Composition based features

Debasis_Sengupta · April 5, 2019, 1:54pm

Hi

I have gone through the tutorial in the following link

https://nbviewer.jupyter.org/github/hackingmaterials/matminer_examples/blob/master/matminer_examples/machine_learning-nb/bulk_modulus.ipynb

And it worked fine. Now I am trying to generate other composition based features. However, I have difficulty in following the what needs to passed as argument in featurize_dataframe other than “df”.

For example

from matminer.featurizers.composition import AtomicOrbitals
ao_feat = AtomicOrbitals()
ao_feat.featurize_dataframe(df,“composition”)

The above did not work (I used the data in the above link and followed the steps upto generating oxidation states)

I have tried a couple of other composition based featurizer, but haven’t been able to run them successfully. The documentation is perhaps very good for users with advanced knowledge of python class, but not quite easy to follow for me. For example, the following did not work either:

from matminer.featurizers.composition import ElectronAffinity
ea_feat = ElectronAffinity()
ea_feat.featurize_dataframe(df,“composition”)

An example for each of these feature generation would be really helpful for user with no so advanced knowledge of python.

Any help is really appreciated on how to generate those composition based features.

Thanks

Debasis

Logan_Ward · April 5, 2019, 2:36pm

Hello Debasis,

For each of these featurizer operations, the “featurize_dataframe” operation takes the dataframe “df” and the name of the column of that dataset (Pandas dataframes represent tables with rows and columns).

The error that I think might be happening here is that you did generate the composition with oxidation states, but are not calling “featurize_dataframe” with the column of the dataframe that contains the oxidized compositions (I think that defaults to “oxid_composition”).

One that that would help us debug your problem (now and in general), would be to paste the error that you are receiving. Generally, the last few line of the error message contains the information that is most relevant to us. Python returns that type of the error after telling you exactly where the program was (in terms of which function called by whichever other function). So, if you tell us the type of error and the last line in that “stack trace,” we can help you at lot easier.

Also, thanks for letting us know the docs were unclear for those without much Python experience. We’ll figure out how to revise them.

Best,

Logan

···

From: [email protected]
Sent: Friday, April 5, 2019 8:54 AM
To: matminer
Subject: Composition based features

Hi

I have gone through the tutorial in the following link

https://nbviewer.jupyter.org/github/hackingmaterials/matminer_examples/blob/master/matminer_examples/machine_learning-nb/bulk_modulus.ipynb

And it worked fine. Now I am trying to generate other composition based features. However, I have difficulty in following the what needs to passed as argument in featurize_dataframe other than “df”.

For example

from matminer.featurizers.composition import AtomicOrbitals
ao_feat = AtomicOrbitals()
ao_feat.featurize_dataframe(df,“composition”)

The above did not work (I used the data in the above link and followed the steps upto generating oxidation states)

I have tried a couple of other composition based featurizer, but haven’t been able to run them successfully. The documentation is perhaps very good for users with advanced knowledge of python class, but not quite easy to follow for me. For example, the following did not work either:

from matminer.featurizers.composition import ElectronAffinity
ea_feat = ElectronAffinity()
ea_feat.featurize_dataframe(df,“composition”)

An example for each of these feature generation would be really helpful for user with no so advanced knowledge of python.

Any help is really appreciated on how to generate those composition based features.

Thanks

Debasis

–
You received this message because you are subscribed to the Google Groups “matminer” group.
To unsubscribe from this group and stop receiving emails from it, send an email to [email protected].
For more options, visit https://groups.google.com/d/optout.

Debasis_Sengupta · April 5, 2019, 3:02pm

Hello Logan,

It seems that column name is “composition_oxid”. That is what df.columns prints.

