Date parsing/conversion

[koeaw]

We need to introduce a date parsing mechanism to our project because some incoming date data does not (have to) follow strict formatting rules.

Given that the current DateParser.py code is old and convoluted – see also PR #125, which sadly didn't make it into the code base – should we continue to use this or is there a better solution?

We don't need very fancy date parsing, meaning we're currently only dealing with dates from the 20th century to now and incoming dates won't contain vague textual descriptions (e.g. "approx. 20-30 years B.C.") but will be number-based and use date formats/variations commonly used in AT (+ maybe ISO format). Examples: 1.5.1970, 02.04.85, 23/12/1999, 2024-01-01.

[b1rger]

I guess the date should be stored in a field and the original input string can be thrown away? You can use DateParser.py, but you might have better luck with dateparser

The advantage of DateParser.py is, that it calculates timespans of date (i.e. if you only know the year of a date), but that does not seem to be needed in this case

[koeaw]

I guess the date should be stored in a field and the original input string can be thrown away?

Almost! We're planning to keep the original input strings around and editable by content managers and further updates/corrections to them should get propagated to the (non-editable) field for the streamlined date.

Hmm, we don't need to handle time spans per se, but ah, should have clarified above that we indeed will also have to deal with year-only dates. I'll look at dateparser to see if it can handle incomplete dates as well, thank you!

[koeaw]

I'll look at dateparser to see if it can handle incomplete dates as well, thank you!

Yup, seems to be the case: https://dateparser.readthedocs.io/en/latest/introduction.html#incomplete-dates

[koeaw]

Closing this as the question was answered. Thx, @b1rger!