From 4ef559b2a0e08e4b3235d7a82c64d1d582b03497 Mon Sep 17 00:00:00 2001
From: Daniel Hornung <d.hornung@indiscale.com>
Date: Tue, 19 Mar 2024 13:14:20 +0100
Subject: [PATCH] WIP for fill_xlsx: More graceful handling of incomplete data.
---
.../table_json_conversion/fill_xlsx.py | 13 ++++++++++---
1 file changed, 10 insertions(+), 3 deletions(-)
diff --git a/src/caosadvancedtools/table_json_conversion/fill_xlsx.py b/src/caosadvancedtools/table_json_conversion/fill_xlsx.py
index c3859ac1..511c006d 100644
--- a/src/caosadvancedtools/table_json_conversion/fill_xlsx.py
+++ b/src/caosadvancedtools/table_json_conversion/fill_xlsx.py
@@ -27,6 +27,7 @@ import pathlib
from collections import OrderedDict
from types import SimpleNamespace
from typing import Any, Dict, List, Optional, TextIO, Union
+from warnings import warn
from jsonschema import FormatChecker, validate
from jsonschema.exceptions import ValidationError
@@ -134,8 +135,9 @@ def _read_or_dict(data: Union[dict, str, TextIO]) -> dict:
class TemplateFiller:
"""Class to fill XLSX templates. Has an index for all relevant columns."""
- def __init__(self, workbook: Workbook):
+ def __init__(self, workbook: Workbook, graceful: bool = False):
self._workbook = workbook
+ self._graceful = graceful
self._create_index()
@property
@@ -305,6 +307,7 @@ out: union[dict, None]
# collecting the data
assert isinstance(content, list)
+ content = [str(x) for x in content]
value = ";".join(content) # TODO we need escaping of values
path_str = p2s(path)
assert path_str not in insertables
@@ -318,6 +321,9 @@ out: union[dict, None]
insert_row = None
sheet = None
for path_str, value in insertables.items():
+ if self._graceful and path_str not in self._sheet_index:
+ warn(f"Ignoring path with missing sheet index: {path_str}")
+ continue
sheet_meta = self._sheet_index[path_str]
if sheet is None:
sheet = sheet_meta.sheet
@@ -359,7 +365,8 @@ result: str
Path for the result XLSX.
validation_schema: dict, optional
If given, validate the date against this schema first. This raises an exception if the validation
- fails.
+ fails. If no validation schema is given, try to ignore more errors in the data when filling the
+ XLSX template.
"""
data = _read_or_dict(data)
assert isinstance(data, dict)
@@ -377,7 +384,7 @@ validation_schema: dict, optional
# Filling the data
result_wb = load_workbook(template)
- template_filler = TemplateFiller(result_wb)
+ template_filler = TemplateFiller(result_wb, graceful=(validation_schema is None))
template_filler.fill_data(data=data)
parentpath = pathlib.Path(result).parent
--
GitLab