Hi there, may I know how to customize a matminer.dataset which is not included in the list. I tried to edit the dataset_metadata.json, but I stuck at the ‘hash’ label. Could anyone provide some insights?
I tried to manually modifiy the utils.py, dataset_retrieval.py and dataset_metadata.json.
However, I still met an error message: "ValueError: Unrecognized dataset name: ", which is due to the missing definition in the global variable "_dataset_dict ".
I have no idea where the "_dataset_dict " is pre-defined?Does anyone know it?
Are you planning on contributing this for a PR to matminer or for other reasons?
If you are just doing it for convenience sake (i.e., not for a PR to matminer), I’m not sure customizing a matminer.dataset retrieval is your best option.
If you are planning on contributing it to matminer (which we appreciate!!), I recommend you open a PR there which I and @alex can help you troubleshoot.
Hi Alex,
Thank you very much for the instant reply!
Actually I was trying obtain and analyze a dataset that exists on the MPcontribs Portal. Here is the URL: https://portal.mpcontribs.org/2dmatpedia/.
However, I wasn’t able to find that dataset in Matminer datasets list.
In addition, I did have future plan contributing my dataset but somehow I need to compare part of my results with the available dataset that mentioned above.
Do you know how to retrieve that dataset through Matminer?
Thank you very much!
@ardunn and I are working on getting matminer linked up to the MPContribs API. I’m currently reorganizing the 2dmatpedia dataset. Once I’m done (~next week) and until we integrate with matminer, you can use the mpcontribs-client directly to retrieve the dataset. Please see the API Docs and the example notebooks for details. Please feel free to apply for a MPContribs dataset when you’re ready.
Ping this thread again in a few weeks if you are having trouble accessing the dataset through matminer and we can help you troubleshoot. As @tschaume said we are currently working on linking the two.