In fact I used “composition_oxid” column for generating atomic orbital and electron affinity features. Following is the last few lines of errors. In addition, it also gives a warning at the beginning:

/global/homes/d/dxs/.local/cori/3.6-anaconda-4.4/lib/python3.6/site-packages/matminer/featurizers/composition.py:390: UserWarning: AtomicOrbitals: Al3+1 Co2+1 Co3+1 Si4-2 truncated to Al(CoSi)2

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
File “”, line 1, in
File “/global/homes/d/dxs/.local/cori/3.6-anaconda-4.4/lib/python3.6/site-packages/matminer/featurizers/base.py”, line 340, in featurize_dataframe
pbar=pbar)
File “/global/homes/d/dxs/.local/cori/3.6-anaconda-4.4/lib/python3.6/site-packages/matminer/featurizers/base.py”, line 467, in featurize_many
return p.map(func, entries, chunksize=self.chunksize)
File “/usr/common/software/python/3.6-anaconda-4.4/lib/python3.6/multiprocessing/pool.py”, line 266, in map
return self._map_async(func, iterable, mapstar, chunksize).get()
File “/usr/common/software/python/3.6-anaconda-4.4/lib/python3.6/multiprocessing/pool.py”, line 644, in get
raise self._value
TypeError: ‘NoneType’ object is not subscriptable
To skip errors when featurizing specific compounds, consider running the batch featurize() operation (e.g., featurize_many(), featurize_dataframe(), etc.) with ignore_errors=True

···

On Friday, April 5, 2019 at 10:37:01 AM UTC-4, Logan Ward wrote:

Hello Debasis,

For each of these featurizer operations, the “featurize_dataframe” operation takes the dataframe “df” and the name of the column of that dataset (Pandas dataframes represent tables with rows and columns).

The error that I think might be happening here is that you did generate the composition with oxidation states, but are not calling “featurize_dataframe” with the column of the dataframe that contains the oxidized compositions (I think that defaults to “oxid_composition”).

One that that would help us debug your problem (now and in general), would be to paste the error that you are receiving. Generally, the last few line of the error message contains the information that is most relevant to us. Python returns that type of the error after telling you exactly where the program was (in terms of which function called by whichever other function). So, if you tell us the type of error and the last line in that “stack trace,” we can help you at lot easier.

Also, thanks for letting us know the docs were unclear for those without much Python experience. We’ll figure out how to revise them.

Best,

Logan

From: [email protected]
Sent: Friday, April 5, 2019 8:54 AM
To: matminer
Subject: Composition based features

Hi

I have gone through the tutorial in the following link

https://nbviewer.jupyter.org/github/hackingmaterials/matminer_examples/blob/master/matminer_examples/machine_learning-nb/bulk_modulus.ipynb

And it worked fine. Now I am trying to generate other composition based features. However, I have difficulty in following the what needs to passed as argument in featurize_dataframe other than “df”.

For example

from matminer.featurizers.composition import AtomicOrbitals
ao_feat = AtomicOrbitals()
ao_feat.featurize_dataframe(df,“composition”)

The above did not work (I used the data in the above link and followed the steps upto generating oxidation states)

I have tried a couple of other composition based featurizer, but haven’t been able to run them successfully. The documentation is perhaps very good for users with advanced knowledge of python class, but not quite easy to follow for me. For example, the following did not work either:

from matminer.featurizers.composition import ElectronAffinity
ea_feat = ElectronAffinity()
ea_feat.featurize_dataframe(df,“composition”)

An example for each of these feature generation would be really helpful for user with no so advanced knowledge of python.

Any help is really appreciated on how to generate those composition based features.

Thanks

Debasis

–
You received this message because you are subscribed to the Google Groups “matminer” group.
To unsubscribe from this group and stop receiving emails from it, send an email to [email protected].
For more options, visit https://groups.google.com/d/optout.

Debasis_Sengupta · April 10, 2019, 1:59pm

Hello Logan,

Would you be able to see what mistake I am doing for computing atomic orbital. I have added the error (in my last message).

Thanks

Debasis

···

On Friday, April 5, 2019 at 11:02:37 AM UTC-4, [email protected] wrote:

Hello Logan,

It seems that column name is “composition_oxid”. That is what df.columns prints.

