D-Net
/
dnet-hadoop
14
5
Fork 11
Code Issues 14 Pull Requests 13 Projects 2 Releases 2 Wiki Activity

Add filtering based on citation source

This commit is contained in:
Ilias Kanellos 2023-03-20 15:38:36 +02:00
parent 720fd19b39
commit b5c252865c
1 changed files with 9 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

11
dhp-workflows/dhp-impact-indicators/src/main/resources/create_openaire_ranking_graph.py
View File

@ -126,12 +126,19 @@ oa_objects_df = oa_objects_df.drop('deletedbyinference').drop('invisible').disti
# Collect only valid citations i.e., invisible = false & deletedbyinference=false # Collect only valid citations i.e., invisible = false & deletedbyinference=false
cites_df = spark.read.json(graph_folder + "/relation")\ cites_df = spark.read.json(graph_folder + "/relation")\
.select(F.col('source').alias('citing'), F.col('target').alias('cited'), 'relClass', 'dataInfo.deletedbyinference', 'dataInfo.invisible')\ .select(F.col('source').alias('citing'), F.col('target').alias('cited'), 'collectedfrom.value', 'relClass', 'dataInfo.deletedbyinference', 'dataInfo.invisible')\
.where( (F.col('relClass') == "Cites") \ .where( (F.col('relClass') == "Cites") \
& (F.col('dataInfo.deletedbyinference') == "false")\ & (F.col('dataInfo.deletedbyinference') == "false")\
& (F.col('dataInfo.invisible') == "false"))\ & (F.col('dataInfo.invisible') == "false"))\
.drop('dataInfo.deletedbyinference').drop('dataInfo.invisible')\ .drop('dataInfo.deletedbyinference').drop('dataInfo.invisible')\
.repartition(num_partitions, 'citing').drop('relClass') .repartition(num_partitions, 'citing').drop('relClass')\
.withColumn('collected_lower', F.expr('transform(collectedfrom.value, x -> lower(x))'))\
.drop('collectedfrom.value')\
.where(
(F.array_contains(F.col('collected_lower'), "opencitations"))
| (F.array_contains(F.col('collected_lower'), "crossref"))
| (F.array_contains(F.col('collected_lower'), "mag"))
).drop('collected_lower')
# print ("Cited df has: " + str(cites_df.count()) + " entries") # print ("Cited df has: " + str(cites_df.count()) + " entries")
# DEPRECATED # DEPRECATED