init
Browse files- .gitattributes +2 -0
- README.md +14 -14
- check_predicate.py +10 -0
- create_split.py +0 -10
- data/stats.data_size.csv +3 -5
- data/t_rex.filter.sample.csv +3 -0
- filtering_purify.py +8 -6
.gitattributes
CHANGED
@@ -61,3 +61,5 @@ data/t_rex.filter.validation.jsonl filter=lfs diff=lfs merge=lfs -text
|
|
61 |
data/t_rex.filter.test.jsonl filter=lfs diff=lfs merge=lfs -text
|
62 |
data/t_rex.filter.train.jsonl filter=lfs diff=lfs merge=lfs -text
|
63 |
data/t_rex.filter.jsonl filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
61 |
data/t_rex.filter.test.jsonl filter=lfs diff=lfs merge=lfs -text
|
62 |
data/t_rex.filter.train.jsonl filter=lfs diff=lfs merge=lfs -text
|
63 |
data/t_rex.filter.jsonl filter=lfs diff=lfs merge=lfs -text
|
64 |
+
data/stats.data_size.csv filter=lfs diff=lfs merge=lfs -text
|
65 |
+
data/t_rex.filter.sample.csv filter=lfs diff=lfs merge=lfs -text
|
README.md
CHANGED
@@ -37,20 +37,20 @@ we choose top-`max predicate` triples based on the frequency of the subject and
|
|
37 |
|
38 |
- number of triples in each configuration
|
39 |
|
40 |
-
| min entity / max predicate | 10 |
|
41 |
-
|
42 |
-
|
|
43 |
-
|
|
44 |
-
|
|
45 |
-
|
|
46 |
-
|
47 |
-
|
48 |
-
| min entity
|
49 |
-
|
50 |
-
|
|
51 |
-
|
|
52 |
-
|
|
53 |
-
|
|
54 |
|
55 |
- distribution of entities
|
56 |
|
|
|
37 |
|
38 |
- number of triples in each configuration
|
39 |
|
40 |
+
| min entity / max predicate | 10 | 25 | 50 | 100 |
|
41 |
+
|-----------------------------:|-----:|-----:|------:|------:|
|
42 |
+
| 4 | 4501 | 9245 | 15196 | 23936 |
|
43 |
+
| 8 | 3557 | 7291 | 11804 | 18699 |
|
44 |
+
| 12 | 3132 | 6346 | 10155 | 16115 |
|
45 |
+
| 16 | 2769 | 5559 | 9014 | 14499 |
|
46 |
+
|
47 |
+
|
48 |
+
| min entity | predicate |
|
49 |
+
|-------------:|-----:|
|
50 |
+
| 4 | 516 |
|
51 |
+
| 8 | 409 |
|
52 |
+
| 12 | 366 |
|
53 |
+
| 16 | 321 |
|
54 |
|
55 |
- distribution of entities
|
56 |
|
check_predicate.py
ADDED
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import json
|
2 |
+
import numpy as np
|
3 |
+
import pandas as pd
|
4 |
+
|
5 |
+
with open("data/t_rex.filter.jsonl") as f:
|
6 |
+
data = pd.DataFrame([json.loads(i) for i in f.read().split('\n') if len(i) > 0])
|
7 |
+
tmp = data.groupby("predicate").sample(10, replace=True)
|
8 |
+
tmp = tmp.drop_duplicates()
|
9 |
+
tmp.to_csv("data/t_rex.filter.sample.csv", index=False)
|
10 |
+
|
create_split.py
CHANGED
@@ -1,10 +0,0 @@
|
|
1 |
-
import json
|
2 |
-
import numpy as np
|
3 |
-
|
4 |
-
with open("data/t_rex.filter.jsonl") as f:
|
5 |
-
data = [json.loads(i) for i in f.read().split('\n') if len(i) > 0]
|
6 |
-
|
7 |
-
p, c = np.unique([i['predicate'] for i in data], return_counts=True)
|
8 |
-
d = dict(zip(p.tolist(), c.tolist()))
|
9 |
-
with open("data/t_rex.filter.predicate.json", 'w') as f:
|
10 |
-
json.dump(d, f)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
data/stats.data_size.csv
CHANGED
@@ -1,5 +1,3 @@
|
|
1 |
-
|
2 |
-
|
3 |
-
|
4 |
-
3,4836,9943,16449,26001,557
|
5 |
-
4,4501,9245,15196,23936,516
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:e9957f1edf8f50f209ee22c5b2c99aedf8f432dfc6d92beb13541ac6d8d81c06
|
3 |
+
