-
Notifications
You must be signed in to change notification settings - Fork 0
/
preprocess.py
173 lines (145 loc) · 7.81 KB
/
preprocess.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
import pandas as pd
import numpy as np
from tqdm import tqdm
from glob import glob
from utilities import add_disirable_label
evidence_data = pd.read_excel("./data/college/EvidenceSentencePairData.xlsx")
reason_data = pd.read_excel("./data/college/ReasoningSentencePairData.xlsx")
doc_paths = glob("./data/college/revisions/*/*.xlsx")
eval_gold = []
eval_predict = []
purpose_dict = set()
for path in tqdm(doc_paths):
print("Data file: ", path)
if "esl" in path:
essay_name = str(777777) + path.split("/")[-1].split(".")[0].split("_")[-1].replace("esl", "")
essay_id = path.split("/")[-1].split(".")[0].split("_")[-1]
else:
essay_name = str(666666) + path.split("/")[-1].split(".")[0].split("_")[-1].replace("native", "")
essay_id = path.split("/")[-1].split(".")[0].split("_")[-1]
# if "e23." in path or "e127." in path:
# print()
# continue
# if "e10." in path:
# print()
# if "e153." in path or "e152." in path or "e109." in path or "e23." in path or "e52." in path or "e124." in path or "e50." in path or "e127." in path or "e55." in path:
# # continue
# print("skip")
# continue
old_draft = pd.read_excel(path, sheet_name="Old Draft")
new_draft = pd.read_excel(path, sheet_name="New Draft")
purpose_dict.update(old_draft["Revision Purpose Level 0"].tolist())
purpose_dict.update(new_draft["Revision Purpose Level 0"].tolist())
if "General Content Development,Conventions/Grammar/Spelling" in old_draft["Revision Purpose Level 0"].tolist():
print("----------")
old_draft_list = old_draft["Sentence Content"].tolist()
new_draft_list = new_draft["Sentence Content"].tolist()
tmp_old_gold_align = old_draft["Aligned Index"].apply(lambda x: x-1 if type(x)!=str else x).apply(lambda x: str(int(x)-1) if type(x)==str and x not in ["ADD", "DELETE"] else x).tolist()
tmp_new_gold_align = new_draft["Aligned Index"].apply(lambda x: x-1 if type(x)!=str else x).apply(lambda x: str(int(x)-1) if type(x)==str and x not in ["ADD", "DELETE"] else x).tolist()
old_sent_index = old_draft["Sentence Index"].apply(lambda x: x-1 if type(x)!=str else x).apply(lambda x: str(int(x)-1) if type(x)==str and x not in ["ADD", "DELETE"] else x).tolist()
new_sent_index = new_draft["Sentence Index"].apply(lambda x: x-1 if type(x)!=str else x).apply(lambda x: str(int(x)-1) if type(x)==str and x not in ["ADD", "DELETE"] else x).tolist()
old_gold_align = old_draft["Aligned Index"].apply(lambda x: x-1 if type(x)!=str else x).apply(lambda x: str(int(x)-1) if type(x)==str and x not in ["ADD", "DELETE"] else x).tolist()
new_gold_align = new_draft["Aligned Index"].apply(lambda x: x-1 if type(x)!=str else x).apply(lambda x: str(int(x)-1) if type(x)==str and x not in ["ADD", "DELETE"] else x).tolist()
old_purpose = old_draft["Revision Purpose Level 0"].tolist()
new_purpose = new_draft["Revision Purpose Level 0"].tolist()
# new draft
latest_align_num = 0
inserted_num = 0
for i in range(len(tmp_new_gold_align)):
tmp_align = tmp_new_gold_align[i]
if str(tmp_align) != "ADD":
latest_align_num = int(tmp_align)
else:
j = latest_align_num + inserted_num + 1
# if delete and insert are conjuncted, do delete first then insert
while j < len(old_gold_align) and old_gold_align[j] == "DELETE":
j += 1
old_draft_list.insert(j, np.NaN)
old_sent_index.insert(j, np.NaN)
