negative value at Die(5): -1

In[12]: dt.DiceTable({1: 1}, dt.DiceRecord({'a': 2.0}))
dicetables.eventsbases.eventerrors.DiceRecordError: input must be {ProtoDie: int, ...}

Calling combine_by_flattened_list can be risky:

In [36]: x = dt.AdditiveEvents({1:1, 2: 5})

In [37]: x = x.combine_by_flattened_list(dt.AdditiveEvents({1: 2, 3: 4}), 5)

In [39]: x = x.combine_by_flattened_list(dt.AdditiveEvents({1: 2, 3: 4*10**10}), 5)
MemoryError

In [42]: x = x.combine_by_flattened_list(dt.AdditiveEvents({1: 2, 3: 4*10**700}))
OverflowError: cannot fit 'int' into an index-sized integer

Top

CHANGES

0.4_to_1.0

1.0_to_2.0

from version 0.4.6 to version 1.0

There are several major changes. An important side-effect of these changes is that dicetables is now much more modular and ready for change. It is now possible to speed up the algorithms and push those changes without further affecting the API. (speeds have already been doubled for large adds).

Version 1.0 is just an intermediary to allow some of the useful changes without breaking the API too badly. If any really astounding changes happen, I will try to adapt them to the 1.X versions too, but any further work will be done on 2.X versions.

• Modules and classes and methods got renamed. see the dictionary at the bottom. There are new classes

• DiceTable.__init__() now takes arguments. The class method DiceTable.new() creates an empty table.

• DiceTable and its parent AdditiveEvents are no longer responsible for obtaining any but the most basic information.

• All the calculations and information are now done by EventsInformation and EventsCalculations

• Aside from the above two classes, every other object is now a child of IntegerEvents.

• Dice classes no longer have “tuple_list()” method. They use the same “get_dict()” method that all IntegerEvents use

The following modules and classes have been renamed.

• longintmath.py: baseevents.py

• dicestats.py: dieevents.py, dicetable.py

• tableinfo.py: eventsinfo.py

• LongIntTable: AdditiveEvents

The following classes have been added:

• baseevents.InvalidEventsError

• dicetable.DiceRecordError

• baseevents.IntegerEvents

• dicetable.DetailedDiceTable

• eventsinfo.EventsInformation

• eventsinfo.EventsCalculations

DiceTable.__init__() now takes two arguments - a dictionary of {event: occurrences} and a list of [(die, number), ]. to create a new table, call the class method DiceTable.new(). This change allows easy creation of a new dice table from data. new_table = DiceTable(old_table.get_dict(), old_table.get_list()) or new_table = DiceTable(stored_dict, stored_dice_list). To create a DiceTable with no dice, use DiceTable.new().

The base class of DiceTable is now called AdditiveEvents and not LongIntTable. If any IntegerEvents events is instantiated in a way that would cause bugs, it raises an error; the same is true for any dice.

AdditiveEvents.combine/remove take any IntegerEvents as an argument whereas LongIntTable.add/remove took a list of tuples as an argument. the methods for getting basic information from LongIntTable are now in EventsInformation. mean() and stddev() are part of EventsCalculations object. These objects work on ANY kind of IntegerEvents, not just DiceTable.

all of tableinfo was rewritten as objects. although they are deprecated, the following still exist as wrapper functions for those objects:

• events_range

• format_number

• full_table_string

• graph_pts

• graph_pts_overflow

• mean

• percentage_axes

• percentage_points

• safe_true_div

• stats

• stddev

the new objects are:

• NumberFormatter

• EventsInformation

• EventsCalculations

for details, see their headings in the README.

For output: stats() now shows tiny percentages, and if infinite, shows ‘Infinity’. Any exponent between 10 and -10 has that extraneous zero removed: ‘1.2e+05’ is now ‘1.2e+5’.

Any subclass of ProtoDie no longer has the .tuple_list() method. It has been replaced by the .get_dict() method which returns a dictionary and not a list of tuples. The string for StrongDie now puts parentheses around the multiplier.

CONVERSIONS = {
    'DiceTable()': 'DiceTable.new()',
    'LongIntTable.add': 'AdditiveEvents.combine',
    'LongIntTable.frequency': 'EventsInformation(event).get_event',
    'LongIntTable.frequency_all': 'EventsInformation(event).all_events',
    'LongIntTable.frequency_highest': 'EventsInformation(event).biggest_event',
    'LongIntTable.frequency_range': 'EventsInformation(event).get_range_of_events',
    'LongIntTable.mean': 'EventsCalculations(event).mean',
    'LongIntTable.merge': 'GONE',
    'LongIntTable.remove': 'AdditiveEvents.remove',
    'LongIntTable.stddev': 'EventsCalculations(event).stddev',
    'LongIntTable.total_frequency': 'EventsInformation(event).total_occurrences',
    'LongIntTable.update_frequency': 'GONE',
    'LongIntTable.update_value_add': 'GONE',
    'LongIntTable.update_value_ow': 'GONE',
    'LongIntTable.values': 'EventsInformation(event).event_keys',
    'LongIntTable.values_max': 'EventsInformation(event).event_range[0]',
    'LongIntTable.values_min': 'EventsInformation(event).event_range[1]',
    'LongIntTable.values_range': 'EventsInformation(event).event_range',
    'DiceTable.update_list': 'GONE (DiceTable owns a DiceRecord object that handles this)',
    'ProtoDie.tuple_list': ('sorted(ProtoDie.get_dict().items)', 'EventsInformation(ProtoDie).all_events'),
    'scinote': 'NumberFormatter.format',
    'full_table_string', 'EventsCalculations(event).full_table_string',
    'stats', 'EventsCalculations(event).stats_strings',
    'long_int_div': 'safe_true_div',
    'graph_pts': ('EventsCalculations(event).percentage_points',
                  'EventsCalculations(event).percentage_points_exact',
                  'EventsCalculations(event).percentage_axes',
                  'EventsCalculations(event).percentage_axes_exact',
                  'EventsInformation(events).all_events',
                  'EventsInformation(events).all_events_include_zeroes')
    }

Top

from version 1.0 to version 2.0

There are 3 large changes. They affect speed slightly for the worse in the microseconds range (small numbers of adds) and double the speed in the large adds range.

• add_die, remove_die, all combines and remove now follow the signature add_die(die, times=1). It takes care of inconsistencies in the API: all output was (events, times) and now input is the same. It allows for easily adding/removing one time without confusion. Since the other changes were so drastic, one more couldn’t hurt.

• all children of AdditiveEvents are immutable. This can have some interesting inheritance effects. See Inheritance.

• DiceTable does not take a list of [(die, number), …]. It now takes DiceRecord({die: number}). To get the correct data to build a new table, use DiceTable.get_dict() and DiceTable.dice_data() .

• Removed wrapper functions: graph_pts, graph_pts_overflow, format_number

in [12]: new = dt.AdditiveEvents.new()

in [12]: new.combine(dt.AdditiveEvents({1: 1, 2: 5}), 2)
Out[13]: <dicetables.baseevents.AdditiveEvents at 0x5e73828>

In [14]: dt.DiceTable({1: 1, 2: 1, 3: 1}, dt.DiceRecord({dt.Die(3): 1}))
Out[14]: <dicetables.dicetable.DiceTable at 0x5eddef0>