-
Notifications
You must be signed in to change notification settings - Fork 0
/
skb_bridge.py
325 lines (288 loc) · 14.3 KB
/
skb_bridge.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
import difflib
from collections.abc import Iterable
from stark_qa.skb import SKB
from logger import Logger
from pathfinding import find_edge_type
from utils import edge_type2str
from vss import VSS
def create_node_dict_prime(skb: SKB):
nodes_alias2id = {}
for n_type in skb.node_type_lst():
nodes_alias2id[n_type] = {}
for i in range(skb.num_nodes()):
node = skb.node_info[i]
n_type = node["type"]
n_name = node['name'].lower()
nodes_alias2id[n_type][n_name] = i
if 'details' in node:
if 'alias' in node['details']:
alias = node['details']['alias']
if isinstance(alias, list):
for a in alias:
nodes_alias2id[n_type][a.lower()] = i
else:
nodes_alias2id[n_type][alias.lower()] = i
return nodes_alias2id
def create_node_dict_mag(skb: SKB):
nodes_alias2id = {}
for n_type in skb.node_type_lst():
nodes_alias2id[n_type] = {}
for i in range(skb.num_nodes()):
node = skb.node_info[i]
n_type = node["type"]
if 'title' in node:
nodes_alias2id[n_type][node['title'].lower()] = i
elif 'DisplayName' in node and node['DisplayName'] != -1 and node['DisplayName'] != "-1":
nodes_alias2id[n_type][node['DisplayName'].lower()] = i
return nodes_alias2id
def create_node_dict_amazon(skb: SKB):
raise NotImplementedError("ID: 456klj23hed (create_node_dict_amazon not implemented)")
def find_closest_nodes_mag(targets: Iterable, nodes_alias2id: {}, all_node_types: list[str], cutoff: float) -> set:
raise NotImplementedError("Not further implemented, because nodes are not sorted by node type at the current state"
"of development.")
node_ids = []
# TODO /// targets ... /// get their types then search
for n_name, n_type in targets:
if n_type == -1:
n_type = all_node_types
else:
n_type = [n_type]
closest_key = difflib.get_close_matches(target.lower(), nodes_alias2id[n_type], n=1, cutoff=0.9)
if len(closest_key) > 0:
closest_key = closest_key[0]
node_ids.append(nodes_alias2id[closest_key])
return node_ids
class SKBbridge:
def __init__(self, dataset_name: str, skb: SKB = None, vss: VSS = None):
if dataset_name not in ['prime', 'mag', 'amazon']:
raise ValueError(f"Dataset {dataset_name} not found. It should be in ['prime', 'mag,', 'amazon']")
self.name = dataset_name
self.vss = vss
self.skb = skb
self.nodes_alias2id = None
self.create_node_dict() # replaces self.nodes_alias2id
self.is_directed = False
if self.name == 'prime':
self.is_directed = True
def create_node_dict(self):
if self.name == 'prime':
self.nodes_alias2id = create_node_dict_prime(self.skb)
elif self.name == 'mag':
self.nodes_alias2id = create_node_dict_mag(self.skb)
elif self.name == 'amazon':
self.nodes_alias2id = create_node_dict_amazon(self.skb)
else:
raise ValueError(f"dataset name should be in ['prime', 'mag,', 'amazon'], but '{self.name}' is given")
def find_closest_nodes(self, targets: Iterable, cutoff: float, drop_duplicates: bool) -> set:
# if self.name == "mag":
# return find_closest_nodes_mag(targets, self.nodes_alias2id, cutoff)
node_ids = []
for target in targets:
n_type = target[1]
n_name = target[0].lower()
if n_type is None:
closest_node_name = ""
highest_similarity = 0.
