Copyright 2016-2022 Caleb Evans
Released under the MIT license

Automata is a Python 3 library which implements the structures and algorithms for finite automata, pushdown automata, and Turing machines. The library requires Python 3.7 or newer.

Huge thanks to @YtvwlD, @dengl11, @Tagl, @lewiuberg, @abhinavsinha‑adrino, and @eliotwrobson for their invaluable code contributions to this project! 🎉

If you wish to migrate to Automata v6 from an older version, please follow the migration guide.

You can install the latest version of Automata via pip:

pip install automata-lib

• class Automaton
  • Finite Automaton (FA)
    • Deterministic (DFA)
    • Non-Deterministic (NFA)
    • Generalized Non-Deterministic (GNFA)
  • Pushdown Automaton (PDA)
    • Deterministic (DPDA)
    • Non-Deterministic (NPDA)
  • Turing Machine (TM)
    • Deterministic (DTM)
    • Non-Deterministic (NTM)
    • Multi-Tape Non-Deterministic (MNTM)
• Regular Expressions
• Base exception classes
• Turing machine exceptions

The Automaton class is an abstract base class from which all automata (including Turing machines) inherit. As such, it cannot be instantiated on its own; you must use a defined subclass instead (or you may create your own subclass if you're feeling adventurous). The Automaton class can be found under automata/base/automaton.py.

If you wish to subclass Automaton, you can import it like so:

from automata.base.automaton import Automaton

The following methods are common to all Automaton subtypes, and must be implemented if you create your own subclass:

Reads an input string into the automaton, returning the automaton's final configuration (according to its subtype). If the input is rejected, the method raises a RejectionException.

Reads an input string like read_input(), except instead of returning the final configuration, the method returns a generator. The values yielded by this generator depend on the automaton's subtype.

If the string is rejected by the automaton, the method still raises a RejectionException.

Reads an input string like read_input(), except it returns a boolean instead of returning the automaton's final configuration (or raising an exception). That is, the method always returns True if the input is accepted, and it always returns False if the input is rejected.

Checks whether the automaton is actually a valid automaton (according to its subtype). It returns True if the automaton is valid; otherwise, it will raise the appropriate exception (e.g. the state transition is missing for a particular symbol).

This method is automatically called when the automaton is initialized, so it's only really useful if a automaton object is modified after instantiation.

Returns a deep copy of the automaton according to its subtype.

The FA class is an abstract base class from which all finite automata inherit. The FA class can be found under automata/fa/fa.py.

If you wish to subclass FA, you can import it like so:

from automata.fa.fa import FA

The DFA class is a subclass of FA and represents a deterministic finite automaton. It can be found under automata/fa/dfa.py.

Every DFA has the following (required) properties:

1. states: a set of the DFA's valid states, each of which must be represented as a string

2. input_symbols: a set of the DFA's valid input symbols, each of which must also be represented as a string

3. transitions: a dict consisting of the transitions for each state. Each key is a state name and each value is a dict which maps a symbol (the key) to a state (the value).

4. initial_state: the name of the initial state for this DFA

5. final_states: a set of final states for this DFA

6. allow_partial: by default, each DFA state must have a transition to every input symbol; if allow_partial is True, you can disable this characteristic (such that any DFA state can have fewer transitions than input symbols)

from automata.fa.dfa import DFA
# DFA which matches all binary strings ending in an odd number of '1's
dfa = DFA(
    states={'q0', 'q1', 'q2'},
    input_symbols={'0', '1'},
    transitions={
        'q0': {'0': 'q0', '1': 'q1'},
        'q1': {'0': 'q0', '1': 'q2'},
        'q2': {'0': 'q2', '1': 'q1'}
    },
    initial_state='q0',
    final_states={'q1'}
)

Returns the final state the DFA stopped on, if the input is accepted.

dfa.read_input('01')  # returns 'q1'

dfa.read_input('011')  # raises RejectionException

Yields each state reached as the DFA reads characters from the input string, if the input is accepted.

dfa.read_input_stepwise('0111')
# yields:
# 'q0'
# 'q0'
# 'q1'
# 'q2'
# 'q1'

if dfa.accepts_input(my_input_str):
    print('accepted')
else:
    print('rejected')

Returns True if the DFA is valid, otherwise the appropriate exception is raised.

