Closed GraemeWatt closed 4 months ago
I have found further related issues this week as we just discussed @ItIsJordan:
/data/<int:recid>/<int:data_recid>/<int:version>/<int:load_all>
route twice with load_all = 0
, each time reading a YAML data file from disk. This is inefficient.load_all = 0
) then the table appears after clicking "Load Table" with load_all = 1
. This is the correct behaviour. However, problems appear if other tables are loaded before clicking on the large data table. In this case, multiple requests to the route are made with load_all = 1
corresponding to different tables and often the wrong table is displayed together with overlapping display of the number of values like in issue 2. above. An example record is https://www.hepdata.net/record/ins2755637 where the first table is a large one.@ItIsJordan : another related bug just reported by a user. The JSON links for individual tables are currently broken because they link to /data/<int:recid>/<int:data_recid>/<int:version>
without the extra <int:load_all>
argument. This route should be reinstated with a default value of load_all = 1
if not specified explicitly. See converter/views.py for examples of functions with multiple routes.
I noticed a couple of issues with https://www.hepdata.net/record/ins2752118 where the second table is larger than the automatic loading threshold and its description contains LaTeX encoding. There seem to be a couple of bugs that were missed in the original testing of #733.
Normally, table descriptions containing LaTeX encoding are rendered using MathJax, but this does not happen before clicking "Load Table":
If other (smaller) tables are loaded before the second (larger) table, there is overlap between the display of the number of values:
Could you please fix these two bugs, @ItIsJordan?