Mislabelled information about "has bandstructure" column


#1

I wrote a short script to go through a list of the compound with “Has Bandstructure = True” to collect their bandstructure to see if they have a direct or indirect bandgap. The image location is typically stored in https://materialsproject.org/electronic_structure/bandstructure/plot/ followed by the materials id
e.g. https://materialsproject.org/electronic_structure/bandstructure/plot/mp-25708. for Bi2MoO6

However, I noticed that for some of these compound the bandstructure is not included. Below is the list of IDs:
mp-565435 mp-561113 mp-753594 mp-504878 mp-561766 mp-561823 mp-555769 mp-1069079 mp-561102 mp-567075 mp-655337

In particular, searching mp-565435 will give the material for Bi2MoO6 mp-25708. Same for searching mp-655337 which gives mp-24942. mp-24942 also suffers from the same problem as above, while mp-25708 has a bandstructure diagram.


#2

Hi @Vincent_Chung,

Please use the API to download band structures rather than scraping the image URL, which is not officially supported (@dwinston, please correct me if I’m wrong on this).

Regarding your specific issue, mp-ids refer to individual tasks (e.g. a relaxation calculation, a band structure calculation, etc.), of these tasks one mp-id is chosen to be the canonical mp-id for that material. That’s why mp-565435 (a specific task id) redirects to mp-25708 (the canonical mp-id for that material) on the website. You will only be able to retrieve the band structure for the canonical mp-id. If you use the API, this should just work.

Hope this helps,

Matt


#3

Hi @Vincent_Chung,

@mkhorton is correct that material detail pages fetch image urls after first resolving canonical material ids. However, it does seem that there is an issue with fetching some of these images. We are looking into this now. Thank you for posting about this.

Best,
Donny