Rooba
/
spaCy
mirror of https://github.com/explosion/spaCy.git
1
1
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Revert "Changes to strings.pyx for new StringStore scheme" 

This reverts commit 22d4752d64.
This commit is contained in:
Matthew Honnibal 2016-09-30 20:19:42 +02:00
parent 9e09b39b9f
commit de01e427fd
2 changed files with 11 additions and 17 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
spacy/strings.pxd
View File

@ -28,6 +28,6 @@ cdef class StringStore:
cpdef int remove_oov_map(self, Pool mem) except -1 cpdef int remove_oov_map(self, Pool mem) except -1
cpdef attr_t intern(self, basestring base_string, Pool mem=*) except -1 cdef hash_t intern(self, unicode py_string, Pool mem=*) except UINT64_MAX
cdef const Utf8Str* _intern_utf8(self, const unsigned char* utf8_string, cdef const Utf8Str* _intern_utf8(self, const unsigned char* utf8_string,
int length) except NULL int length) except NULL

26
spacy/strings.pyx
View File

@ -86,7 +86,7 @@ cdef class StringStore:
self.size = 1 self.size = 1
if strings is not None: if strings is not None:
for string in strings: for string in strings:
self.intern(string) _ = self[string]
property size: property size:
def __get__(self): def __get__(self):
@ -115,7 +115,7 @@ cdef class StringStore:
byte_string = <bytes>string_or_id byte_string = <bytes>string_or_id
if len(byte_string) == 0: if len(byte_string) == 0:
return 0 return 0
key = <attr_t>_hash_utf8(byte_string, len(byte_string)) key = _hash_utf8(byte_string, len(byte_string))
utf8str = <Utf8Str*>self._map.get(key) utf8str = <Utf8Str*>self._map.get(key)
if utf8str is NULL: if utf8str is NULL:
raise KeyError(byte_string) raise KeyError(byte_string)
@ -124,7 +124,7 @@ cdef class StringStore:
elif isinstance(string_or_id, unicode): elif isinstance(string_or_id, unicode):
if len(<unicode>string_or_id) == 0: if len(<unicode>string_or_id) == 0:
return 0 return 0
key = <attr_t>hash_string(string_or_id) key = hash_string(string_or_id)
utf8str = <Utf8Str*>self._map.get(key) utf8str = <Utf8Str*>self._map.get(key)
if utf8str is NULL: if utf8str is NULL:
raise KeyError(string_or_id) raise KeyError(string_or_id)
@ -136,7 +136,7 @@ cdef class StringStore:
def __contains__(self, unicode string not None): def __contains__(self, unicode string not None):
if len(string) == 0: if len(string) == 0:
return True return True
cdef attr_t key = <attr_t>hash_string(string) cdef hash_t key = hash_string(string)
return self._map.get(key) is not NULL return self._map.get(key) is not NULL
def __iter__(self): def __iter__(self):
@ -154,17 +154,12 @@ cdef class StringStore:
strings.append(py_string) strings.append(py_string)
return (StringStore, (strings,), None, None, None) return (StringStore, (strings,), None, None, None)
cpdef attr_t intern(self, basestring base_string, Pool mem=None) except -1: cdef hash_t intern(self, unicode py_string, Pool mem=None) except UINT64_MAX:
if len(base_string) == 0:
return 0
if mem is None: if mem is None:
mem = self.mem mem = self.mem
if isinstance(base_string, unicode):
byte_string = base_string.encode('utf8')
else:
byte_string = base_string
cdef hash_t map_key = id(mem) cdef hash_t map_key = id(mem)
cdef attr_t key = <attr_t>_hash_utf8(byte_string, len(byte_string)) cdef bytes byte_string = py_string.encode('utf8')
cdef hash_t key = _hash_utf8(byte_string, len(byte_string))
cdef const Utf8Str* utf8str = <Utf8Str*>self._map.get(key) cdef const Utf8Str* utf8str = <Utf8Str*>self._map.get(key)
cdef hash_t map_id = id(mem) cdef hash_t map_id = id(mem)
cdef MapStruct* oov_map cdef MapStruct* oov_map
@ -184,7 +179,7 @@ cdef class StringStore:
map_set(mem, oov_map, key, new_utf8str) map_set(mem, oov_map, key, new_utf8str)
return key return key
def decode_int(self, attr_t int_, Pool mem=None): def decode_int(self, hash_t int_, Pool mem=None):
cdef hash_t map_key cdef hash_t map_key
if int_ == 0: if int_ == 0:
return u'' return u''
@ -218,16 +213,15 @@ cdef class StringStore:
int length) except NULL: int length) except NULL:
if self.size == self._resize_at: if self.size == self._resize_at:
self._realloc() self._realloc()
cdef attr_t key = <attr_t>_hash_utf8(utf8_string, length) key = _hash_utf8(utf8_string, length)
self.c[self.size] = _allocate(self.mem, utf8_string, length) self.c[self.size] = _allocate(self.mem, utf8_string, length)
self._map.set(key, <void*>&self.c[self.size]) self._map.set(key, <void*>&self.c[self.size])
self.size += 1 self.size += 1
return &self.c[self.size-1] return &self.c[self.size-1]
cpdef int remove_oov_map(self, Pool mem) except -1: cpdef int remove_oov_map(self, Pool mem) except -1:
print("Remove mem", id(mem))
cdef hash_t key = id(mem) cdef hash_t key = id(mem)
self.oov_maps.pop(key) self._maps.pop(key)
def dump(self, file_): def dump(self, file_):
# TODO: Is it problematic that we don't save the OOV strings? No, right? # TODO: Is it problematic that we don't save the OOV strings? No, right?