In fact I used “composition_oxid” column for generating atomic orbital and electron affinity features. Following is the last few lines of errors. In addition, it also gives a warning at the beginning:

/global/homes/d/dxs/.local/cori/3.6-anaconda-4.4/lib/python3.6/site-packages/matminer/featurizers/composition.py:390: UserWarning: AtomicOrbitals: Al3+1 Co2+1 Co3+1 Si4-2 truncated to Al(CoSi)2

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
File “”, line 1, in
File “/global/homes/d/dxs/.local/cori/3.6-anaconda-4.4/lib/python3.6/site-packages/matminer/featurizers/base.py”, line 340, in featurize_dataframe
pbar=pbar)
File “/global/homes/d/dxs/.local/cori/3.6-anaconda-4.4/lib/python3.6/site-packages/matminer/featurizers/base.py”, line 467, in featurize_many
return p.map(func, entries, chunksize=self.chunksize)
File “/usr/common/software/python/3.6-anaconda-4.4/lib/python3.6/multiprocessing/pool.py”, line 266, in map
return self._map_async(func, iterable, mapstar, chunksize).get()
File “/usr/common/software/python/3.6-anaconda-4.4/lib/python3.6/multiprocessing/pool.py”, line 644, in get
raise self._value
TypeError: ‘NoneType’ object is not subscriptable
To skip errors when featurizing specific compounds, consider running the batch featurize() operation (e.g., featurize_many(), featurize_dataframe(), etc.) with ignore_errors=True

On Friday, April 5, 2019 at 10:37:01 AM UTC-4, Logan Ward wrote:

Hello Debasis,

For each of these featurizer operations, the “featurize_dataframe” operation takes the dataframe “df” and the name of the column of that dataset (Pandas dataframes represent tables with rows and columns).

The error that I think might be happening here is that you did generate the composition with oxidation states, but are not calling “featurize_dataframe” with the column of the dataframe that contains the oxidized compositions (I think that defaults to “oxid_composition”).

One that that would help us debug your problem (now and in general), would be to paste the error that you are receiving. Generally, the last few line of the error message contains the information that is most relevant to us. Python returns that type of the error after telling you exactly where the program was (in terms of which function called by whichever other function). So, if you tell us the type of error and the last line in that “stack trace,” we can help you at lot easier.

Also, thanks for letting us know the docs were unclear for those without much Python experience. We’ll figure out how to revise them.

Best,

Logan

From: [email protected]
Sent: Friday, April 5, 2019 8:54 AM
To: matminer
Subject: Composition based features

Hi

I have gone through the tutorial in the following link

https://nbviewer.jupyter.org/github/hackingmaterials/matminer_examples/blob/master/matminer_examples/machine_learning-nb/bulk_modulus.ipynb

And it worked fine. Now I am trying to generate other composition based features. However, I have difficulty in following the what needs to passed as argument in featurize_dataframe other than “df”.

For example

from matminer.featurizers.composition import AtomicOrbitals
ao_feat = AtomicOrbitals()
ao_feat.featurize_dataframe(df,“composition”)

The above did not work (I used the data in the above link and followed the steps upto generating oxidation states)

I have tried a couple of other composition based featurizer, but haven’t been able to run them successfully. The documentation is perhaps very good for users with advanced knowledge of python class, but not quite easy to follow for me. For example, the following did not work either:

from matminer.featurizers.composition import ElectronAffinity
ea_feat = ElectronAffinity()
ea_feat.featurize_dataframe(df,“composition”)

An example for each of these feature generation would be really helpful for user with no so advanced knowledge of python.

Any help is really appreciated on how to generate those composition based features.

Thanks

Debasis

–
You received this message because you are subscribed to the Google Groups “matminer” group.
To unsubscribe from this group and stop receiving emails from it, send an email to [email protected].
For more options, visit https://groups.google.com/d/optout.

Logan_Ward · April 10, 2019, 2:21pm

I’m not sure what’s happening for this one.

Could you add “ao_feat.set_n_jobs(1)” after “ao_feat = AtomicOrbitals()” and re-run to get a clearer error message (multiprocessing tends to eat the most useful error messages)?