dfa.validate()  # returns True or raises an exception

dfa.copy()  # returns deep copy of dfa

Creates a minimal DFA which accepts the same inputs as the old one. Unreachable states are removed and equivalent states are merged. States are renamed by default.

minimal_dfa = dfa.minify()
minimal_dfa_with_old_names = dfa.minify(retain_names=True)

Creates a DFA which accepts an input if and only if the old one does not.

complement_dfa = ~dfa

Given two DFAs which accept the languages A and B respectively, creates a DFA which accepts the union of A and B. Minifies by default.

minimal_union_dfa = dfa | other_dfa
union_dfa = dfa.union(other_dfa, minify=False)

Given two DFAs which accept the languages A and B respectively, creates a DFA which accepts the intersection of A and B. Minifies by default.

minimal_intersection_dfa = dfa & other_dfa
intersection_dfa = dfa.intersection(other_dfa, minify=False)

Given two DFAs which accept the languages A and B respectively, creates a DFA which accepts the set difference of A and B, often denoted A \ B or A - B. Minifies by default.

minimal_difference_dfa = dfa - other_dfa
difference_dfa = dfa.difference(other_dfa, minify=False)

Given two DFAs which accept the languages A and B respectively, creates a DFA which accepts the symmetric difference of A and B. Minifies by default.

minimal_symmetric_difference_dfa = dfa ^ other_dfa
symmetric_difference_dfa = dfa.symmetric_difference(other_dfa, minify=False)

Given two DFAs which accept the languages A and B respectively, returns True of the A is a subset of B, False otherwise.

dfa <= other_dfa
dfa.issubset(other_dfa)

Given two DFAs which accept the languages A and B respectively, returns True of the A is a superset of B, False otherwise.

dfa >= other_dfa
dfa.issuperset(other_dfa)

Given two DFAs which accept the languages A and B respectively, returns True of A and B are disjoint, False otherwise.

dfa.isdisjoint(other_dfa)

Returns True if the DFA does not accept any inputs, False otherwise.

dfa.isempty()

Returns True if the DFA accepts a finite language, False otherwise.

dfa.isfinite()

Creates a DFA that is equivalent to the given NFA.

from automata.fa.dfa import DFA
from automata.fa.nfa import NFA
dfa = DFA.from_nfa(nfa)  # returns an equivalent DFA

dfa.show_diagram(path='./dfa1.png')

The NFA class is a subclass of FA and represents a nondeterministic finite automaton. It can be found under automata/fa/nfa.py.

Every NFA has the same five DFA properties: state, input_symbols, transitions, initial_state, and final_states. However, the structure of the transitions object has been modified slightly to accommodate the fact that a single state can have more than one transition for the same symbol. Therefore, instead of mapping a symbol to one end state in each sub-dict, each symbol is mapped to a set of end states.

from automata.fa.nfa import NFA
# NFA which matches strings beginning with 'a', ending with 'a', and containing
# no consecutive 'b's
nfa = NFA(
    states={'q0', 'q1', 'q2'},
    input_symbols={'a', 'b'},
    transitions={
        'q0': {'a': {'q1'}},
        # Use '' as the key name for empty string (lambda/epsilon) transitions
        'q1': {'a': {'q1'}, '': {'q2'}},
        'q2': {'b': {'q0'}}
    },
    initial_state='q0',
    final_states={'q1'}
)

Returns a set of final states the FA stopped on, if the input is accepted.

nfa.read_input('aba')  # returns {'q1', 'q2'}

nfa.read_input('abba')  # raises RejectionException

Yields each set of states reached as the NFA reads characters from the input string, if the input is accepted.

nfa.read_input_stepwise('aba')
# yields:
# {'q0'}
# {'q1', 'q2'}
# {'q0'}
# {'q1', 'q2'}

if nfa.accepts_input(my_input_str):
    print('accepted')
else:
    print('rejected')

Returns True if the NFA is valid, otherwise the appropriate exception is raised.

nfa.validate()  # returns True or raises an exception

nfa.copy()  # returns deep copy of nfa

nfa.reverse()

reversed(nfa)

nfa1 + nfa2

nfa1.concatenate(nfa2)

nfa1.kleene_star()

Returns union of two NFAs

new_nfa = nfa1.union(nfa2)

new_nfa = nfa1 | nfa2

Removes epsilon transitions from the NFA which recognizes the same language.

nfa1.eliminate_lambda()

Creates an NFA that is equivalent to the given DFA.

from automata.fa.nfa import NFA
from automata.fa.dfa import DFA
nfa = NFA.from_dfa(dfa)  # returns an equivalent NFA

Returns a new NFA instance from the given regular expression.

nfa1 = NFA.from_regex('ab(c|d)*ba?')

