kensho-technologies / qwikidata Goto Github PK

Hello,

I am using your qwikidata for the last part of my thesis. The function return_sparql_query_results returns an exception for all queries since today afternoon. When I run the SPARQL from Wikidata Query service interface, it returns the results.
Could you check this function ?

Thank in advance.
Best Regards.

Hi, thank you for your so useful library.

I get incorrect results relying on WikidataItem.get_truthy_claim_groups method.
Let's take for example Kochi city from India:

`kochi = WikidataItem(get_entity_dict_from_api('Q1800'))

located_in = item.get_truthy_claim_groups()['P131']
`

P131 is the property of located in the administrative territorial entity.
As result we will have 3 entities, 2 of which have end time.
It means that P582 qualifier is not taken into consideration.
Perhaps it was not your intention, but if so the name get_truthy_claim_groups looks to be incorrect.

Thank you

The Time datavalue type has "precision", which affects how the time is interpreted or rendered.
The following example is taken from the comments in qwikidata/datavalue.py

That is, 1988-07-13T00:00:00 with
      precision 8 (decade) will be interpreted as 198?-??-?? and rendered as "1980s".

However, the Time class does not seem to have a method to render the time value in such way according to its precision. The __str__() method simply prints the precision, while the get_parsed_datetime_dict() method ignores the precision.

Am I missing anything? Thanks.

I sometimes come across the following error:

ValueError: required qualifier_dict keys are ['hash', 'snaktype', 'property', 'datavalue', 'datatype']. only found ['snaktype', 'property', 'hash', 'datatype']

This arises when running command get_truthy_claim_group, it seems that for some properties datavalue is not available as a dictionary key.

https://github.com/kensho-technologies/qwikidata/blob/develop/qwikidata/utils.py#L33

ent_hi will be undefined

There are some wikimedia services (e.g. https://tools.wmflabs.org/) that would benefit from Python 3.5 support. This would require,

• removing f-strings (e.g. f"the name of this thing is {name}")
• making inline typehints comments (e.g. thisint: int = 2 -> thisint = 2 # type: int)

When using the example code from https://github.com/kensho-technologies/qwikidata/tree/develop/examples/basic_json_dump.py (which appears to be the same as the docs), I get the following error messages:

File "test-qwikidata.py", line 56, in <module>
    dump_entities_to_json(politicians, out_fname)
File "path/to/py35/lib/python3.5/site-packages/qwikidata/utils.py", line 31, in dump_entities_to_json
    ent = next(entities, None)
TypeError: 'list' object is not an iterator

Used dump latest-all.json.bz2 downloaded 11/4/22 from here
Occurs in both python 3.5 and 3.10.

Is there a way around this?

Is there any way to get the claims for an item whithout specifiying the property id? I tried item.get_claim_group() but it gives me : TypeError: get_claim_group() missing 1 required positional argument: 'property_id'.

Thank you for the help.

Is there a need for a feature to filter on some condition when writing the chunks? For example, one might be interested only in geographical information or data on humans? This would substantially reduce the (filtered) dump.