size 164
|
|
|
|
data/t_rex.filter.sample.csv
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:e6733c3707cdbd270a070fd94f5717495c0f6d4e452dc6d4d4986fab741b35f7
|
3 |
+
size 6123997
|
filtering_purify.py
CHANGED
@@ -14,6 +14,8 @@ from datasets import Dataset
|
|
14 |
|
15 |
parameters_min_e_freq = [4, 8, 12, 16]
|
16 |
parameters_max_p_freq = [100, 50, 25, 10]
|
|
|
|
|
17 |
sns.set_theme(style="whitegrid")
|
18 |
|
19 |
# load filtered data
|
@@ -119,9 +121,9 @@ if __name__ == '__main__':
|
|
119 |
df_p.columns = [f"min entity: {mef}, max predicate: {mpf}" for mef, mpf in candidates]
|
120 |
fig, axes = plt.subplots(2, 2, constrained_layout=True)
|
121 |
fig.suptitle('Predicate Distribution over Different Configurations')
|
122 |
-
for (x, y), mpf in zip([(0, 0), (0, 1), (1, 0), (1, 1)],
|
123 |
-
_df = df_p[[f"min entity: {mef}, max predicate: {mpf}" for mef in
|
124 |
-
_df.columns = [f"min entity: {mef}" for mef in
|
125 |
ax = sns.lineplot(ax=axes[x, y], data=_df, linewidth=1)
|
126 |
if mpf != 100:
|
127 |
ax.legend_.remove()
|
@@ -136,9 +138,9 @@ if __name__ == '__main__':
|
|
136 |
df_e.columns = [f"min entity: {mef}, max predicate: {mpf}" for mef, mpf in candidates]
|
137 |
fig, axes = plt.subplots(2, 2, constrained_layout=True)
|
138 |
fig.suptitle('Entity Distribution over Different Configurations')
|
139 |
-
for (x, y), mpf in zip([(0, 0), (0, 1), (1, 0), (1, 1)],
|
140 |
-
_df = df_e[[f"min entity: {mef}, max predicate: {mpf}" for mef in
|
141 |
-
_df.columns = [f"min entity: {mef}" for mef in
|
142 |
ax = sns.lineplot(ax=axes[x, y], data=_df, linewidth=1)
|
143 |
ax.set(xscale='log')
|
144 |
if mpf != 100:
|
|
|
14 |
|
15 |
parameters_min_e_freq = [4, 8, 12, 16]
|
16 |
parameters_max_p_freq = [100, 50, 25, 10]
|
17 |
+
assert len(parameters_min_e_freq) == 4
|
18 |
+
assert len(parameters_max_p_freq) == 4
|
19 |
sns.set_theme(style="whitegrid")
|
20 |
|
21 |
# load filtered data
|
|
|
121 |
df_p.columns = [f"min entity: {mef}, max predicate: {mpf}" for mef, mpf in candidates]
|
122 |
fig, axes = plt.subplots(2, 2, constrained_layout=True)
|
123 |
fig.suptitle('Predicate Distribution over Different Configurations')
|
124 |
+
for (x, y), mpf in zip([(0, 0), (0, 1), (1, 0), (1, 1)], parameters_max_p_freq):
|
125 |
+
_df = df_p[[f"min entity: {mef}, max predicate: {mpf}" for mef in parameters_min_e_freq]]
|
126 |
+
_df.columns = [f"min entity: {mef}" for mef in parameters_min_e_freq]
|
127 |
ax = sns.lineplot(ax=axes[x, y], data=_df, linewidth=1)
|
128 |
if mpf != 100:
|
129 |
ax.legend_.remove()
|
|
|
138 |
df_e.columns = [f"min entity: {mef}, max predicate: {mpf}" for mef, mpf in candidates]
|
139 |
fig, axes = plt.subplots(2, 2, constrained_layout=True)
|
140 |
fig.suptitle('Entity Distribution over Different Configurations')
|
141 |
+
for (x, y), mpf in zip([(0, 0), (0, 1), (1, 0), (1, 1)], parameters_max_p_freq):
|
142 |
+
_df = df_e[[f"min entity: {mef}, max predicate: {mpf}" for mef in parameters_min_e_freq]]
|
143 |
+
_df.columns = [f"min entity: {mef}" for mef in parameters_min_e_freq]
|
144 |
ax = sns.lineplot(ax=axes[x, y], data=_df, linewidth=1)
|
145 |
ax.set(xscale='log')
|
146 |
if mpf != 100:
|