Return a regular expression (string) equivalent to given NFA.

regex = nfa1.to_regex()

nfa1.show_diagram(path='./abc.png')

The GNFA class is a subclass of NFA and represents a generalized nondeterministic finite automaton. It can be found under automata/fa/gnfa.py. Its main usage is for conversion of DFAs and NFAs to regular expressions.

Every GNFA has the same five NFA properties: states, input_sympols, transitions, initial_state, and final_state. However a GNFA has several differences with respect to NFA

• The initial_state has transitions going to every other state but no transitions coming in from any other state.
• There is only a single final_state, and it has transitions coming in from every other state but no transitions going to any other state. Futhermore, the final_state is not the same has initial_state.
• Except for initial_state and final_state, one transition arrow goes from every state to every other state and also from each state to itself. To accommodate this, transitions can be regular expressions and None also in addition to normal symbols.

GNFA is modified with respect to NFA in the following parameters:

1. final_state: a string (single state).
2. transitions: (its structure is changed from NFA) a dict consisting of the transitions for each state except final_state. Each key is a state name and each value is dict which maps a state (the key) to the transition expression (the value).
   • value: a regular expression (string) consisting of input_symbols and the following symbols only: *, |, ?, (). Check Regular Expressions

from automata.fa.gnfa import GNFA
# GNFA which matches strings beginning with 'a', ending with 'a', and containing
# no consecutive 'b's
gnfa = GNFA(
    states={'q_in', 'q_f', 'q0', 'q1', 'q2'},
    input_symbols={'a', 'b'},
    transitions={
        'q0': {'q1': 'a', 'q_f': None, 'q2': None, 'q0': None},
        'q1': {'q1': 'a', 'q2': '', 'q_f': '', 'q0': None},
        'q2': {'q0': 'b', 'q_f': None, 'q2': None, 'q1': None},
        'q_in': {'q0': '', 'q_f': None, 'q2': None, 'q1': None}
    },
    initial_state='q_in',
    final_state='q_f'
)

Initialize this GNFA as one equivalent to the given DFA.

from automata.fa.gnfa import GNFA
from automata.fa.dfa import DFA
gnfa = GNFA.from_dfa(dfa) # returns an equivalent GNFA

Initialize this GNFA as one equivalent to the given NFA.

from automata.fa.gnfa import GNFA
from automata.fa.nfa import NFA
gnfa = GNFA.from_nfa(nfa) # returns an equivalent GNFA

Returns True if the GNFA is valid, otherwise the appropriate exception is raised.

gnfa.validate()

Returns a deep copy of GNFA.

gnfa2 = gnfa1.copy()

Convert GNFA to regular expression.

gnfa.to_regex() # returns a regular expression (string)

Writes a visual diagram of the GNFA to an image file.

1. path: the path of the image to be saved
2. show_None: A boolean indicating when to show or hide None transitions; defaults to True.

gnfa.show_diagram(path='./gnfa.png', show_None=False)

The PDA class is an abstract base class from which all pushdown automata inherit. It can be found under automata/pda/pda.py.

The DPDA class is a subclass of PDA and represents a deterministic finite automaton. It can be found under automata/pda/dpda.py.

Every DPDA has the following (required) properties:

1. states: a set of the DPDA's valid states, each of which must be represented as a string

2. input_symbols: a set of the DPDA's valid input symbols, each of which must also be represented as a string

3. stack_symbols: a set of the DPDA's valid stack symbols

4. transitions: a dict consisting of the transitions for each state; see the example below for the exact syntax

5. initial_state: the name of the initial state for this DPDA

6. initial_stack_symbol: the name of the initial symbol on the stack for this DPDA

