Creating archives from inside another archive's normalizer
Hey Nomad,
these days I am using a workaround for creating Archives inside the normalizer of another archive. The usecase is, I have a batch, which contains samples and i need to create the batch and automatically create the samles. I know the code is not pretty, it is just a placeholder for something better.
```python
class HySprint_Batch(BatchWithID, EntryData):
m_def = Section(
a_eln=dict(hide=['lab_id', 'author', 'user']),
a_template=dict(institute="HZB_Hysprint"))
number_of_new_substrates = Quantity(
type=np.dtype(np.int64),
description='The number of substrates in the batch.',
default=0,
a_eln=dict(component='NumberEditQuantity')
)
number_of_substrates = Quantity(
type=np.dtype(np.int64),
description='The number of substrates in the batch.',
default=0
)
samples = Quantity(
type=Reference(HySprint_Sample.m_def),
shape=['*'],
descriptions='The samples in the batch.',
# a_eln=dict(component='ReferenceEditQuantity')
)
create_samples = Quantity(
type=bool,
default=False,
a_eln=dict(component='BoolEditQuantity')
)
substrate = Quantity(
type=Reference(HySprint_Substrate.m_def),
a_eln=dict(component='ReferenceEditQuantity'))
def normalize(self, archive, logger):
super(HySprint_Batch, self).normalize(archive, logger)
import json
from nomad.utils import hash
upload_id = self.m_root().metadata.upload_id
if self.number_of_new_substrates > 0 and self.create_samples and self.batch_id:
sample_refs = []
for idx in range(
self.number_of_substrates,
self.number_of_substrates + self.number_of_new_substrates):
# For each sample number, create instance of sample.
sample = HySprint_Sample(
institute=self.institute if self.institute else '',
sample_owner=self.batch_owner if self.batch_owner else '',
creation_datetime=self.creation_datetime if self.creation_datetime else '',
sample_short_name=self.batch_short_name + "_" + str(idx) if self.batch_short_name else str(idx),
substrate=self.substrate)
sample_entry = sample.m_to_dict(with_root_def=True)
with archive.m_context.raw_file(self.batch_id + '_' + str(idx) + ".archive.json", 'w') as outfile:
json.dump({"data": sample_entry}, outfile)
entry_id = hash(upload_id, os.path.basename(outfile.name))
print(entry_id)
sample_refs.append(
f'../uploads/{upload_id}/archive/{entry_id}#data')
self.number_of_substrates += self.number_of_new_substrates
self.number_of_new_substrates = 0
self.samples = sample_refs
self.create_samples = False
```
As you can see I convert the class ```Hyprint_Sample``` to dict, hash it and append a reference to it. If I reprocess the upload the samples are picked up.
There are some aspects here:
1. Are you planning on having sth like this which is easier to use then this hack?
2. Will it be possible to process the archive when created automatically?
3. Is it related to the ability to create an empty archive?
Best Micha
issue