Skip to content

Commit c748e09

Browse files
committed
add pxd files
1 parent f4c6fda commit c748e09

File tree

13 files changed

+55
-50
lines changed

13 files changed

+55
-50
lines changed

wikipedia2vec/dictionary.pxd

Lines changed: 12 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,12 @@
1+
cdef class Item:
2+
cdef readonly int index
3+
cdef readonly int count
4+
cdef readonly int doc_count
5+
6+
7+
cdef class Word(Item):
8+
cdef readonly str text
9+
10+
11+
cdef class Entity(Item):
12+
cdef readonly str title

wikipedia2vec/dictionary.py

Lines changed: 5 additions & 13 deletions
Original file line numberDiff line numberDiff line change
@@ -24,10 +24,6 @@
2424

2525
@cython.cclass
2626
class Item:
27-
index = cython.declare(cython.int, visibility="readonly")
28-
count = cython.declare(cython.int, visibility="readonly")
29-
doc_count = cython.declare(cython.int, visibility="readonly")
30-
3127
def __init__(self, index: int, count: int, doc_count: int):
3228
self.index = index
3329
self.count = count
@@ -36,8 +32,6 @@ def __init__(self, index: int, count: int, doc_count: int):
3632

3733
@cython.cclass
3834
class Word(Item):
39-
text = cython.declare(str, visibility="readonly")
40-
4135
def __init__(self, text: str, index: int, count: int, doc_count: int):
4236
super().__init__(index, count, doc_count)
4337
self.text = text
@@ -51,8 +45,6 @@ def __repr__(self):
5145

5246
@cython.cclass
5347
class Entity(Item):
54-
title = cython.declare(str, visibility="readonly")
55-
5648
def __init__(self, title: str, index: int, count: int, doc_count: int):
5749
super().__init__(index, count, doc_count)
5850
self.title = title
@@ -72,9 +64,9 @@ def __init__(
7264
redirect_dict: RecordTrie,
7365
word_stats: np.ndarray,
7466
entity_stats: np.ndarray,
75-
language: str,
76-
lowercase: bool,
77-
build_params: dict,
67+
language: Optional[str] = None,
68+
lowercase: Optional[bool] = None,
69+
build_params: Optional[dict] = None,
7870
min_paragraph_len: int = 0,
7971
uuid: str = "",
8072
):
@@ -342,8 +334,8 @@ def load(target: Union[str, dict], mmap: bool = True) -> "Dictionary":
342334
return Dictionary(word_dict, entity_dict, redirect_dict, word_stats, entity_stats, **target["meta"])
343335

344336

345-
_dump_db: Optional[DumpDB] = None
346-
_tokenizer: Optional[BaseTokenizer] = None
337+
_dump_db: DumpDB
338+
_tokenizer: BaseTokenizer
347339

348340

349341
def _init_worker(dump_db: DumpDB, tokenizer: BaseTokenizer):

wikipedia2vec/dump_db.pxd

Lines changed: 10 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,10 @@
1+
cdef class WikiLink:
2+
cdef readonly str title
3+
cdef readonly str text
4+
cdef readonly int start
5+
cdef readonly int end
6+
7+
cdef class Paragraph:
8+
cdef readonly str text
9+
cdef readonly list[WikiLink] wiki_links
10+
cdef readonly bint abstract

wikipedia2vec/dump_db.py

Lines changed: 0 additions & 9 deletions
Original file line numberDiff line numberDiff line change
@@ -21,11 +21,6 @@
2121

2222
@cython.cclass
2323
class WikiLink:
24-
title = cython.declare(str, visibility="readonly")
25-
text = cython.declare(str, visibility="readonly")
26-
start = cython.declare(cython.int, visibility="readonly")
27-
end = cython.declare(cython.int, visibility="readonly")
28-
2924
def __init__(self, title: str, text: str, start: int, end: int):
3025
self.title = title
3126
self.text = text
@@ -45,10 +40,6 @@ def __repr__(self):
4540

4641
@cython.cclass
4742
class Paragraph:
48-
text = cython.declare(str, visibility="readonly")
49-
wiki_links = cython.declare(List[WikiLink], visibility="readonly")
50-
abstract = cython.declare(cython.bint, visibility="readonly")
51-
5243
def __init__(self, text: str, wiki_links: List[WikiLink], abstract: bool):
5344
self.text = text
5445
self.wiki_links = wiki_links

wikipedia2vec/link_graph.py

Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -127,8 +127,8 @@ def build(
127127
return LinkGraph(dictionary, matrix.indices, matrix.indptr, build_params, uuid)
128128

129129

130-
_dump_db: Optional[DumpDB] = None
131-
_dictionary: Optional[Dictionary] = None
130+
_dump_db: DumpDB
131+
_dictionary: Dictionary
132132

133133

134134
def _init_worker(dump_db: DumpDB, dictionary_obj: dict):

wikipedia2vec/mention_db.pxd

Lines changed: 9 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,9 @@
1+
cdef class Mention:
2+
cdef readonly str text
3+
cdef readonly int index
4+
cdef readonly int link_count
5+
cdef readonly int total_link_count
6+
cdef readonly int doc_count
7+
cdef readonly int start
8+
cdef readonly int end
9+
cdef _dictionary

wikipedia2vec/mention_db.py

Lines changed: 4 additions & 13 deletions
Original file line numberDiff line numberDiff line change
@@ -23,15 +23,6 @@
2323

2424
@cython.cclass
2525
class Mention:
26-
text = cython.declare(str, visibility="readonly")
27-
index = cython.declare(int, visibility="readonly")
28-
link_count = cython.declare(int, visibility="readonly")
29-
total_link_count = cython.declare(int, visibility="readonly")
30-
doc_count = cython.declare(int, visibility="readonly")
31-
start = cython.declare(int, visibility="readonly")
32-
end = cython.declare(int, visibility="readonly")
33-
_dictionary: cython.declare(Dictionary)
34-
3526
def __init__(
3627
self,
3728
dictionary: Dictionary,
@@ -258,10 +249,10 @@ def load(target: Union[str, dict], dictionary: Dictionary) -> "MentionDB":
258249
return MentionDB(mention_trie, data_trie, dictionary, **target["kwargs"])
259250

260251

261-
_dictionary: Optional[Dictionary] = None
262-
_dump_db: Optional[DumpDB] = None
263-
_tokenizer: Optional[BaseTokenizer] = None
264-
_name_trie: Optional[Trie] = None
252+
_dictionary: Dictionary
253+
_dump_db: DumpDB
254+
_tokenizer: BaseTokenizer
255+
_name_trie: Trie
265256

266257

267258
def _init_worker(dump_db: DumpDB, dictionary_obj: dict, tokenizer: BaseTokenizer, name_trie: Trie = None):
Lines changed: 4 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,4 @@
1+
cdef class Sentence:
2+
cdef readonly str text
3+
cdef readonly int start
4+
cdef readonly int end

wikipedia2vec/utils/sentence_detector/sentence.py

Lines changed: 0 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -5,10 +5,6 @@
55

66
@cython.cclass
77
class Sentence:
8-
text = cython.declare(str, visibility="readonly")
9-
start = cython.declare(cython.int, visibility="readonly")
10-
end = cython.declare(cython.int, visibility="readonly")
11-
128
def __init__(self, text: str, start: int, end: int):
139
self.text = text
1410
self.start = start
Lines changed: 4 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,4 @@
1+
cdef class Token:
2+
cdef readonly str text
3+
cdef readonly int start
4+
cdef readonly int end

0 commit comments

Comments
 (0)