Work on matcher2

This commit is contained in:
Matthew Honnibal 2018-02-12 10:17:43 +01:00
parent ab35ac4e6f
commit fae5c0dc18

399
spacy/matcher2.pyx Normal file
View File

@ -0,0 +1,399 @@
# cython: infer_types=True
from libcpp.vector cimport vector
from libc.stdint cimport int32_t, uint64_t
from preshed.maps cimport PreshMap
from cymem.cymem cimport Pool
from murmurhash.mrmr cimport hash64
from .typedefs cimport attr_t, hash_t
from .structs cimport TokenC
from .lexeme cimport attr_id_t
from .vocab cimport Vocab
from .tokens.doc cimport Doc
from .tokens.doc cimport get_token_attr
from .attrs cimport ID, attr_id_t, NULL_ATTR
from .attrs import IDS
cdef enum quantifier_t:
ZERO
ZERO_ONE
ZERO_PLUS
ONE
ONE_PLUS
cdef struct AttrValueC:
attr_id_t attr
attr_t value
cdef struct TokenPatternC:
AttrValueC* attrs
int32_t nr_attr
quantifier_t quantifier
hash_t key
cdef struct ActionC:
char is_match
char keep_state
char advance_state
cdef struct PatternStateC:
TokenPatternC* state
int32_t pattern_id
int32_t start
ActionC last_action
cdef struct MatchC:
int32_t pattern_id
int32_t start
int32_t end
cdef find_matches(TokenPatternC** patterns, int n, Doc doc):
cdef vector[PatternStateC] init_states
cdef ActionC null_action = ActionC(-1, -1, -1)
for i in range(n):
init_states.push_back(PatternStateC(patterns[i], i, -1, last_action=null_action))
cdef vector[PatternStateC] curr_states
cdef vector[PatternStateC] nexts
cdef vector[MatchC] matches
cdef PreshMap cache = PreshMap()
cdef Pool mem = Pool()
# TODO: Prefill this with the extra attribute values.
extra_attrs = <attr_t**>mem.alloc(len(doc), sizeof(attr_t*))
for i in range(doc.length):
nexts.clear()
for j in range(curr_states.size()):
action = get_action(curr_states[j], &doc.c[i], extra_attrs[i], cache)
transition(matches, nexts,
action, curr_states[j], i)
for j in range(init_states.size()):
action = get_action(init_states[j], &doc.c[i], extra_attrs[i], cache)
transition(matches, nexts,
action, init_states[j], i)
nexts, curr_states = curr_states, nexts
# Filter out matches that have a longer equivalent.
longest_matches = {}
for i in range(matches.size()):
key = matches[i].pattern_id, matches[i].start
length = matches[i].end - matches[i].start
if key not in longest_matches or length > longest_matches[key]:
longest_matches[key] = length
return [(pattern_id, start, length-start)
for (pattern_id, start), length in longest_matches]
cdef void transition(vector[MatchC]& matches, vector[PatternStateC]& nexts,
ActionC action, PatternStateC state, int token) except *:
if state.start == -1:
state.start = token
if action.is_match:
matches.push_back(
MatchC(pattern_id=state.pattern_id, start=state.start, end=token+1))
if action.keep_state:
nexts.push_back(PatternStateC(pattern_id=pattern_id,
start=state.start, state=state.state, last_action=action))
if action.advance_state:
nexts.push_back(PatternStateC(pattern_id=pattern_id,
start=state.start, state=state.state+1, last_action=action))
cdef ActionC get_action(PatternStateC state, const TokenC* token, const attr_t* extra_attrs,
PreshMap cache) except *:
'''We need to consider:
a) Does the token match the specification? [Yes, No]
b) What's the quantifier? [1, 0+, ?]
c) Is this the last specification? [final, non-final]
We therefore have 12 cases to consider. For each case, we need to know
whether to emit a match, whether to keep the current state in the partials,
and whether to add an advanced state to the partials.
We therefore have eight possible results for these three booleans, which
we'll code as 000, 001 etc.
1:
- Match, final:
100
- Match, non-final:
001
- No match:
000
0+:
- Match, final:
100
- Match, non-final:
011
- Non-match, final:
100
- Non-match, non-final:
010
Problem: If a quantifier is matching, we're adding a lot of open partials
Question: Is it worth doing a lookahead, to see if we add?
