List of candidate terms has weird behavior, crashes or is too small
Testing 0.0.7.4.9
Summary
When creating a corpus from scratch :
-
There seems to be a regression in 0.0.7.4.9. some terms list do not load, other have weird behavior.
-
In general, the list of map terms and candidate terms is quite small. It was 350 before for map, now is seem more around 270. Also the initial list of candidates it even smaller (e.g. 124 for a corpora of 6k docs) while it should be much more than the map terms, between 2 and 4k terms.
The number of terms for candidates and maps should be hard coded somewhere because this is a preference of the instance. We set 350 for mapterms which is a good start and allow the user to review them quickly ; and few hundreds for candidates which allows to search for expression and add interesting terms by batches. I will make an issue about this.
Steps to reproduce
There is an example at https://dev.sub.gargantext.org/#/share/NodeList/189358
To reproduce the same corpora, make an API search on "water" in HAL by filtering with "IMT all".
What is the current bug behavior?
- I have created a corpus where the candidate terms list has a weird behavior when setting the per page to 10 : it is empty on page one, has 3 results on page 4, 7 on page 3, etc. ... Test on both FF and Chrome. I also had sometime the error
"SendResponseError (ResponseBodyError (ForeignError "JSON.parse: unexpected character at line 1 column 1 of the JSON data") (rf)"
This has been reproduced by others members of the team.
- Map and term list are too small.
What is the expected correct behavior?
-
Lists load well with the right number of item per page.
-
There should be an initial list of 350 map terms and, let's try 2k candidates (we have to check that this is OK for most front end computers).
Relevant logs and/or screenshots