7. final_states: a set of final states for this DPDA

8. acceptance_mode: a string defining whether this DPDA accepts by 'final_state', 'empty_stack', or 'both'; the default is 'both'

from automata.pda.dpda import DPDA
# DPDA which which matches zero or more 'a's, followed by the same
# number of 'b's (accepting by final state)
dpda = DPDA(
    states={'q0', 'q1', 'q2', 'q3'},
    input_symbols={'a', 'b'},
    stack_symbols={'0', '1'},
    transitions={
        'q0': {
            'a': {'0': ('q1', ('1', '0'))}  # transition pushes '1' to stack
        },
        'q1': {
            'a': {'1': ('q1', ('1', '1'))},
            'b': {'1': ('q2', '')}  # transition pops from stack
        },
        'q2': {
            'b': {'1': ('q2', '')},
            '': {'0': ('q3', ('0',))}  # transition does not change stack
        }
    },
    initial_state='q0',
    initial_stack_symbol='0',
    final_states={'q3'},
    acceptance_mode='final_state'
)

Returns a PDAConfiguration object representing the DPDA's config. This is basically a tuple containing the final state the DPDA stopped on, the remaining input (an empty string) as well as a PDAStack object representing the DPDA's stack (if the input is accepted).

dpda.read_input('ab')  # returns PDAConfiguration('q3', '', PDAStack(('0',)))

dpda.read_input('aab')  # raises RejectionException

Yields sets of PDAConfiguration objects. These are basically tuples containing the current state, the remaining input and the current stack as a PDAStack object, if the input is accepted.

dpda.read_input_stepwise('ab')
# yields:
# PDAConfiguration('q0', 'ab', PDAStack(('0')))
# PDAConfiguration('q1', 'a', PDAStack(('0', '1')))
# PDAConfiguration('q3', '', PDAStack(('0')))

if dpda.accepts_input(my_input_str):
    print('accepted')
else:
    print('rejected')

dpda.validate()  # returns True

dpda.copy()  # returns deep copy of dpda

The NPDA class is a subclass of PDA and represents a nondeterministic pushdown automaton. It can be found under automata/pda/npda.py.

Every NPDA has the following (required) properties:

1. states: a set of the NPDA's valid states, each of which must be represented as a string

2. input_symbols: a set of the NPDA's valid input symbols, each of which must also be represented as a string

3. stack_symbols: a set of the NPDA's valid stack symbols

4. transitions: a dict consisting of the transitions for each state; see the example below for the exact syntax

5. initial_state: the name of the initial state for this NPDA

6. initial_stack_symbol: the name of the initial symbol on the stack for this NPDA

7. final_states: a set of final states for this NPDA

8. acceptance_mode: a string defining whether this NPDA accepts by 'final_state', 'empty_stack', or 'both'; the default is 'both'

from automata.pda.npda import NPDA
# NPDA which matches palindromes consisting of 'a's and 'b's
# (accepting by final state)
# q0 reads the first half of the word, q1 the other half, q2 accepts.
# But we have to guess when to switch.
npda = NPDA(
    states={'q0', 'q1', 'q2'},
    input_symbols={'a', 'b'},
    stack_symbols={'A', 'B', '#'},
    transitions={
        'q0': {
            '': {
                '#': {('q2', '#')},
            },
            'a': {
                '#': {('q0', ('A', '#'))},
                'A': {
                    ('q0', ('A', 'A')),
                    ('q1', ''),
                },
                'B': {('q0', ('A', 'B'))},
            },
            'b': {
                '#': {('q0', ('B', '#'))},
                'A': {('q0', ('B', 'A'))},
                'B': {
                    ('q0', ('B', 'B')),
                    ('q1', ''),
                },
            },
        },
        'q1': {
            '': {'#': {('q2', '#')}},
            'a': {'A': {('q1', '')}},
            'b': {'B': {('q1', '')}},
        },
    },
    initial_state='q0',
    initial_stack_symbol='#',
    final_states={'q2'},
    acceptance_mode='final_state'
)

Returns a set of PDAConfigurations representing all of the NPDA's configurations. Each of these is basically a tuple containing the final state the NPDA stopped on, the remaining input (an empty string) as well as a PDAStack object representing the NPDA's stack (if the input is accepted).

npda.read_input("aaaa") # returns {PDAConfiguration('q2', '', PDAStack('#',))}

npda.read_input('ab')  # raises RejectionException

Yields sets of PDAConfiguration object. Each of these is basically a tuple containing the current state, the remaining input and the current stack as a PDAStack object, if the input is accepted.

npda.read_input_stepwise('aa')
# yields:
# {PDAConfiguration('q0', 'aa', PDAStack('#',))}
# {PDAConfiguration('q0', 'a', PDAStack('#', 'A')), PDAConfiguration('q2', 'aa', PDAStack('#',))}
# {PDAConfiguration('q0', '', PDAStack('#', 'A', 'A')), PDAConfiguration('q1', '', PDAStack('#',))}
# {PDAConfiguration('q2', '', PDAStack('#',))}

if npda.accepts_input(my_input_str):
    print('accepted')
else:
    print('rejected')

npda.validate()  # returns True

npda.copy()  # returns deep copy of npda

The TM class is an abstract base class from which all Turing machines inherit. It can be found under automata/tm/tm.py.

