ENH: Improving labeling and the execution script #36
Conversation
heaven00
changed the title
| ''' | ||
| import re | ||
| import pandas as pd | ||
| import joblib |
There was a problem hiding this comment.
Add it to the requirements.txt
| if ('Actuals' in row['text'] or | ||
| 'Budget' in row['text'] or | ||
| 'Revised' in row['text'] or | ||
| 'Estimate' in row['text']): |
There was a problem hiding this comment.
Make a regex wherever you can, and add it on the top as a constant!
| # check capitalization of letters | ||
| if row.is_text and row.text.isupper() and pd.isnull(row.label): | ||
| if ('REVENUE EXPENDITURE' in row.text or 'DETAILED ACCOUNT' in | ||
| row.text or 'ABSTRACT ACCOUNT' in row.text or 'CAPITAL EXPENDITURE' |
There was a problem hiding this comment.
Make a regex wherever you can, and add it on the top as a constant!
| if row.is_text and row.text.isupper() and pd.isnull(row.label): | ||
| if ('REVENUE EXPENDITURE' in row.text or 'DETAILED ACCOUNT' in | ||
| row.text or 'ABSTRACT ACCOUNT' in row.text or 'CAPITAL EXPENDITURE' | ||
| in row.text or 'LOAN EXPENDITURE' in row.text): |
There was a problem hiding this comment.
Make a regex wherever you can, and add it on the top as a constant!
| row.text or 'ABSTRACT ACCOUNT' in row.text or 'CAPITAL EXPENDITURE' | ||
| in row.text or 'LOAN EXPENDITURE' in row.text): | ||
| row['label'] = 'title' | ||
| if 'demand no' in row.text.lower(): |
There was a problem hiding this comment.
Make a regex wherever you can, and add it on the top as a constant!
| def __init__(self, img, block_features, page_num, target_folder): | ||
| def __init__(self, img, block_features, page_num, target_folder, | ||
| default_headers): | ||
| self.img = img |
There was a problem hiding this comment.
Please explain all the init arguments
| ''' | ||
| Check which pdfs are already generated and remove them from the complete | ||
| list of pdfs. | ||
| ''' |
There was a problem hiding this comment.
Args and return type documentation is missing?
| default_headers): | ||
| '''Process a folder of demand draft pdfs and store the output in the output | ||
| folder. | ||
| ''' |
There was a problem hiding this comment.
Args and return type documentation is missing?
| vertical_ratio, | ||
| page_num, | ||
| pdf_file_path, | ||
| (25, 20), |
There was a problem hiding this comment.
Declare 25,20 as class variables and explain its rationale
| block_features.to_csv('{0}/{1}.csv'.format(features_log_folder, | ||
| page_num), index=False) | ||
| # Blank page check | ||
| if len(block_features.index) > 3: |
There was a problem hiding this comment.
Declare constant 3 as Class variable?
Labeling
Execution Script