These are chat archives for frictionlessdata/chat

28th
Jun 2017
Paul Walsh
@pwalsh
Jun 28 2017 04:04

hey @georgeslabreche_twitter . yes i think the test suite is out of date.

One note: In Java, there seems to be several ~mature libraries for JSON Schema ( https://github.com/java-json-tools/json-schema-validator https://github.com/everit-org/json-schema https://github.com/networknt/json-schema-validator ). As noted in the implementation notes, we use JSON Schema to do the validation of the descriptors in the Python and JavaScript libraries. It means less code for that part of the work. For example, in this commit ( frictionlessdata/datapackage-java@89ac7aa ) you are essentially writing the spec in Java, whereas using JSON Schema you dont need to write that code ( ref. http://specs.frictionlessdata.io/schemas/data-package.json https://github.com/frictionlessdata/datapackage-js/tree/master/src/profiles https://github.com/frictionlessdata/datapackage-js/blob/master/src/validate.js )

Georges L J Labrèche
@georgeslabreche_twitter
Jun 28 2017 08:09
Thank you for the resources, will check them out!
are there any updated test suites that can be used?
roll
@roll
Jun 28 2017 09:17
@georgeslabreche_twitter not yet. But I'm going to provide at least testing datasets pretty soon
Also please take a look on type tests - https://github.com/frictionlessdata/tableschema-py/tree/master/tests/types - it's a good mini testsuite for type casting behaviour which could be ported to different languages (updated to lang specifics of course)
roll
@roll
Jun 28 2017 09:24
@IsmiKin As Paul's said new report structure is a something to go forward with. We already use it as an interface between different systems in goodtables ecosystem and it should be considered as a stable API already (report structure). We're about to make official v1 release pretty soon
Ismael Ors
@IsmiKin
Jun 28 2017 09:40
@roll thank you, it's look great.
Now I'm using tableschema lib to infer schemas, but for some reason it always detect all rows as string. I tried to wrapper content between double quotes and without them in the .csv, but I still get all fields infered as string. Are there any constraints to allow the infer method work properly?
roll
@roll
Jun 28 2017 09:45
@IsmiKin could you please create a tableschema issue with this incorrect infer behavior example?
Ismael Ors
@IsmiKin
Jun 28 2017 09:45
yes, I guess I'm doing something wrong., When I run the cli infer method it works
Gabriela Rodríguez Berón
@gabelula
Jun 28 2017 21:03
hey! I'm trying to use http://goodtables.io/ for other repository but I do not see it in the repos when I sync sources