qurator-spk/modstool
1
0
Fork 0
mirror of https://github.com/qurator-spk/modstool.git synced 2025-11-04 11:24:14 +01:00
Code Issues Releases Wiki Activity

🐛 Fix converting/writing out per-page information (e.g. structure information)

Browse source
This commit is contained in:
Mike Gerber 2024-07-31 10:27:46 +02:00
parent 1bf86bfb4c
commit a1f333f4a4
2 changed files with 14 additions and 12 deletions
Download patch file Download diff file Expand all files Collapse all files

15
src/mods4pandas/mods4pandas.py
View file

@ -373,8 +373,8 @@ def pages_to_dict(mets, raise_errors=True) -> List[Dict]:
default='mods_info_df.parquet', show_default=True)
@click.option('--output-csv', type=click.Path(), help='Output CSV file')
@click.option('--output-xlsx', type=click.Path(), help='Output Excel .xlsx file')
@click.option('--page-info', is_flag=True, show_default=True, default=False, help='Save page info')
def process(mets_files: List[str], output_file: str, output_csv: str, output_xlsx: str, page_info: bool):
@click.option('--output-page-info', type=click.Path(), help='Save page info')
def process(mets_files: List[str], output_file: str, output_csv: str, output_xlsx: str, output_page_info: str):
"""
A tool to convert the MODS metadata in INPUT to a pandas DataFrame.
@ -420,11 +420,11 @@ def process(mets_files: List[str], output_file: str, output_csv: str, output_xls
d['mets_file'] = mets_file
# METS - per-page
if page_info:
if output_page_info:
page_info_doc: list[dict] = pages_to_dict(mets, raise_errors=True)
mods_info.append(d)
if page_info:
if output_page_info:
page_info.extend(page_info_doc)
if caught_warnings:
@ -450,12 +450,11 @@ def process(mets_files: List[str], output_file: str, output_csv: str, output_xls
mods_info_df.to_excel(output_xlsx)
# Convert page_info
# XXX hardcoded filenames + other formats
if page_info:
if output_page_info:
page_info_df = dicts_to_df(page_info, index_column=("ppn", "ID"))
# Save the DataFrame
logger.info('Writing DataFrame to {}'.format("page_info_df.parquet"))
page_info_df.to_parquet("page_info_df.parquet")
logger.info('Writing DataFrame to {}'.format(output_page_info))
page_info_df.to_parquet(output_page_info)
def main():