return {"feedback": True}
elif s == "n":
return {"feedback": False}
elif feedback_request["message"] == "give-examples":
attribute = feedback_request["attribute"]
sys.stdout.flush()
sys.stderr.flush()
print("Provide examples for '{}', leave empty to continue:".format(attribute.name))
examples = []
while True:
s = input("- ")
if s == "":
return {"examples": examples}
else:
examples.append(s)
else:
assert False, f"Unknown message '{feedback_request['message']}'!"
matching_phase(document_base, feedback_fn, status_function, statistics["matching"])
# store the result
with open("cache/matched-document-base.bson", "wb") as file:
file.write(document_base.to_bson())
with open("cache/statistics.json", "w") as file:
json.dump(statistics.to_serializable(), file, indent=4)
results["f1_score"] = (
2 * results["precision"] * results["recall"] /
(results["precision"] + results["recall"]))
# compute the results as the median
for attribute in dataset.ATTRIBUTES:
for score in ["recall", "precision", "f1_score"]:
values = [
res["results"][attribute][score]
for res in statistics["matching"]["runs"].all_values()
]
statistics["matching"]["results"][attribute][
score] = np.median(values)
################################################################################################################
# store the results
################################################################################################################
if not os.path.isdir(
os.path.join(os.path.dirname(__file__), "..", "results",
f"{dataset.NAME}")):
os.makedirs(
os.path.join(os.path.dirname(__file__), "..", "results",
f"{dataset.NAME}"),
exist_ok=True,
)
with open(
os.path.join(os.path.dirname(__file__), "results",
f"{dataset.NAME}", "aset-stanza-ranking.json"),
"w") as file:
json.dump(statistics.to_serializable(), file)