style: json에 교정문장 추가, score 포맷 수정

[#12]
boostcampaitech6 · Mar 24, 2024 · c8454b9 · c8454b9
1 parent 15359bf
commit c8454b9
Show file tree

Hide file tree

Showing 2 changed files with 11 additions and 5 deletions.
diff --git a/models/utils/gram_metrics.py b/models/utils/gram_metrics.py
@@ -1,6 +1,6 @@
 import string
-from typing import List, Dict, Literal, get_args
-from .gram_out_json import get_cleaned_token_list, get_scrs_tok
+from typing import Dict, Literal, get_args
+from out_json import get_cleaned_token_list, get_scrs_tok
 
 
 def get_error_count(
@@ -33,7 +33,7 @@ def get_error_rate_sen(
     error_count = get_error_count(checker_data=checker_data)
     sentence_count = len(og_list)
 
-    return error_count / sentence_count
+    return round(error_count / sentence_count, 2)
 
 
 def get_error_rate_word(
@@ -47,12 +47,14 @@ def get_error_rate_word(
         # remove punctuations
         new_sen = sen.translate(str.maketrans('', '', string.punctuation))
         word_count += len(new_sen.split(" "))
+        result = (1 - (error_count / word_count)) * 100
 
-    return 1 - (error_count / word_count)
+    return round(result, 2)
 
 # ec = error count, psc = per sentence count, pwc = per word count
 _TYPES = Literal["ec", "psc", "pwc"]
 
+
 def get_score(
         checker_data: Dict,
         score_type: _TYPES = "pwc",

diff --git a/models/utils/gram_out_json.py b/models/utils/gram_out_json.py
@@ -205,6 +205,7 @@ def get_phase_1_data(
 def get_phase_2_inner_data(
         sid: int, 
         sent: str, 
+        corr_sent: str,
         edited: bool,
         ref_word_list: List,
         tag_list: List,
@@ -213,6 +214,7 @@ def get_phase_2_inner_data(
     inner = {
         "sid": int,
         "sentence": str,
+        "corrected_sentence": str,
         "edited": False,
         "ref_word": [],
         "category": [],
@@ -221,6 +223,7 @@ def get_phase_2_inner_data(
     }
     inner["sid"] = sid
     inner["sentence"] = sent
+    inner["corrected_sentence"] = corr_sent
     inner["edited"] = edited
     inner["ref_word"] = ref_word_list
     inner["tag"] = tag_list
@@ -251,11 +254,12 @@ def get_phase_2_data(
         gector_dict = get_scrs_tok(inner_dict, ctl)
         sid = sentence_list.index(og_sent)
         sent = og_sent
+        corr_sent = inner_dict["fin_sentence"]
         edited = inner_dict["edited"]
         ref_word_list = gector_dict["og_word"]
         tag_list = gector_dict["full_tag"]
         if edited == True:
-            inner = get_phase_2_inner_data(sid=sid, sent=sent, edited=edited, ref_word_list=ref_word_list, 
+            inner = get_phase_2_inner_data(sid=sid, sent=sent, corr_sent=corr_sent, edited=edited, ref_word_list=ref_word_list, 
                                         tag_list=tag_list, tag_grammar=tag_grammar)
             tag_gram_dict["tag_grammar_info"].append(inner)