Importing ecospold1 processes exported from openLCA #126
Comments
|
hi @sc-gcoste , did you find a way to import agribalyse? I am facing the same issue... thanks! |
|
Hi @renaud, unfortunately no... |
proposed fix for brightway-lca#126
|
same problem here |
1 similar comment
|
same problem here |
|
same here |
|
I experienced a similar issue, but not with agribalyse. Would recommend looking through your XML files for any apparent issues from OpenLCA. In my case, some tags were not linked and there were a few empty XML files and tweaking functions ``
@classmethod
def is_valid_ecospold1(cls, dataset):
try:
ref_func = dataset.metaInformation.processInformation.referenceFunction
name = ref_func.get("name").strip()
unit = ref_func.get("unit")
categories = [ref_func.get("category"), ref_func.get("subCategory")]
code = int(dataset.get("number"))
location = dataset.metaInformation.processInformation.technology.get("text")
technology = dataset.metaInformation.processInformation.technology.get("text")
# time_period = getattr2(dataset.metaInformation.processInformation, "timePeriod").get("text")
production_volume = getattr2(dataset.metaInformation.modellingAndValidation, "representativeness").get("productionVolume")
# sampling = getattr2(dataset.metaInformation.modellingAndValidation, "representativeness").get("samplingProcedure"),
# extrapolations = getattr2(dataset.metaInformation.modellingAndValidation, "representativeness").get("extrapolations")
# uncertainty = getattr2(dataset.metaInformation.modellingAndValidation, "representativeness").get("uncertaintyAdjustments")
# Checking exchanges
for exc in dataset.flowData.iterchildren():
if exc.tag == "comment":
continue
if exc.tag in ("{http://www.EcoInvent.org/EcoSpold01}exchange", "exchange"):
if hasattr(exc, "outputGroup"):
if exc.outputGroup.text in {"0", "2", "3"}:
pass
elif exc.outputGroup.text == "1":
pass
elif exc.outputGroup.text == "4":
pass
else:
raise ValueError(
"Can't understand output group {}".format(exc.outputGroup.text)
)
else:
if exc.inputGroup.text in {"1", "2", "3", "5"}:
kind = "technosphere"
elif exc.inputGroup.text == "4":
kind = "biosphere" # Resources
else:
raise ValueError(
"Can't understand input group {}".format(exc.inputGroup.text)
)
elif exc.tag in (
"{http://www.EcoInvent.org/EcoSpold01}allocation",
"allocation",
):
reference = int(exc.get("referenceToCoProduct")),
fraction = float(exc.get("fraction")),
exchanges = [int(c.text) for c in exc.iterchildren() if c.tag != "comment"],
else:
raise ValueError("Flow data type %s not understood" % exc.tag)
return True
except Exception as e:
print(f"Error message: {e}")
return False
# except AttributeError:
# return False
@classmethod
def process_dataset(cls, dataset, filename, db_name):
ref_func = dataset.metaInformation.processInformation.referenceFunction
def get_comment():
try:
comments = [
ref_func.get("generalComment"),
ref_func.get("includedProcesses"),
(
"Location: ",
dataset.metaInformation.processInformation.geography.get("text"),
),
(
"Technology: ",
dataset.metaInformation.processInformation.technology.get("text"),
),
(
"Time period: ",
getattr2(dataset.metaInformation.processInformation, "timePeriod").get(
"text"
),
),
(
"Production volume: ",
getattr2(
dataset.metaInformation.modellingAndValidation, "representativeness"
).get("productionVolume"),
),
(
"Sampling: ",
getattr2(
dataset.metaInformation.modellingAndValidation, "representativeness"
).get("samplingProcedure"),
),
(
"Extrapolations: ",
getattr2(
dataset.metaInformation.modellingAndValidation, "representativeness"
).get("extrapolations"),
),
(
"Uncertainty: ",
getattr2(
dataset.metaInformation.modellingAndValidation, "representativeness"
).get("uncertaintyAdjustments"),
),
]
comment = "\n".join(
[
(" ".join(x) if isinstance(x, tuple) else x)
for x in comments
if (x[1] if isinstance(x, tuple) else x)
]
)
return comment
except:
return []
def get_authors():
try:
ai = dataset.metaInformation.administrativeInformation
data_entry = []
for elem in ai.iterchildren():
if "dataEntryBy" in elem.tag:
data_entry.append(elem.get("person"))
