jacob-mathison/doczyai-pipelines
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Activity

page number fixed

Browse Source
This commit is contained in:
Mayank Aamseek
2024-04-08 15:34:26 +05:30
parent 6962c1574b
commit 895493419a
2 changed files with 32 additions and 63660 deletions
Download Patch File Download Diff File Expand all files Collapse all files
streamlit/interface_3.py
+32 -7
View File
@@ -481,7 +481,8 @@ if st.session_state.user_info['mail'] in user_list:
# date_list.append(extracted_date)
# answer_list = date_list
try:
answer_list = [answer.strip("\n").strip().strip("{").strip("}").strip('"').rstrip('"') for answer in answer_list]
answer_list = [answer.strip("\n").strip().strip("{").strip("}").strip('"').rstrip('"').strip(
' ') if answer is not None else None for answer in answer_list]
if llm_selected in ['Llama 2 Chat 13B', 'Llama 2 Chat 70B']:
answer_list = [answer.rstrip(".") for answer in answer_list]
answer_list = [answer if "I don't know" not in str(answer) else " " for answer in answer_list]
@@ -492,8 +493,18 @@ if st.session_state.user_info['mail'] in user_list:
answer_list = [answer if "Not applicable" not in str(answer) else " " for answer in answer_list]
elif llm_selected in ['Claude 2', 'Claude Instant']:
answer_list = [answer if "do not have" not in str(answer) else " " for answer in answer_list]
answer_list = [answer if "do not see" not in str(answer) else " " for answer in answer_list]
answer_list = [answer if "does not specify" not in str(answer) else " " for answer in answer_list]
answer_list = [answer if "does not explicitly" not in str(answer) else " " for answer in answer_list]
answer_list = [answer if "N/A" not in str(answer) else " " for answer in answer_list]
answer_list = [answer if "don't know" not in str(answer) else " " for answer in answer_list]
answer_list = [answer if "do not see" not in str(answer) else " " for answer in answer_list]
answer_list = [answer if "Not specified" not in str(answer) else " " for answer in answer_list]
answer_list = [answer if "Don't know" not in str(answer) else " " for answer in answer_list]
answer_list = [answer if "don't see" not in str(answer) else " " for answer in answer_list]
answer_list = [answer if "don't have" not in str(answer) else " " for answer in answer_list]
answer_list = [answer if "Don't know" not in str(answer) else " " for answer in answer_list]
answer_list = [answer if "Don't know" not in str(answer) else " " for answer in answer_list]
answer_list = [answer.rstrip(".") for answer in answer_list]
else:
answer_list = [answer.rstrip(".") for answer in answer_list]
@@ -513,6 +524,8 @@ if st.session_state.user_info['mail'] in user_list:
df['Confidence Level'] = ' '
df['Snippet'] = snippet_list
df['New Page Number'] = page_no_list
df['New Page Number'] = df['New Page Number'].apply(lambda x: re.search(r'\d+', x).group(
) if isinstance(x, str) and re.search(r'\d+', x) is not None else " ")
df['Revised Prompt'] = [prompt] * len(contract_list_f)
df = pd.merge(df, fields[['Field Name', 'SF_DB_COL_NAME']], how ='left', on ='Field Name')
@@ -534,18 +547,30 @@ if st.session_state.user_info['mail'] in user_list:
field_values_2 = pd.concat([field_values_2, field_values_1], ignore_index = True)
df = pd.merge(df, field_values_2, how ='left', on =['Contract Name', 'SF_DB_COL_NAME'])
df['Original Page Number'] = df['Original Page Number'].apply(lambda x: re.search(r'\d+', x).group(
) if isinstance(x, str) and re.search(r'\d+', x) is not None else " ")
answer_list = list(df['New Extracted value'])
# if 'Date' in field:
# df['Actual Value Stored'] = pd.to_datetime(df['Actual Value Stored'],errors='coerce').dt.date
df_date = df[df['SF_DB_COL_NAME'].str.contains('_DT')]
df_others = df[~df['SF_DB_COL_NAME'].str.contains('_DT')]
df_date['Actual Value Stored'] = pd.to_datetime(df_date['Actual Value Stored'],errors='coerce').dt.strftime('%Y-%m-%d').fillna(" ")
df_date['New Extracted value'] = pd.to_datetime(df_date['New Extracted value'],errors='coerce').dt.strftime('%Y-%m-%d').fillna(" ")
df = pd.concat([df_date, df_others], ignore_index = True)
df.sort_values(['SF_DB_COL_NAME', 'Contract Name'], inplace=True)
df.fillna(" ", inplace=True)
df['Actual Value Stored'] = df['Actual Value Stored'].apply(lambda x: x.strip() if isinstance(x, str) else '')
df['New Extracted value'] = df['New Extracted value'].apply(lambda x: x.strip() if isinstance(x, str) else '')
actual_value_list = list(df['Actual Value Stored'])
result_list = [i==j for i, j in zip(actual_value_list, answer_list)]
answer_list = list(df['New Extracted value'])
actual_value_list = [s.replace('-', '').replace(' ', '').replace('[', '').replace(']', '').lower() for s in actual_value_list]
answer_list = [s.replace('-', '').replace(' ', '').replace('[', '').replace(']', '').lower() for s in answer_list]
result_list = [(i in j) or (j in i) if isinstance(i, str) and isinstance(
j, str) and ((i != '') == (j != '')) else False for i, j in zip(actual_value_list, answer_list)]
df['Result'] = [str(x) for x in result_list]
df = df[~df['Contract ID'].isnull()]
# if 'Original Page Number' not in df.columns:
# df['Original Page Number'] = ' '
df = df[['Contract Name','Contract ID', 'SF_DB_COL_NAME', 'Actual Value Stored','New Extracted value','Confidence Level'
,'Snippet','Original Page Number', 'New Page Number', 'Revised Prompt', 'Result']]
streamlit/results.csv
-63653
View File
File diff suppressed because it is too large Load Diff