closest_node_type = ""
for n_type in self.skb.node_type_lst():
potential_closest_node = difflib.get_close_matches(n_name, self.nodes_alias2id[n_type],
n=1, cutoff=cutoff)
if len(potential_closest_node) > 0:
similarity = difflib.SequenceMatcher(None, n_name, potential_closest_node[0]).ratio()
if similarity > highest_similarity:
closest_node_name = potential_closest_node[0]
highest_similarity = similarity
closest_node_type = n_type
if closest_node_name != "":
node_ids.append(self.nodes_alias2id[closest_node_type][closest_node_name])
elif self.vss is not None:
node_ids.append(self.vss.get_top_k_nodes(n_name, node_id_mask=None, max_k=1)[0])
else:
closest_node_name = difflib.get_close_matches(n_name, self.nodes_alias2id[n_type], n=1, cutoff=cutoff)
if len(closest_node_name) > 0:
closest_node_name = closest_node_name[0]
node_ids.append(self.nodes_alias2id[n_type][closest_node_name])
else:
filtered_node_ids = self.skb.get_node_ids_by_type(n_type)
node_ids.append(self.vss.get_top_k_nodes(n_name, node_id_mask=filtered_node_ids, max_k=1)[0])
if drop_duplicates:
node_ids = list(set(node_ids))
return node_ids
def find_closest_nodes_w_VSS(self, targets: Iterable, cutoff: float, drop_duplicates: bool) -> set:
# if self.name == "mag":
# return find_closest_nodes_mag(targets, self.nodes_alias2id)
node_ids = []
for target in targets:
n_type = target[1]
n_name = target[0].lower()
if n_type is None:
closest_node_name = ""
highest_similarity = 0.
closest_node_type = ""
for n_type in self.skb.node_type_lst():
potential_closest_node = difflib.get_close_matches(n_name, self.nodes_alias2id[n_type],
n=1, cutoff=cutoff)
if len(potential_closest_node) > 0:
similarity = difflib.SequenceMatcher(None, n_name, potential_closest_node[0]).ratio()
if similarity > highest_similarity:
closest_node_name = potential_closest_node[0]
highest_similarity = similarity
closest_node_type = n_type
if closest_node_name != "":
node_ids.append(self.nodes_alias2id[closest_node_type][closest_node_name])
else:
print('TODO') # vss.
else:
closest_node_name = difflib.get_close_matches(n_name, self.nodes_alias2id[n_type], n=1, cutoff=cutoff)
if len(closest_node_name) > 0:
closest_node_name = closest_node_name[0]
node_ids.append(self.nodes_alias2id[n_type][closest_node_name])
if drop_duplicates:
node_ids = list(set(node_ids))
return node_ids
def entity_id2name(self, id: int):
if self.name == 'prime':
return self.skb.node_info[id]['name']
if self.name == 'mag':
node = self.skb.node_info[id]
if 'title' in node:
return node['title']
elif 'DisplayName' in node and node['DisplayName'] != -1 and node['DisplayName'] != "-1":
return node['DisplayName']
else:
return f"node without name. id: {id}"
raise NotImplementedError(f"Not implemented for dataset {self.name}")
def path2str(self, path):
out = ""
for i in range(len(path) - 1):
edge_type = find_edge_type(path[i], path[i + 1], self.skb)
out += (f"{self.skb.node_info[path[i]]['name']} {edge_type2str(self.name, edge_type)} "
f"{self.skb.node_info[path[i + 1]]['name']}, ")
return out[:-2]
def nodes2str(self, node_ids: int | list[str]):
if isinstance(node_ids, list):
out = []
for node_id in node_ids:
out.append(self.skb.get_doc_info(node_id, add_rel=False, compact=False))
return out
else:
return self.skb.get_doc_info(node_ids, add_rel=False, compact=False)
def get_node_type_from_key_str(self, key_str, logger: Logger):
type_of_unknown = None
try:
key_str = int(key_str)
if key_str in self.skb.node_type_dict:
type_of_unknown = self.skb.node_type_dict[key_str]
except ValueError("Input should be natural number or -1."):
logger.log("ValueError(Input should be natural number or -1.)")
