asahi417 commited on
Commit
5399d79
1 Parent(s): fe6c94b
data/tempo_wic/test.jsonl CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:c1ccb104f713ad9d7852640d15b6d0c286327f571970b3edce9c26509651afdd
3
- size 1416017
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:975f67ba599362ef3bea5d749932abe572f2443e73b1d8e4e8a753cd5acd6162
3
+ size 1284151
data/tempo_wic/train.jsonl CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:ff7b7b23ee1c8b3b7a87b80d460b145beaed7a7c8f1d2722cd56b2072aa141e7
3
- size 1574854
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3972efcca73ec01c81feceec04e8fa2efa9e00d33f827decb8e8a7033c48507d
3
+ size 1439969
data/tempo_wic/validation.jsonl CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:42052a317896a2722fabf71b797673cd91e5d98b5330623abdf414c3c60d6f1a
3
- size 383998
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b44124d18d379d1b8330a2c3872cd5ea0eec0b6fdac2d3eb8d268f652a274b6c
3
+ size 347362
data/tweet_intimacy/test.jsonl CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:bb6d7c7be0302e3bc65a3d81458ca3c9fb86b0c9917c6660ace797a139893d9d
3
- size 40914
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:38e723fe91f2f18ec346c82aff75ac7323fc4098b789d9b266d581945c1be37a
3
+ size 38638
data/tweet_intimacy/train.jsonl CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:86b44b6a7877310f0964dd47dbd6c83427d795177583b17bacd4aadb31fd3568
3
- size 127091
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1af59bd18d4443e1276a7462e8fd15ddf62f62236a16b54770203e7ac840879b
3
+ size 119898
data/tweet_intimacy/validation.jsonl CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:471770bd25edb1b968492e2ae15b89c820d2854f7785d13ac10c6575f86cecc8
3
- size 42661
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:56f3b4d36f87ca43a7eb3033775194169ef3e48bf11ac6c4b943ae194f035357
3
+ size 40087
process/tempo_wic.py CHANGED
@@ -22,12 +22,17 @@ for s in ['train', 'validation', 'test']:
22
  data_jl = []
23
  for _, i in data.iterrows():
24
  i = i.to_dict()
25
- tmp = {"id": i['id'], "word": i["word"], "label_binary": i["label"]}
26
  tmp.update({f"{k}_1": v for k, v in i['tweet1'].items()})
27
- tmp.update({f"{k}_2": v for k, v in i['tweet2'].items()})
28
  tmp['text_1_tokenized'] = tmp.pop('tokens_1')
 
29
  tmp['text_2_tokenized'] = tmp.pop('tokens_2')
 
 
 
 
 
30
  data_jl.append(tmp)
 
31
  with open(f"data/tempo_wic/{s}.jsonl", "w") as f:
32
  f.write("\n".join([json.dumps(i) for i in data_jl]))
33
-
 
22
  data_jl = []
23
  for _, i in data.iterrows():
24
  i = i.to_dict()
25
+ tmp = {"word": i["word"], "gold_label_binary": i["label"]}
26
  tmp.update({f"{k}_1": v for k, v in i['tweet1'].items()})
 
27
  tmp['text_1_tokenized'] = tmp.pop('tokens_1')
28
+ tmp.update({f"{k}_2": v for k, v in i['tweet2'].items()})
29
  tmp['text_2_tokenized'] = tmp.pop('tokens_2')
30
+ tmp.pop("id")
31
+ tmp.pop("text_start_1")
32
+ tmp.pop("text_end_1")
33
+ tmp.pop("text_start_2")
34
+ tmp.pop("text_end_2")
35
  data_jl.append(tmp)
36
+
37
  with open(f"data/tempo_wic/{s}.jsonl", "w") as f:
38
  f.write("\n".join([json.dumps(i) for i in data_jl]))
 
process/tweet_intimacy.py CHANGED
@@ -9,14 +9,14 @@ df_test = df_test[df_test['language'] == 'English']
9
  df_test.pop("language")
10
  test = [i.to_dict() for _, i in df_test.iterrows()]
11
  for i in test:
12
- i['label_float'] = i.pop("label")
13
 