'''
cached_match = <uint64_t>cache.get(state.state.key)
cdef char is_match
if cached_match == 0:
is_match = get_is_match(state, token, extra_attrs)
cached_match = is_match + 1
cache.set(state.state.key, <void*>cached_match)
elif cached_match == 1:
is_match = 0
else:
is_match = 1
quantifier = get_quantifier(state, token)
is_final = get_is_final(state, token)
if quantifier == ONE:
if not is_match:
return ActionC(is_match=0, keep_state=0, advance_state=0)
elif is_final:
return ActionC(is_match=1, keep_state=0, advance_state=0)
else:
return ActionC(is_match=0, keep_state=0, advance_state=1)
elif quantifier == ZERO_PLUS:
if is_final:
return ActionC(is_match=1, keep_state=0, advance_state=0)
elif is_match:
return ActionC(is_match=0, keep_state=1, advance_state=1)
else:
return ActionC(is_match=0, keep_state=1, advance_state=0)
elif quantifier == ZERO_ONE:
if is_final:
return ActionC(is_match=1, keep_state=0, advance_state=0)
elif is_match:
if state.last_action.keep_state:
return ActionC(is_match=0, keep_state=0, advance_state=1)
else:
return ActionC(is_match=0, keep_state=1, advance_state=1)
else:
print(quantifier, is_match, is_final)
raise ValueError
cdef char get_is_match(PatternStateC state, const TokenC* token, const attr_t* extra_attrs) nogil:
spec = state.state
for attr in spec.attrs[:spec.nr_attr]:
if get_token_attr(token, attr.attr) != attr.value:
return 0
else:
return 1
cdef char get_is_final(PatternStateC state, const TokenC* token) nogil:
if state.state[1].attrs[0].attr == ID and state.state[1].nr_attr == 0:
return 1
else:
return 0
cdef char get_quantifier(PatternStateC state, const TokenC* token) nogil:
return state.state.quantifier
cdef TokenPatternC* init_pattern(Pool mem, attr_t entity_id,
object token_specs) except NULL:
pattern = <TokenPatternC*>mem.alloc(len(token_specs) + 1, sizeof(TokenPatternC))
cdef int i
for i, (quantifier, spec) in enumerate(token_specs):
pattern[i].quantifier = quantifier
pattern[i].attrs = <AttrValueC*>mem.alloc(len(spec), sizeof(AttrValueC))
pattern[i].nr_attr = len(spec)
for j, (attr, value) in enumerate(spec):
pattern[i].attrs[j].attr = attr
pattern[i].attrs[j].value = value
pattern[i].key = hash64(pattern[i].attrs, pattern[i].nr_attr * sizeof(AttrValueC), 0)
i = len(token_specs)
pattern[i].attrs = <AttrValueC*>mem.alloc(2, sizeof(AttrValueC))
pattern[i].attrs[0].attr = ID
pattern[i].attrs[0].value = entity_id
pattern[i].nr_attr = 0
return pattern
cdef attr_t get_pattern_key(const TokenPatternC* pattern) nogil:
while pattern.nr_attr != 0:
pattern += 1
id_attr = pattern[0].attrs[0]
return id_attr.value
def _convert_strings(token_specs, string_store):
# Support 'syntactic sugar' operator '+', as combination of ONE, ZERO_PLUS
operators = {'*': (ZERO_PLUS,), '+': (ONE, ZERO_PLUS),
'?': (ZERO_ONE,), '1': (ONE,)}
tokens = []
op = ONE
for spec in token_specs:
if not spec:
# Signifier for 'any token'
tokens.append((ONE, [(NULL_ATTR, 0)]))
continue
token = []
ops = (ONE,)
for attr, value in spec.items():
if isinstance(attr, basestring) and attr.upper() == 'OP':
if value in operators:
ops = operators[value]
else:
msg = "Unknown operator '%s'. Options: %s"
raise KeyError(msg % (value, ', '.join(operators.keys())))
if isinstance(attr, basestring):
attr = IDS.get(attr.upper())
if isinstance(value, basestring):
value = string_store.add(value)
if isinstance(value, bool):
value = int(value)
if attr is not None:
token.append((attr, value))
for op in ops:
tokens.append((op, token))
return tokens
cdef class Matcher:
"""Match sequences of tokens, based on pattern rules."""
cdef Pool mem
cdef vector[TokenPatternC*] patterns
cdef readonly Vocab vocab
cdef public object _patterns
cdef public object _entities
cdef public object _callbacks
def __init__(self, vocab):
"""Create the Matcher.
vocab (Vocab): The vocabulary object, which must be shared with the
documents the matcher will operate on.