The DTM class is a subclass of TM and represents a deterministic Turing machine. It can be found under automata/tm/dtm.py.

Every DTM has the following (required) properties:

1. states: a set of the DTM's valid states, each of which must be represented as a string

2. input_symbols: a set of the DTM's valid input symbols represented as strings

3. tape_symbols: a set of the DTM's valid tape symbols represented as strings

4. transitions: a dict consisting of the transitions for each state; each key is a state name and each value is a dict which maps a symbol (the key) to a state (the value)

5. initial_state: the name of the initial state for this DTM

6. blank_symbol: a symbol from tape_symbols to be used as the blank symbol for this DTM

7. final_states: a set of final states for this DTM

from automata.tm.dtm import DTM
# DTM which matches all strings beginning with '0's, and followed by
# the same number of '1's
dtm = DTM(
    states={'q0', 'q1', 'q2', 'q3', 'q4'},
    input_symbols={'0', '1'},
    tape_symbols={'0', '1', 'x', 'y', '.'},
    transitions={
        'q0': {
            '0': ('q1', 'x', 'R'),
            'y': ('q3', 'y', 'R')
        },
        'q1': {
            '0': ('q1', '0', 'R'),
            '1': ('q2', 'y', 'L'),
            'y': ('q1', 'y', 'R')
        },
        'q2': {
            '0': ('q2', '0', 'L'),
            'x': ('q0', 'x', 'R'),
            'y': ('q2', 'y', 'L')
        },
        'q3': {
            'y': ('q3', 'y', 'R'),
            '.': ('q4', '.', 'R')
        }
    },
    initial_state='q0',
    blank_symbol='.',
    final_states={'q4'}
)

The direction N (for no movement) is also supported.

Returns a TMConfiguration. This is basically a tuple containing the final state the machine stopped on, as well as a TMTape object representing the DTM's internal tape (if the input is accepted).

dtm.read_input('01')  # returns TMConfiguration('q4', TMTape('xy..', 3))

Calling config.print() will produce a more readable output:

dtm.read_input('01').print()
# q4: xy..
#        ^

dtm.read_input('011')  # raises RejectionException

Yields sets of TMConfiguration objects. Those are basically tuples containing the current state and the current tape as a TMTape object.

dtm.read_input_stepwise('01')
# yields:
# TMConfiguration('q0', TMTape('01', 0))
# TMConfiguration('q1', TMTape('x1', 1))
# TMConfiguration('q2', TMTape('xy', 0))
# TMConfiguration('q0', TMTape('xy', 1))
# TMConfiguration('q3', TMTape('xy.', 2))
# TMConfiguration('q4', TMTape('xy..', 3))

if dtm.accepts_input(my_input_str):
    print('accepted')
else:
    print('rejected')

dtm.validate()  # returns True

dtm.copy()  # returns deep copy of dtm

The NTM class is a subclass of TM and represents a nondeterministic Turing machine. It can be found under automata/tm/ntm.py.

Every NTM has the following (required) properties:

1. states: a set of the NTM's valid states, each of which must be represented as a string

2. input_symbols: a set of the NTM's valid input symbols represented as strings

3. tape_symbols: a set of the NTM's valid tape symbols represented as strings

4. transitions: a dict consisting of the transitions for each state; each key is a state name and each value is a dict which maps a symbol (the key) to a set of states (the values)

5. initial_state: the name of the initial state for this NTM

6. blank_symbol: a symbol from tape_symbols to be used as the blank symbol for this NTM