Logan

···

From: [email protected]
Sent: Wednesday, April 10, 2019 8:59 AM
To: matminer
Subject: Re: Composition based features

Hello Logan,

Would you be able to see what mistake I am doing for computing atomic orbital. I have added the error (in my last message).

Thanks

Debasis

On Friday, April 5, 2019 at 11:02:37 AM UTC-4, [email protected] wrote:

Hello Logan,

It seems that column name is “composition_oxid”. That is what df.columns prints.

In fact I used “composition_oxid” column for generating atomic orbital and electron affinity features. Following is the last few lines of errors. In addition, it also gives a warning at the beginning:

/global/homes/d/dxs/.local/cori/3.6-anaconda-4.4/lib/python3.6/site-packages/matminer/featurizers/composition.py:390: UserWarning: AtomicOrbitals: Al3+1 Co2+1 Co3+1 Si4-2 truncated to Al(CoSi)2

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
File “”, line 1, in
File “/global/homes/d/dxs/.local/cori/3.6-anaconda-4.4/lib/python3.6/site-packages/matminer/featurizers/base.py”, line 340, in featurize_dataframe
pbar=pbar)
File “/global/homes/d/dxs/.local/cori/3.6-anaconda-4.4/lib/python3.6/site-packages/matminer/featurizers/base.py”, line 467, in featurize_many
return p.map(func, entries, chunksize=self.chunksize)
File “/usr/common/software/python/3.6-anaconda-4.4/lib/python3.6/multiprocessing/pool.py”, line 266, in map
return self._map_async(func, iterable, mapstar, chunksize).get()
File “/usr/common/software/python/3.6-anaconda-4.4/lib/python3.6/multiprocessing/pool.py”, line 644, in get
raise self._value
TypeError: ‘NoneType’ object is not subscriptable
To skip errors when featurizing specific compounds, consider running the batch featurize() operation (e.g., featurize_many(), featurize_dataframe(), etc.) with ignore_errors=True

On Friday, April 5, 2019 at 10:37:01 AM UTC-4, Logan Ward wrote:

Hello Debasis,

For each of these featurizer operations, the “featurize_dataframe” operation takes the dataframe “df” and the name of the column of that dataset (Pandas dataframes represent tables with rows and columns).

The error that I think might be happening here is that you did generate the composition with oxidation states, but are not calling “featurize_dataframe” with the column of the dataframe that contains the oxidized compositions (I think that defaults to “oxid_composition”).

One that that would help us debug your problem (now and in general), would be to paste the error that you are receiving. Generally, the last few line of the error message contains the information that is most relevant to us. Python returns that type of the error after telling you exactly where the program was (in terms of which function called by whichever other function). So, if you tell us the type of error and the last line in that “stack trace,” we can help you at lot easier.

Also, thanks for letting us know the docs were unclear for those without much Python experience. We’ll figure out how to revise them.

Best,

Logan

From: [email protected]
Sent: Friday, April 5, 2019 8:54 AM
To: matminer
Subject: Composition based features

Hi

I have gone through the tutorial in the following link

https://nbviewer.jupyter.org/github/hackingmaterials/matminer_examples/blob/master/matminer_examples/machine_learning-nb/bulk_modulus.ipynb

And it worked fine. Now I am trying to generate other composition based features. However, I have difficulty in following the what needs to passed as argument in featurize_dataframe other than “df”.

For example

from matminer.featurizers.composition import AtomicOrbitals
ao_feat = AtomicOrbitals()
ao_feat.featurize_dataframe(df,“composition”)

The above did not work (I used the data in the above link and followed the steps upto generating oxidation states)

I have tried a couple of other composition based featurizer, but haven’t been able to run them successfully. The documentation is perhaps very good for users with advanced knowledge of python class, but not quite easy to follow for me. For example, the following did not work either:

from matminer.featurizers.composition import ElectronAffinity
ea_feat = ElectronAffinity()
ea_feat.featurize_dataframe(df,“composition”)

An example for each of these feature generation would be really helpful for user with no so advanced knowledge of python.