fields = [
("address", "address"),
("company", "companyCode"),
("country", "countryCode"),
("email", "email"),
("name", "name"),
]
authors = []
for elem in ai.iterchildren():
if "person" in elem.tag and elem.get("number") in data_entry:
authors.append({label: elem.get(code) for label, code in fields})
return authors
except:
return []
data = {
"categories": [ref_func.get("category"), ref_func.get("subCategory")],
"code": int(dataset.get("number")),
"comment": get_comment(),
"authors": get_authors(),
"database": db_name,
"exchanges": cls.process_exchanges(dataset),
"filename": filename,
"location": dataset.metaInformation.processInformation.geography.get(
"location"
),
"name": ref_func.get("name").strip(),
"type": "process",
"unit": ref_func.get("unit"),
}
try:
allocation_exchanges = [
exc for exc in data["exchanges"] if exc.get("reference")
]
except:
allocation_exchanges = []
if allocation_exchanges != []:
data["allocations"] = allocation_exchanges
data["exchanges"] = [exc for exc in data["exchanges"] if exc.get("type")]
return data
``Hope this helps! |
|
Dear everyone, apologies for not seeing this or responding earlier. We just merged new import xmlschema
import pyecospold
from pathlib import Path
xsd = Path(pyecospold.__file__).parent / "schemas" / "v1" / "EcoSpold01Dataset.xsd"
def get_validation_errors(xml_file: Path, xsd_file: Path):
schema = xmlschema.XMLSchema(xsd_file)
validation_error_iterator = schema.iter_errors(open(xml_file).read())
for idx, validation_error in enumerate(validation_error_iterator, start=1):
print(f'[{idx}]\n\tpath: {validation_error.path}\n\treason: {validation_error.reason}')
get_validation_errors(
"process_000f29c8-0b4b-32f7-96f7-e0f29530d2fb.xml",
xsd
)Gives the following errors:
My inclination is to not support files which are very invalid - it would mean writing much more complicated code and would also make testing quite difficult. Note that openLCA is the only ones publishing invalid ecospold 1/2 files - even the big boys do it sometimes. However, we can make adjustments to the schema if there is a good reason. You can find the @tngTUDOR @jsvgoncalves FYI and feel free to express your opinion. |
|
It is not so visible in the user interface, but in the EcoSpold 1 export wizard in openLCA there is a second page when you click
When I import the attached example dataset above and export it again with this option, it will generate a default start- and end-date: <timePeriod dataValidForEntirePeriod="true" text="Unspecified">
<startDate>9999-01-01+01:00</startDate>
<endDate>9999-12-31+01:00</endDate>
</timePeriod>and also a default person which is then linked as data generator etc.: <administrativeInformation>
<dataEntryBy person="1"/>
<dataGeneratorAndPublication
person="1"
dataPublishedIn="0"
copyright="true"
accessRestrictedTo="0"/>
<person
number="1"
name="default"
address="Created for EcoSpold 1 compatibility"
telephone="000"
companyCode="default"
countryCode="CH"/>
</administrativeInformation>edit: I think the dataset is then valid against the updated schema of |
|
Thanks a lot @msrocka! It might make sense to have that field checked by default - I am not sure what the specific business stories are to emit data which doesn't validate against the schema, but probably the default should be a valid file, even if some data is not usable. |
Agreed, would not try to fix very invalid files. But maybe we could try improving the error/exception information to make it a bit more obvious that the file is very invalid. |
I have an ecospold1 dataset extracted from openLCA and I would like to import it into Brightway2. Using the SingleOutputEcospold1Importer should read the ecospold files but apparently something is wrong in the file schema.
Code:
Output:
Process from Agribalyse3 imported to EcoSpold1 with openLCA (to unzip and place in the directory searched by the importer)
process_000f29c8-0b4b-32f7-96f7-e0f29530d2fb.zip
NB: When using
use_mp=TrueI get multipleMultiprocessingErrorinviting to rerun withuse_mp=False.The text was updated successfully, but these errors were encountered: