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

🐛 Write mods_info Parquet file again

Browse source
This commit is contained in:
Mike Gerber 2024-11-28 18:27:39 +01:00
parent abb20b8ba9
commit 11a04916f3
1 changed files with 12 additions and 9 deletions
Download patch file Download diff file Expand all files Collapse all files

19
src/mods4pandas/mods4pandas.py
View file

@ -399,19 +399,24 @@ def process(mets_files: List[str], output_file: str, output_page_info: str):
mets_files_real.append(m) mets_files_real.append(m)
# Prepare output files
# Process METS files with contextlib.suppress(FileNotFoundError):
os.remove(output_file)
output_file_sqlite3 = output_file + ".sqlite3" output_file_sqlite3 = output_file + ".sqlite3"
with contextlib.suppress(FileNotFoundError): with contextlib.suppress(FileNotFoundError):
os.remove(output_file_sqlite3) os.remove(output_file_sqlite3)
logger.info('Writing SQLite DB to {}'.format(output_file_sqlite3))
con = sqlite3.connect(output_file_sqlite3) con = sqlite3.connect(output_file_sqlite3)
if output_page_info: if output_page_info:
output_page_info_sqlite3 = output_page_info + ".sqlite3" output_page_info_sqlite3 = output_page_info + ".sqlite3"
logger.info('Writing SQLite DB to {}'.format(output_page_info_sqlite3))
with contextlib.suppress(FileNotFoundError): with contextlib.suppress(FileNotFoundError):
os.remove(output_page_info_sqlite3) os.remove(output_page_info_sqlite3)
con_page_info = sqlite3.connect(output_page_info_sqlite3) con_page_info = sqlite3.connect(output_page_info_sqlite3)
# Process METS files
with open(output_file + '.warnings.csv', 'w') as csvfile: with open(output_file + '.warnings.csv', 'w') as csvfile:
csvwriter = csv.writer(csvfile) csvwriter = csv.writer(csvfile)
mods_info = [] mods_info = []
@ -454,14 +459,12 @@ def process(mets_files: List[str], output_file: str, output_page_info: str):
except Exception as e: except Exception as e:
logger.exception('Exception in {}'.format(mets_file)) logger.exception('Exception in {}'.format(mets_file))
# Convert the mods_info List[Dict] to a pandas DataFrame # Convert the mods_info SQL to a pandas DataFrame
# TODO mods_info_df = pd.read_sql_query("SELECT * FROM mods_info", con, index_col="recordInfo_recordIdentifier")
# mods_info_df = dicts_to_df(mods_info, index_column="recordInfo_recordIdentifier")
# Save the DataFrame # Save the DataFrame
# TODO logger.info('Writing DataFrame to {}'.format(output_file))
#logger.info('Writing DataFrame to {}'.format(output_file)) mods_info_df.to_parquet(output_file)
#mods_info_df.to_parquet(output_file)
# Convert page_info # Convert page_info
# TODO # TODO