Any help is really appreciated on how to generate those composition based features.

Thanks

Debasis

–
You received this message because you are subscribed to the Google Groups “matminer” group.
To unsubscribe from this group and stop receiving emails from it, send an email to [email protected].
For more options, visit https://groups.google.com/d/optout.

–
You received this message because you are subscribed to the Google Groups “matminer” group.
To unsubscribe from this group and stop receiving emails from it, send an email to [email protected].
For more options, visit https://groups.google.com/d/optout.

Debasis_Sengupta · April 11, 2019, 9:08pm

Hi Logan,

I did as you recommended. Here is the script followed by the errors. Note that composition_oxid column has been generated before this following the tutorial.

from matminer.featurizers.composition import AtomicOrbitals
ao_feat = AtomicOrbitals()
ao_feat.set_n_jobs(1)
ao_feat.featurize_dataframe(df,“composition_oxid”)
Traceback (most recent call last):
File “”, line 1, in
File “/home/dxs/.local/lib/python3.6/site-packages/matminer/featurizers/base.py”, line 337, in featurize_dataframe
features = self.featurize_many(df[col_id].values,
File “/home/dxs/.local/lib/python3.6/site-packages/pandas/core/frame.py”, line 2934, in getitem
raise_missing=True)
File “/home/dxs/.local/lib/python3.6/site-packages/pandas/core/indexing.py”, line 1354, in _convert_to_indexer
return self._get_listlike_indexer(obj, axis, **kwargs)[1]
File “/home/dxs/.local/lib/python3.6/site-packages/pandas/core/indexing.py”, line 1161, in _get_listlike_indexer
raise_missing=raise_missing)
File “/home/dxs/.local/lib/python3.6/site-packages/pandas/core/indexing.py”, line 1246, in _validate_read_indexer
key=key, axis=self.obj._get_axis_name(axis)))
KeyError: “None of [Index([‘composition_oxid’], dtype=‘object’)] are in the [columns]”

···

On Wednesday, April 10, 2019 at 10:21:50 AM UTC-4, Logan Ward wrote:

I’m not sure what’s happening for this one.

Could you add “ao_feat.set_n_jobs(1)” after “ao_feat = AtomicOrbitals()” and re-run to get a clearer error message (multiprocessing tends to eat the most useful error messages)?

Logan

From: [email protected]
Sent: Wednesday, April 10, 2019 8:59 AM
To: matminer
Subject: Re: Composition based features

Hello Logan,

Would you be able to see what mistake I am doing for computing atomic orbital. I have added the error (in my last message).

Thanks

Debasis

On Friday, April 5, 2019 at 11:02:37 AM UTC-4, [email protected] wrote:

Hello Logan,

It seems that column name is “composition_oxid”. That is what df.columns prints.

In fact I used “composition_oxid” column for generating atomic orbital and electron affinity features. Following is the last few lines of errors. In addition, it also gives a warning at the beginning:

/global/homes/d/dxs/.local/cori/3.6-anaconda-4.4/lib/python3.6/site-packages/matminer/featurizers/composition.py:390: UserWarning: AtomicOrbitals: Al3+1 Co2+1 Co3+1 Si4-2 truncated to Al(CoSi)2

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
File “”, line 1, in
File “/global/homes/d/dxs/.local/cori/3.6-anaconda-4.4/lib/python3.6/site-packages/matminer/featurizers/base.py”, line 340, in featurize_dataframe
pbar=pbar)
File “/global/homes/d/dxs/.local/cori/3.6-anaconda-4.4/lib/python3.6/site-packages/matminer/featurizers/base.py”, line 467, in featurize_many
return p.map(func, entries, chunksize=self.chunksize)
File “/usr/common/software/python/3.6-anaconda-4.4/lib/python3.6/multiprocessing/pool.py”, line 266, in map
return self._map_async(func, iterable, mapstar, chunksize).get()
File “/usr/common/software/python/3.6-anaconda-4.4/lib/python3.6/multiprocessing/pool.py”, line 644, in get
raise self._value
TypeError: ‘NoneType’ object is not subscriptable
To skip errors when featurizing specific compounds, consider running the batch featurize() operation (e.g., featurize_many(), featurize_dataframe(), etc.) with ignore_errors=True