return type_of_unknown
def find_unknowns_from_triplets(self, triplets: list[str], logger: Logger, cutoff: float) -> {}:
skb = self.skb
unknowns = {}
unknowns_sizes = []
nothing_changed = False
rounds = 0
while (not nothing_changed):
logger.log(f"Round {rounds}:")
for triplet in triplets:
triplet = triplet.split("->")
if len(triplet) != 3:
logger.log("triplet" + str(triplet) + "is not valid.")
continue
u = triplet[0].strip()
e = triplet[1].strip()
v = triplet[2].strip()
if e not in skb.edge_type_dict.values():
logger.log("Warning: Edge type " + e + " not found in knowledge base. Replacing it with None.")
e = None
# u:
if "'" in u:
u = u.split("'")
u_type = u[0].strip()
u_name = u[1].strip()
if u_type not in skb.node_type_lst():
logger.log("Warning: Node type " + u_type + " not found in knowledge base. Replacing it with None.")
u_type = None
node_ids = self.find_closest_nodes([(u_name, u_type)],
cutoff=cutoff,
# experiment.config["find_closest_nodes_cut_off"],
drop_duplicates=True)
if len(node_ids) == 0:
logger.log("Node " + u + " not found in knowledge base.")
continue
u = node_ids[0]
u_is_constant = True
logger.log(f"Found entity: {self.entity_id2name(u)} ({u_type})")
elif "|" in u:
u_is_constant = False
u = u.split("|")
u_type = u[0].strip()
u_name = u[1].strip()
else:
logger.log("first node " + u + "is not valid.")
continue
# v:
if "'" in v:
v = v.split("'")
v_type = v[0].strip()
v_name = v[1].strip()
if v_type not in skb.node_type_lst():
v_type = None
node_ids = self.find_closest_nodes([(v_name, v_type)],
cutoff=cutoff,
# experiment.config["find_closest_nodes_cut_off"],
drop_duplicates=True)
if len(node_ids) == 0:
logger.log("Node " + v + " not found in knowledge base.")
continue
v = node_ids[0]
v_is_constant = True
logger.log(f"Found entity: {self.entity_id2name(v)} ({v_type})")
elif "|" in v:
v = v.split("|")
v_type = v[0].strip()
v_name = v[1].strip()
v_is_constant = False
else:
logger.log("second node " + v + "is not valid.")
continue
if u_is_constant and not v_is_constant:
candidates = set(skb.get_neighbor_nodes(u, e))
if v_name in unknowns:
unknowns[v_name].intersection_update(candidates)
else:
unknowns[v_name] = candidates
logger.log(
f"Found triplet: {self.entity_id2name(u)} ({u_type}) -> {e} -> {len(unknowns[v_name])} candidates of type {v_type}")
elif not u_is_constant and v_is_constant:
candidates = set(skb.get_neighbor_nodes(v, e))
if u_name in unknowns:
unknowns[u_name].intersection_update(candidates)
else:
unknowns[u_name] = candidates
logger.log(
f"Found triplet: {len(unknowns[u_name])} candidates of type {u_type} -> {e} -> {self.entity_id2name(v)} ({v_type})")
elif not u_is_constant and not v_is_constant:
if u_name in unknowns:
candidates = set()
for u_candidate in unknowns[u_name]:
candidates = candidates.union(set(skb.get_neighbor_nodes(u_candidate, e)))
if v_name in unknowns:
unknowns[v_name].intersection_update(candidates)
else:
unknowns[v_name] = candidates
if v_name in unknowns:
candidates = set()
for v_candidate in unknowns[v_name]:
candidates = candidates.union(set(skb.get_neighbor_nodes(v_candidate, e)))
if u_name in unknowns:
unknowns[u_name].intersection_update(candidates)
else:
unknowns[u_name] = candidates
logger.log(f"Found {len(candidates)} candidates for triplet: type {u_type} -> {e} -> {v_type}")
new_unknown_lengths = [len(x) for x in unknowns.values()]
if unknowns_sizes == new_unknown_lengths:
nothing_changed = True
else:
unknowns_sizes = new_unknown_lengths
return unknowns