Encyclopedia fixlist
Now that the encyclopedia is using the current production data, several issues have popped up:
-
Labels in DOS plot are sometimes messed up (1) (Due to invalid DOS normalization, now uses scientific notation.) -
Some functional types not correctly displayed in the Methodology overview box, although they show up properly when the methodology view is opened (1, 2) -
The electronic properties view is not displaying correctly when opened without first accessing the overview page (1) -
The band gap summary is not being displayed properly (1) -
When browsing through materials, the results of the previous material are not correctly cleared. -
Loading some band structures causes a server error, see e.g. calculation YDeISdnCy6L649OT5IKTrhGdVtgB and Uzy9_4hDSLrvDKBQV_QSYW7GIP8X (1, 2) These calculations are missing k-points and energies. Will be fixed by the newly introduced check for the existence of k-points and energies. Both FHI-aims and CASTEP seem to be sometimes opening empty section_k_bands
. -
Some functional type is not processed correctly and results in null
shown in the GUI (1) This was caused by elastic calculations for which the method information is not available. -
Loading the overview for some materials is causing an internal server error. (1, 2) This is caused by encountering data with old metainfo layout (section_encyclopedia vs section_metadata/encyclopedia). Should be fixed after successfull reprocess. -
Sorting on methodology page is not working correctly for all columns -
Searching for materials with both thermal properties and electronic properties is not returning any results -
Some 1D materials seem to have overlapping atoms. Investigate if problem in raw data or normalization. (1) See issue #294 (closed). The WIEN2k parser was doing things wrong, now fixed -
Some 1D materials look suspiciously 2D. Investigate if they are ribbons or 2D. (1). They are graphene nanoribbons and thus labeled correctly. -
The allowed maximum distance between atoms for them to be considered continuous seems too large. Because of this some materials are wrongly classified. Now lowered from 3.1 Å to 2.5 Å and added support for different radii definitions in MatID.(1, 2) -
The idealized cell of some 1D materials is wrapped incorrectly. (1) Now fixed in matid 0.6.1. -
Method overview not showing up for some 1D materials -
Some 2D materials are incorrectly labeled as bulk in the material list, but show up correctly in the overview. (1) This was caused by overlapping material ids between bulk and 2D entries and due to inconsistent selection of the representative calculations. Both now fixed. -
The chemical species search is not working correctly for 2D systems. E.g. "2D" and "C" is not bringing up graphene, which is found if you simply scroll through the 2D systems. The species and species counts were missing for 2D and 1D. Now added. -
The list of 2D systems includes multiple surface systems. Explore this mixup and remove the surface systems for now as we do not have a way to group them (create a material id) yet. These systems were incorrectly classified as 2D in the processing. With the newest processing pipeline they get correctly processed. I'm unable to locate which change did this, could have been a parser update, matid update or some other config update. -
Various 2D systems seem to have overlapping atoms. (1) These were all incorrectly labeled surface systems that were pushed through the 2D structure pipeline. -
If functional type is missing, the GUI incorrectly includes a "null" functional type in the statistics views. (1) -
The structure viewer is not being correctly loaded when entering into structure details the first time. -
The dropdown for selecting a page is not always showing the electronic structure page even if data is present (1) -
Basis set type search is not working.
Edited by Lauri Himanen