def save_detector_settings_csv(
self, detectors_by_value: Dict[str, List[str]]) -> None:
detector = DocumentFieldMultilineRegexDetector()
detector.document_field = self.document_field
df = pd.DataFrame(columns=['value', 'pattern'])
df.set_index("pattern", inplace=True)
for field_val in detectors_by_value:
for include_reg_value in detectors_by_value[field_val]:
df = df.append(
{
'value': field_val,
'pattern': include_reg_value
},
ignore_index=True)
df.drop_duplicates(subset='pattern', inplace=True)
try:
existing = DocumentFieldMultilineRegexDetector.objects.get(
document_field_id=self.document_field.uid
) # type: DocumentFieldMultilineRegexDetector
except DocumentFieldMultilineRegexDetector.DoesNotExist:
detector.csv_content = df.to_csv()
detector.update_checksum()
detector.save()
return
# just update CSV content and hashsum
if self.drop_previous_field_detectors:
existing.csv_content = df.to_csv()
existing.update_checksum()
existing.save()
return
# join these options with existing one
# overwriting duplicates by detected_value or regexp pattern
existing.combine_with_dataframe(df)
existing.save()
def save_detector_settings(cls,
detector: DocumentFieldMultilineRegexDetector):
detector.save()