mp_ids=['mp-1','mp-42'] # the mp_id list
with MPRester(MP_API_KEY) as mpr:
chgcars = mpr.get_charge_density_from_material_id(mp_ids) # avoiding looping the mp_id list
the above code works fine ,but how can i get the chgcar file for ‘mp-1’ and ‘mp-42’ individually?
Hi @bigtom4coding ,
The get_charge_density_from_material_id function isn’t actually intended to be used with multiple MPIDs input as a list, it’s meant to get a single CHGCAR entry for a given MPID. If you look at the structure attribute of the CHGCAR you pull from MP, it will only be one entry (whichever is newer). That’s a bug (I think).
That being said, if you want to pull CHGCARs from MP and save them either as gzipped JSON (maybe more natural for pymatgen objects) or as numpy arrays, here’s how you can do it.
Try to avoid doing this with many MPIDs (let’s say more than ~50). If you need more CHGCAR data on MP, it might be best to reach out directly to @tschaume or @munrojm.
from mp_api.client import MPRester
import numpy as np
mpids = ["mp-1","mp-42"]
with MPRester() as mpr:
for mpid in mpids:
# not recommended for many MPIDs
chgcars = mpr.get_charge_density_from_material_id(mpid)
# Saving as gzipped JSON
chgcars.save(f"chgcar_{mpid}.json.gz")
# Saving as numpy array
chgcar_as_array = np.array([chgcars.data[k] for k in ["total","diff"] if chgcars.data.get(k) is not None])
np.save(f"chgcar_{mpid}.npy")
I would like to say thank you to @Aaron_Kaplan and @bigtom4coding ,
for your Q&A indirectly solved my query that mp-api does not work for acquiring chgcar.
If one sees the following error:
[... skipping hidden 1 frame]
/usr/local/lib/python3.10/dist-packages/emmet/core/tasks.py in model_post_init(self, _TaskDoc__context)
451 # To determine task and run type, we search for input sets in this order
452 # of precedence: calcs_reversed, inputs, orig_inputs
--> 453 for inp_set in [self.calcs_reversed[0].input, self.input, self.orig_inputs]:
454 if inp_set is not None:
455 break
TypeError: 'NoneType' object is not subscriptable
It is because emmet-core 0.84.1 has a bug. The solution is:
pip install mp-api
pip install emmet-core==0.83.0
#pip install numpy==1.26.4 # numpy 2.0 does not work either
Hi @sc303165, this may be more of an issue with older tasks in MP (we’re working to make sure they are all in the same format), but since the line of code this fails on is from a PR I made, I’ll look into it!
Cheers!
@Aaron_Kaplan Thank you very much ! You answer is very helpful！
@sc303165 Thank you as well for another piece of valuable information!
I encountered the same “NoneType” issue! Your solution works! The emmet-core version was 0.84.2rc2. I downgraded the emmet-core to 0.83.0. It works now! Thanks!
Hi @Aaron_Kaplan, sorry to interrupt again.
I got some relevant questions about your code, and I’d be appreciated if you can elaborate on that.
chgcar_as_array = np.array([chgcars.data[k] for k in ["total","diff"] if chgcars.data.get(k) is not None])
1.What do “total” and “diff” stand for? does “total” stand for CHG file and ‘diff’ for CHGCAR file?
2. Do entries with ‘total’ available still have data for ‘diff’. In other word, do one entry have these two kinds of data available simultaneously？
3. Where can I find relevant documentation？I tried, but failed.
Check out the VASP manual entry for CHG/CHGCAR and let me know if that’s still unclear
@Aaron_Kaplan Thanks for the reference! I read the manual and did a bit of googling ,here’ s what i found:
You are correct about 1. The CHG file is identical to CHGCAR, except the numbers are written to lower precision (some other data blocks needed to restart a calculation are also only present in CHGCAR). But pymatgen.io.vasp.outputs.Chgcar can parse both CHGCAR and CHG files.
@Aaron_Kaplan I got it. thanks very much !!
This should be fixed in emmet-core==v0.84.2rc5 and later versions.