On Friday, April 5, 2019 at 10:37:01 AM UTC-4, Logan Ward wrote:

Hello Debasis,

For each of these featurizer operations, the “featurize_dataframe” operation takes the dataframe “df” and the name of the column of that dataset (Pandas dataframes represent tables with rows and columns).

The error that I think might be happening here is that you did generate the composition with oxidation states, but are not calling “featurize_dataframe” with the column of the dataframe that contains the oxidized compositions (I think that defaults to “oxid_composition”).

One that that would help us debug your problem (now and in general), would be to paste the error that you are receiving. Generally, the last few line of the error message contains the information that is most relevant to us. Python returns that type of the error after telling you exactly where the program was (in terms of which function called by whichever other function). So, if you tell us the type of error and the last line in that “stack trace,” we can help you at lot easier.

Also, thanks for letting us know the docs were unclear for those without much Python experience. We’ll figure out how to revise them.

Best,

Logan

From: [email protected]
Sent: Friday, April 5, 2019 8:54 AM
To: matminer
Subject: Composition based features

Hi

I have gone through the tutorial in the following link

https://nbviewer.jupyter.org/github/hackingmaterials/matminer_examples/blob/master/matminer_examples/machine_learning-nb/bulk_modulus.ipynb

And it worked fine. Now I am trying to generate other composition based features. However, I have difficulty in following the what needs to passed as argument in featurize_dataframe other than “df”.

For example

from matminer.featurizers.composition import AtomicOrbitals
ao_feat = AtomicOrbitals()
ao_feat.featurize_dataframe(df,“composition”)

The above did not work (I used the data in the above link and followed the steps upto generating oxidation states)

I have tried a couple of other composition based featurizer, but haven’t been able to run them successfully. The documentation is perhaps very good for users with advanced knowledge of python class, but not quite easy to follow for me. For example, the following did not work either:

from matminer.featurizers.composition import ElectronAffinity
ea_feat = ElectronAffinity()
ea_feat.featurize_dataframe(df,“composition”)

An example for each of these feature generation would be really helpful for user with no so advanced knowledge of python.

Any help is really appreciated on how to generate those composition based features.

Thanks

Debasis

–
You received this message because you are subscribed to the Google Groups “matminer” group.
To unsubscribe from this group and stop receiving emails from it, send an email to [email protected].
For more options, visit https://groups.google.com/d/optout.

Debasis_Sengupta · April 11, 2019, 9:22pm

Logan,

Ignore my immediate previous message (I forgot to generate “composition_oxid” before running AtomicOribitals). Following is the new error.

from matminer.featurizers.composition import AtomicOrbitals
ao_feat = AtomicOrbitals()
ao_feat.set_n_jobs(1)
ao_feat.featurize_dataframe(df,“composition_oxid”)
AtomicOrbitals: 0%| | 0/1181 [00:00<?, ?it/s]/home/dxs/.local/lib/python3.6/site-packages/matminer/featurizers/composition.py:390: UserWarning:

AtomicOrbitals: Al3+1 Co2+1 Co3+1 Si4-2 truncated to Al(CoSi)2

/home/dxs/.local/lib/python3.6/site-packages/matminer/featurizers/composition.py:390: UserWarning:

AtomicOrbitals: Sb3+1 Pt2-5 Pt2+1 Pt5+1 truncated to SbPt7

/home/dxs/.local/lib/python3.6/site-packages/matminer/featurizers/composition.py:390: UserWarning:

AtomicOrbitals: Sc2+1 Sc3+1 Al3+1 Si4-2 truncated to Sc2AlSi2

Traceback (most recent call last):
File “/home/dxs/.local/lib/python3.6/site-packages/matminer/featurizers/base.py”, line 493, in featurize_wrapper
return self.featurize(*x)
File “/home/dxs/.local/lib/python3.6/site-packages/matminer/featurizers/composition.py”, line 396, in featurize
feat[’{}_character’.format(edge)] = homo_lumo[edge][1][-1]
TypeError: ‘NoneType’ object is not subscriptable

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File “”, line 1, in
File “/home/dxs/.local/lib/python3.6/site-packages/matminer/featurizers/base.py”, line 340, in featurize_dataframe
pbar=pbar)
File “/home/dxs/.local/lib/python3.6/site-packages/matminer/featurizers/base.py”, line 450, in featurize_many
return_errors=return_errors) for x in entries]
File “/home/dxs/.local/lib/python3.6/site-packages/matminer/featurizers/base.py”, line 450, in
return_errors=return_errors) for x in entries]
File “/home/dxs/.local/lib/python3.6/site-packages/matminer/featurizers/base.py”, line 508, in featurize_wrapper
reraise(type(e), type(e)(msg), sys.exc_info()[2])
File “/usr/local/lib/python3.6/site-packages/six.py”, line 692, in reraise
raise value.with_traceback(tb)
File “/home/dxs/.local/lib/python3.6/site-packages/matminer/featurizers/base.py”, line 493, in featurize_wrapper
return self.featurize(*x)
File “/home/dxs/.local/lib/python3.6/site-packages/matminer/featurizers/composition.py”, line 396, in featurize
feat[’{}_character’.format(edge)] = homo_lumo[edge][1][-1]
TypeError: ‘NoneType’ object is not subscriptable
To skip errors when featurizing specific compounds, consider running the batch featurize() operation (e.g., featurize_many(), featurize_dataframe(), etc.) with ignore_errors=True

···

On Thursday, April 11, 2019 at 5:08:32 PM UTC-4, [email protected] wrote:

Hi Logan,

I did as you recommended. Here is the script followed by the errors. Note that composition_oxid column has been generated before this following the tutorial.

from matminer.featurizers.composition import AtomicOrbitals
ao_feat = AtomicOrbitals()
ao_feat.set_n_jobs(1)
ao_feat.featurize_dataframe(df,“composition_oxid”)
Traceback (most recent call last):
File “”, line 1, in
File “/home/dxs/.local/lib/python3.6/site-packages/matminer/featurizers/base.py”, line 337, in featurize_dataframe
features = self.featurize_many(df[col_id].values,
File “/home/dxs/.local/lib/python3.6/site-packages/pandas/core/frame.py”, line 2934, in getitem
raise_missing=True)
File “/home/dxs/.local/lib/python3.6/site-packages/pandas/core/indexing.py”, line 1354, in _convert_to_indexer
return self._get_listlike_indexer(obj, axis, **kwargs)[1]
File “/home/dxs/.local/lib/python3.6/site-packages/pandas/core/indexing.py”, line 1161, in _get_listlike_indexer
raise_missing=raise_missing)
File “/home/dxs/.local/lib/python3.6/site-packages/pandas/core/indexing.py”, line 1246, in _validate_read_indexer
key=key, axis=self.obj._get_axis_name(axis)))
KeyError: “None of [Index([‘composition_oxid’], dtype=‘object’)] are in the [columns]”

On Wednesday, April 10, 2019 at 10:21:50 AM UTC-4, Logan Ward wrote:

I’m not sure what’s happening for this one.

Could you add “ao_feat.set_n_jobs(1)” after “ao_feat = AtomicOrbitals()” and re-run to get a clearer error message (multiprocessing tends to eat the most useful error messages)?

Logan

From: [email protected]
Sent: Wednesday, April 10, 2019 8:59 AM
To: matminer
Subject: Re: Composition based features

Hello Logan,

Would you be able to see what mistake I am doing for computing atomic orbital. I have added the error (in my last message).