old_gold_align.insert(j, np.NaN)
old_purpose.insert(j, np.NaN)
inserted_num += 1
# old draft
latest_align_num = 0
inserted_num = 0
for i in range(len(tmp_old_gold_align)):
tmp_align = tmp_old_gold_align[i]
if str(tmp_align) != "DELETE":
latest_align_num = int(tmp_align)
else:
j = latest_align_num + inserted_num + 1
new_draft_list.insert(j, np.NaN)
new_sent_index.insert(j, np.NaN)
new_gold_align.insert(j, np.NaN)
new_purpose.insert(j, np.NaN)
inserted_num += 1
purpose_mapper = {'Evidence,Warrant/Reasoning/Backing': 'content',
'Warrant/Reasoning/Backing': 'content',
'Conventions/Grammar/Spelling': 'surface',
'Warrant/Reasoning/Backing,General Content Development': 'content',
'Word-Usage/Clarity': 'surface',
'Precision': 'content',
'General Content Development,Conventions/Grammar/Spelling': 'content',
'Organization': 'surface',
'Evidence': 'content',
'Rebuttal/Reservation': 'content',
'Claims/Ideas': 'content',
'General Content Development': 'content'}
merged_purpose = []
for i in range(len(new_purpose)):
old_p = old_purpose[i]
new_p = new_purpose[i]
if type(old_p)==float and np.isnan(old_p):
if type(new_p)==float and np.isnan(new_p):
merged_purpose.append(new_p)
else:
merged_purpose.append(purpose_mapper[new_p])
else:
merged_purpose.append(purpose_mapper[old_p])
reason_evidence_list = [np.NaN]*len(merged_purpose)
fine_grain_label_list = [np.NaN]*len(merged_purpose)
# reason data
selected_reason_data = reason_data[reason_data["ID"]==essay_id]
selected_reason_data = selected_reason_data.to_dict('records')
for i in range(len(selected_reason_data)):
tmp_idx1 = selected_reason_data[i]["index1"]
tmp_idx2 = selected_reason_data[i]["index2"]
if tmp_idx1 == -1:
tmp_idx = new_sent_index.index(tmp_idx2-1)
else:
tmp_idx = old_sent_index.index(tmp_idx1-1)
fine_label = selected_reason_data[i]["Label1"].replace("-","").replace("Repeat","already exists").replace("EVIDENCE","").replace("REASONING","").replace("NonTextBased","not text based").lower().replace("lce", "LCE").strip()
this_label = "reasoning"
reason_evidence_list[tmp_idx] = this_label
fine_grain_label_list[tmp_idx] = fine_label
# evidence data
selected_evidence_data = evidence_data[evidence_data["ID"] == essay_id]
selected_evidence_data = selected_evidence_data.to_dict('records')
for i in range(len(selected_evidence_data)):
tmp_idx1 = selected_evidence_data[i]["index1"]
tmp_idx2 = selected_evidence_data[i]["index2"]
if tmp_idx1 == -1:
tmp_idx = new_sent_index.index(tmp_idx2 - 1)
else:
tmp_idx = old_sent_index.index(tmp_idx1 - 1)
fine_label = selected_evidence_data[i]["Label1"].replace("-","").replace("Repeat","already exists").replace("EVIDENCE","").replace("REASONING","").replace("NonTextBased","not text based").lower().replace("lce", "LCE").strip()
this_label = "evidence"
if "nontextbased" in fine_label:
print()
reason_evidence_list[tmp_idx] = this_label
fine_grain_label_list[tmp_idx] = fine_label
df = pd.DataFrame({"old_sentence_ids": old_sent_index,
"old_sentence_aligned_ids": old_gold_align,
"old_sentences": old_draft_list,
"new_sentences": new_draft_list,
"new_sentence_ids": new_sent_index,
"new_sentence_aligned_ids": new_gold_align,
"coarse_labels": merged_purpose,
"fine_labels": reason_evidence_list,
"purpose_labels": fine_grain_label_list,
})
if "Annotation_eagerstudy_esl40" in path:
print()
df = add_disirable_label(df, is_college=True)
print(df)
df.to_excel(f"./data_clean/revision_purpose/college/{essay_name}.xlsx", index=False)