This commit is contained in:
Pedro Fonseca
2024-01-12 17:40:32 +00:00
4 changed files with 17 additions and 30 deletions

View File

@@ -5,6 +5,9 @@ import re
from dotenv import load_dotenv
from helper.constants import GPT_3_5_TURBO_INSTRUCT
from helper.token_counter import count_tokens
load_dotenv()
openai.api_key = os.getenv("OPENAI_API_KEY")
@@ -237,3 +240,14 @@ def extract_existing_sections_from_body(my_dict, keys_to_extract):
return list(filter(
lambda item: 'code' in item and item['code'] in keys_to_extract and 'grade' in item and 'name' in item,
my_dict['sections']))
def get_misspelled_pairs(text):
message = ('From the given text, extract the misspelled words and put them in the json with the correct word that '
'should be on the text instead. Sample JSON: '
'{"misspelled_words":[{"misspelled": "piza", "correction": "pizza"}] \n The text: "' + text + '"')
token_count = count_tokens(message)["n_tokens"]
response = make_openai_instruct_call(GPT_3_5_TURBO_INSTRUCT, message, token_count, ["misspelled_words"], 0.2)
# Filter out items with the same value for misspelled and correction
filtered_data = [item for item in response["misspelled_words"] if item['misspelled'] != item['correction']]
return filtered_data