Thanks

Debasis

On Friday, April 5, 2019 at 11:02:37 AM UTC-4, [email protected] wrote:

Hello Logan,

It seems that column name is “composition_oxid”. That is what df.columns prints.

In fact I used “composition_oxid” column for generating atomic orbital and electron affinity features. Following is the last few lines of errors. In addition, it also gives a warning at the beginning:

/global/homes/d/dxs/.local/cori/3.6-anaconda-4.4/lib/python3.6/site-packages/matminer/featurizers/composition.py:390: UserWarning: AtomicOrbitals: Al3+1 Co2+1 Co3+1 Si4-2 truncated to Al(CoSi)2

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
File “”, line 1, in
File “/global/homes/d/dxs/.local/cori/3.6-anaconda-4.4/lib/python3.6/site-packages/matminer/featurizers/base.py”, line 340, in featurize_dataframe
pbar=pbar)
File “/global/homes/d/dxs/.local/cori/3.6-anaconda-4.4/lib/python3.6/site-packages/matminer/featurizers/base.py”, line 467, in featurize_many
return p.map(func, entries, chunksize=self.chunksize)
File “/usr/common/software/python/3.6-anaconda-4.4/lib/python3.6/multiprocessing/pool.py”, line 266, in map
return self._map_async(func, iterable, mapstar, chunksize).get()
File “/usr/common/software/python/3.6-anaconda-4.4/lib/python3.6/multiprocessing/pool.py”, line 644, in get
raise self._value
TypeError: ‘NoneType’ object is not subscriptable
To skip errors when featurizing specific compounds, consider running the batch featurize() operation (e.g., featurize_many(), featurize_dataframe(), etc.) with ignore_errors=True

On Friday, April 5, 2019 at 10:37:01 AM UTC-4, Logan Ward wrote:

Hello Debasis,

For each of these featurizer operations, the “featurize_dataframe” operation takes the dataframe “df” and the name of the column of that dataset (Pandas dataframes represent tables with rows and columns).

The error that I think might be happening here is that you did generate the composition with oxidation states, but are not calling “featurize_dataframe” with the column of the dataframe that contains the oxidized compositions (I think that defaults to “oxid_composition”).

One that that would help us debug your problem (now and in general), would be to paste the error that you are receiving. Generally, the last few line of the error message contains the information that is most relevant to us. Python returns that type of the error after telling you exactly where the program was (in terms of which function called by whichever other function). So, if you tell us the type of error and the last line in that “stack trace,” we can help you at lot easier.

Also, thanks for letting us know the docs were unclear for those without much Python experience. We’ll figure out how to revise them.

Best,

Logan

From: [email protected]
Sent: Friday, April 5, 2019 8:54 AM
To: matminer
Subject: Composition based features

Hi

I have gone through the tutorial in the following link

https://nbviewer.jupyter.org/github/hackingmaterials/matminer_examples/blob/master/matminer_examples/machine_learning-nb/bulk_modulus.ipynb

And it worked fine. Now I am trying to generate other composition based features. However, I have difficulty in following the what needs to passed as argument in featurize_dataframe other than “df”.

For example

from matminer.featurizers.composition import AtomicOrbitals
ao_feat = AtomicOrbitals()
ao_feat.featurize_dataframe(df,“composition”)

The above did not work (I used the data in the above link and followed the steps upto generating oxidation states)

I have tried a couple of other composition based featurizer, but haven’t been able to run them successfully. The documentation is perhaps very good for users with advanced knowledge of python class, but not quite easy to follow for me. For example, the following did not work either:

from matminer.featurizers.composition import ElectronAffinity
ea_feat = ElectronAffinity()
ea_feat.featurize_dataframe(df,“composition”)

An example for each of these feature generation would be really helpful for user with no so advanced knowledge of python.

Any help is really appreciated on how to generate those composition based features.

Thanks

Debasis

–
You received this message because you are subscribed to the Google Groups “matminer” group.
To unsubscribe from this group and stop receiving emails from it, send an email to [email protected].
For more options, visit https://groups.google.com/d/optout.