RETURNS (Matcher): The newly constructed object.
"""
self._patterns = {}
self._entities = {}
self._callbacks = {}
self.vocab = vocab
self.mem = Pool()
def __reduce__(self):
data = (self.vocab, self._patterns, self._callbacks)
return (unpickle_matcher, data, None, None)
def __len__(self):
"""Get the number of rules added to the matcher. Note that this only
returns the number of rules (identical with the number of IDs), not the
number of individual patterns.
RETURNS (int): The number of rules.
"""
return len(self._patterns)
def __contains__(self, key):
"""Check whether the matcher contains rules for a match ID.
key (unicode): The match ID.
RETURNS (bool): Whether the matcher contains rules for this match ID.
"""
return self._normalize_key(key) in self._patterns
def add(self, key, on_match, *patterns):
"""Add a match-rule to the matcher. A match-rule consists of: an ID
key, an on_match callback, and one or more patterns.
If the key exists, the patterns are appended to the previous ones, and
the previous on_match callback is replaced. The `on_match` callback
will receive the arguments `(matcher, doc, i, matches)`. You can also
set `on_match` to `None` to not perform any actions.
A pattern consists of one or more `token_specs`, where a `token_spec`
is a dictionary mapping attribute IDs to values, and optionally a
quantifier operator under the key "op". The available quantifiers are:
'!': Negate the pattern, by requiring it to match exactly 0 times.
'?': Make the pattern optional, by allowing it to match 0 or 1 times.
'+': Require the pattern to match 1 or more times.
'*': Allow the pattern to zero or more times.
The + and * operators are usually interpretted "greedily", i.e. longer
matches are returned where possible. However, if you specify two '+'
and '*' patterns in a row and their matches overlap, the first
operator will behave non-greedily. This quirk in the semantics makes
the matcher more efficient, by avoiding the need for back-tracking.
key (unicode): The match ID.
on_match (callable): Callback executed on match.
*patterns (list): List of token descritions.
"""
for pattern in patterns:
if len(pattern) == 0:
msg = ("Cannot add pattern for zero tokens to matcher.\n"
"key: {key}\n")
raise ValueError(msg.format(key=key))
key = self._normalize_key(key)
for pattern in patterns:
specs = _convert_strings(pattern, self.vocab.strings)
self.patterns.push_back(init_pattern(self.mem, key, specs))
self._patterns.setdefault(key, [])
self._callbacks[key] = on_match
self._patterns[key].extend(patterns)
def remove(self, key):
"""Remove a rule from the matcher. A KeyError is raised if the key does
not exist.
key (unicode): The ID of the match rule.
"""
key = self._normalize_key(key)
self._patterns.pop(key)
self._callbacks.pop(key)
cdef int i = 0
while i < self.patterns.size():
pattern_key = get_pattern_key(self.patterns.at(i))
if pattern_key == key:
self.patterns.erase(self.patterns.begin()+i)
else:
i += 1
def has_key(self, key):
"""Check whether the matcher has a rule with a given key.
key (string or int): The key to check.
RETURNS (bool): Whether the matcher has the rule.
"""
key = self._normalize_key(key)
return key in self._patterns
def get(self, key, default=None):
"""Retrieve the pattern stored for a key.
key (unicode or int): The key to retrieve.
RETURNS (tuple): The rule, as an (on_match, patterns) tuple.
"""
key = self._normalize_key(key)
if key not in self._patterns:
return default
return (self._callbacks[key], self._patterns[key])
def __call__(self, Doc doc):
"""Find all token sequences matching the supplied pattern.
doc (Doc): The document to match over.
RETURNS (list): A list of `(key, start, end)` tuples,
describing the matches. A match tuple describes a span
`doc[start:end]`. The `label_id` and `key` are both integers.
"""
matches = find_matches(&self.patterns[0], self.patterns.size(), doc)
return matches
def unpickle_matcher(vocab, patterns, callbacks):
matcher = Matcher(vocab)
for key, specs in patterns.items():
callback = callbacks.get(key, None)
matcher.add(key, callback, *specs)
return matcher