7. final_states: a set of final states for this NTM

from automata.tm.ntm import NTM
# NTM which matches all strings beginning with '0's, and followed by
# the same number of '1's
# Note that the nondeterminism is not really used here.
ntm = NTM(
    states={'q0', 'q1', 'q2', 'q3', 'q4'},
    input_symbols={'0', '1'},
    tape_symbols={'0', '1', 'x', 'y', '.'},
    transitions={
        'q0': {
            '0': {('q1', 'x', 'R')},
            'y': {('q3', 'y', 'R')},
        },
        'q1': {
            '0': {('q1', '0', 'R')},
            '1': {('q2', 'y', 'L')},
            'y': {('q1', 'y', 'R')},
        },
        'q2': {
            '0': {('q2', '0', 'L')},
            'x': {('q0', 'x', 'R')},
            'y': {('q2', 'y', 'L')},
        },
        'q3': {
            'y': {('q3', 'y', 'R')},
            '.': {('q4', '.', 'R')},
        }
    },
    initial_state='q0',
    blank_symbol='.',
    final_states={'q4'}
)

The direction 'N' (for no movement) is also supported.

Returns a set of TMConfigurations. These are basically tuples containing the final state the machine stopped on, as well as a TMTape object representing the NTM's internal tape (if the input is accepted).

ntm.read_input('01')  # returns {TMConfiguration('q4', TMTape('xy..', 3))}

Calling config.print() will produce a more readable output:

ntm.read_input('01').pop().print()
# q4: xy..
#        ^

ntm.read_input('011')  # raises RejectionException

Yields sets of TMConfiguration objects. Those are basically tuples containing the current state and the current tape as a TMTape object.

ntm.read_input_stepwise('01')
# yields:
# {TMConfiguration('q0', TMTape('01', 0))}
# {TMConfiguration('q1', TMTape('x1', 1))}
# {TMConfiguration('q2', TMTape('xy', 0))}
# {TMConfiguration('q0', TMTape('xy', 1))}
# {TMConfiguration('q3', TMTape('xy.', 2))}
# {TMConfiguration('q4', TMTape('xy..', 3))}

if ntm.accepts_input(my_input_str):
    print('accepted')
else:
    print('rejected')

ntm.validate()  # returns True

ntm.copy()  # returns deep copy of ntm

The MNTM class is a subclass of TM and represents a multitape (non)deterministic Turing machine. It can be found under automata/tm/mntm.py.

Every MNTM has the following (required) properties:

1. states: a set of the MNTM's valid states, each of which must be represented as a string

2. input_symbols: a set of the MNTM's valid input symbols represented as strings

3. tape_symbols: a set of the MNTM's valid tape symbols represented as strings

4. n_tapes: an int which dictates the number of tapes of this MNTM

5. transitions: a dict consisting of the transitions for each state; each key is a state name and each value is a dict which maps a symbol (the key) to a set of states (the values)

6. initial_state: the name of the initial state for this MNTM

7. blank_symbol: a symbol from tape_symbols to be used as the blank symbol for this MNTM

8. final_states: a set of final states for this MNTM

from automata.tm.mntm import MNTM
# MNTM which accepts all strings in {0, 1}* and writes all
# 1's from the first tape (input) to the second tape.
self.mntm1 = MNTM(
    states={'q0', 'q1'},
    input_symbols={'0', '1'},
    tape_symbols={'0', '1', '#'},
    n_tapes=2,
    transitions={
        'q0': {
            ('1', '#'): [('q0', (('1', 'R'), ('1', 'R')))],
            ('0', '#'): [('q0', (('0', 'R'), ('#', 'N')))],
            ('#', '#'): [('q1', (('#', 'N'), ('#', 'N')))],
        }
    },
    initial_state='q0',
    blank_symbol='#',
    final_states={'q1'},
)

The direction 'N' (for no movement) is also supported.

Returns a set of MTMConfigurations. These are basically tuples containing the final state the machine stopped on, as well as a list of TMTape objects representing the MNTM's internal tape (if the input is accepted).

mntm.read_input('01')  # returns {MTMConfiguration('q1', [TMTape('01#', 2), TMTape('1#', 1)])}

Calling config.print() will produce a more readable output:

ntm.read_input('01').pop().print()
# q1:
#> Tape 1: 01#
#            ^
#> Tape 2: 1#
#           ^

ntm.read_input('2')  # raises RejectionException

Yields sets of MTMConfiguration objects. Those are basically tuples containing the current state and the list of TMTape objects.