14
  df_train = pd.read_csv("misc/multilingual_tweet_intimacy/train.csv")
15
  df_train = df_train[df_train['language'] == 'English']
16
  df_train.pop("language")
17
  train = [i.to_dict() for _, i in df_train.iterrows()]
18
  for i in train:
19
- i['label_float'] = i.pop("label")
20
  seed(42)
21
  shuffle(train)
22
  val = train[:len(test)]
 
9
  df_test.pop("language")
10
  test = [i.to_dict() for _, i in df_test.iterrows()]
11
  for i in test:
12
+ i['gold_score'] = i.pop("label")
13
 
14
  df_train = pd.read_csv("misc/multilingual_tweet_intimacy/train.csv")
15
  df_train = df_train[df_train['language'] == 'English']
16
  df_train.pop("language")
17
  train = [i.to_dict() for _, i in df_train.iterrows()]
18
  for i in train:
19
+ i['gold_score'] = i.pop("label")
20
  seed(42)
21
  shuffle(train)
22
  val = train[:len(test)]
process/tweet_ner.py CHANGED
@@ -10,10 +10,13 @@ def process(tmp):
10
  tmp.pop("label_name")
11
  tmp = [i.to_dict() for _, i in tmp.iterrows()]
12
  for i in tmp:
13
- i['label_sequence'] = i.pop('tags').tolist()
 
14
  i['text_tokenized'] = i.pop('token').tolist()
 
15
  return tmp
16
 
 
17
  train = process(data["train_2020"].to_pandas())
18
  val = process(data["validation_2020"].to_pandas())
19
  test = process(data["test_2021"].to_pandas())
 
10
  tmp.pop("label_name")
11
  tmp = [i.to_dict() for _, i in tmp.iterrows()]
12
  for i in tmp:
13
+ i.pop("id")
14
+ i['gold_label_sequence'] = i.pop('tags').tolist()
15
  i['text_tokenized'] = i.pop('token').tolist()
16
+ i['text'] = ' '.join(i['text_tokenized'])
17
  return tmp
18
 
19
+
20
  train = process(data["train_2020"].to_pandas())
21
  val = process(data["validation_2020"].to_pandas())
22
  test = process(data["test_2021"].to_pandas())
process/unify_sp_symbol_2.py ADDED
@@ -0,0 +1,44 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import json
2
+ import re
3
+
4
+ from glob import glob
5
+
6
+ for i in glob("data/tweet_intimacy/*.jsonl"):
7
+ with open(i) as f:
8
+ data = [json.loads(j) for j in f.readlines()]
9
+ for d in data:
10
+ for c in ['text']:
11
+ tmp = d[c].replace("{{URL}}", "@url")
12
+ # tmp = re.sub(r"\bhttp\b", "{{URL}}", tmp)
13
+ tmp = re.sub(r"@user", "{{USERNAME}}", tmp)
14
+ tmp = re.sub(r"(@[\S]+)", r"{\1@}", tmp)
15
+ d[c] = tmp
16
+
17
+ with open(i, "w") as f:
18
+ f.write("\n".join([json.dumps(j) for j in data]))
19
+
20
+ # for i in glob("data/tweet_qa/*.jsonl"):
21
+ # with open(i) as f:
22
+ # data = [json.loads(j) for j in f.readlines()]
23
+ # for d in data:
24
+ # for c in ['text', "paragraph", "question", "label_str"]:
25
+ # tmp = d[c]
26
+ # tmp = re.sub(r"(@[\S]+)", r"{\1@}", tmp)
27
+ # tmp = tmp.replace(")@}", '@})')
28
+ # d[c] = tmp
29
+ #
30
+ # with open(i, "w") as f:
31
+ # f.write("\n".join([json.dumps(j) for j in data]))
32
+
33
+ # for i in glob("data/tweet_similarity/*.jsonl"):
34
+ # with open(i) as f:
35
+ # data = [json.loads(j) for j in f.readlines()]
36
+ # for d in data:
37
+ # for c in ['text_1', "text_2"]:
38
+ # tmp = d[c]
39
+ # # tmp = re.sub(r"(@[\S]+)\b", r"{\1@}", tmp)
40
+ # tmp = tmp.replace("{@user@}", "{{USERNAME}}")
41
+ # d[c] = tmp
42
+ #
43
+ # with open(i, "w") as f:
44
+ # f.write("\n".join([json.dumps(j) for j in data]))