ntm.read_input_stepwise('0111')
# yields:
# {MTMConfiguration('q0', (TMTape('0111', 0), TMTape('#', 0)))}
# {MTMConfiguration('q0', (TMTape('0111', 1), TMTape('#', 0)))}
# {MTMConfiguration('q0', (TMTape('0111', 2), TMTape('1#', 1)))}
# {MTMConfiguration('q0', (TMTape('0111', 3), TMTape('11#', 2)))}
# {MTMConfiguration('q0', (TMTape('0111#', 4), TMTape('111#', 3)))}
# {MTMConfiguration('q1', (TMTape('0111#', 4), TMTape('111#', 3)))}

Simulates the MNTM as an NTM by using an extended tape consisting of all tapes of the MNTM separated by a tape_separator_symbol = _ and 'virtual heads' for each 'virtual tape' (which are basically the portions of the extended tape separated by '_'). Each 'virtual head' corresponds to a special symbol (original_symbol + '^') on the extended tape which denotes where the actual head of that tape would be if the MNTM was being run as a MNTM and not a NTM. This is the classic algorithm for performing a single-tape simulation of a multi-tape Turing machine. For more information, visit Sipser's Introduction to the Theory of Computation 3rd Edition, Section 3.2.

ntm.read_input_as_ntm('0111')
# yields:
# {TMConfiguration('q0', TMTape('0^111_#^_', 0))}
# {TMConfiguration('q0', TMTape('01^11_#^_', 8))}
# {TMConfiguration('q0', TMTape('011^1_1#^_', 9))}
# {TMConfiguration('q0', TMTape('0111^_11#^_', 10))}
# {TMConfiguration('q0', TMTape('0111#^_111#^_', 12))}
# {TMConfiguration('q1', TMTape('0111#^_111#^_', 12))}

if mntm.accepts_input(my_input_str):
    print('accepted')
else:
    print('rejected')

ntm.validate()  # returns True

ntm.copy()  # returns deep copy of ntm

A set of tools for working with regular languages. These can be found under automata/base/regex.py

A regular expression with the following operations only are supported in this library:

• *: Kleene star operation. language repeated zero or more times. Ex: a*,(ab)*
• ?: Language repeated zero or one time. Ex: a?
• Concatenation: Ex: abcd
• |: Union. Ex: a|b
• (): Grouping

This is similar to the python RE module but this library does not support any other special character than given above. All regular languages can be written with these.

Preferably the tools for the same can be imported as:

import automata.base.regex as re

Returns True if the regular expression is valid. Otherwise, raise an InvalidRegexError.

re.validate('ab(c|d)*ba?')

Returns True if both regular expressions are equivalent.

re.isequal('aa?', 'a|aa')

Returns True if re1 is a subset of re2.

re.issubset('aa?', 'a*')

Returns True if re1 is a subset of re2.

re.issuperset('a*', 'a?')

The library also includes a number of exception classes to ensure that errors never pass silently (unless explicitly silenced). See automata/base/exceptions.py for these class definitions.

To reference these exceptions (so as to catch them in a try..except block or whatnot), simply import automata.base.exceptions however you'd like:

import automata.base.exceptions as exceptions

A base class from which all other automata exceptions inherit (including finite automata and Turing machines).

Raised if a specified state does not exist within the automaton's states set.

Raised if a specified symbol does not exist within the automaton's symbols set.

Raised if a specified transition definition is missing a defined start state. This error can also be raised if the initial state does not have any transitions defined.

Raised if a given symbol is missing where it would otherwise be required for this type of automaton (e.g. the automaton is missing a transition for one of the listed symbols).

Raised if the initial state fails to meet some required condition for this type of automaton (e.g. if the initial state is also a final state, which is prohibited for Turing machines).

Raised if a final state fails to meet some required condition for this type of automaton (e.g. the final state has transitions to other states, which is prohibited for Turing machines).

Raised if the automaton did not accept the input string after validating (e.g. the automaton stopped on a non-final state after validating input).

A base class for all regular expression related errors.

Raised if the input regular expression is invalid.

Raised if input symbols don't match between two automata but are expected to (e.g. in the creation of a product automaton).

The automata.tm package also includes a module for exceptions specific to Turing machines. You can reference these exception classes like so:

import automata.tm.exceptions as tm_exceptions

A base class from which all other Turing machine exceptions inherit.

Raised if a direction specified in this machine's transition map is not a valid direction (valid directions include 'L', 'R', and 'N').

Raised if the number of tapes defined for the mntm is not consistent with the transitions.

Raised if the extended tape for simulating a mntm as a ntm is not valid.