dimitris.pierrakos
/
dnet-hadoop
forked from D-Net/dnet-hadoop
1
0
Fork 0
Code Pull Requests Activity

Format bip-score based file without doi references

This commit is contained in:
Ilias Kanellos 2023-07-21 13:42:30 +03:00
parent 0c433eccdd
commit 03da965162
1 changed files with 9 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

14
dhp-workflows/dhp-impact-indicators/src/main/resources/eu/dnetlib/dhp/oa/graph/impact_indicators/oozie_app/format_ranking_results.py
View File

@ -213,7 +213,10 @@ if mode == 'bip':
cc_dir = sys.argv[4]
impulse_dir = sys.argv[5]
ram_dir = sys.argv[6]
refs_dir = sys.argv[7]
# NOTE: This was used initial, but @Serafeim told me to remove it since we don't get doi-doi referencew anymore
# In case of emergency, bring this back
# refs_dir = sys.argv[7]
# Score-specific dataframe
pagerank_df = spark.read.schema(float_schema).option('delimiter', '\t').option('header',True).csv(pagerank_dir).repartition(num_partitions, 'id')
@ -221,7 +224,7 @@ if mode == 'bip':
cc_df = spark.read.schema(int_schema).option('delimiter', '\t').option('header',True).csv(cc_dir).repartition(num_partitions, 'id')
impulse_df = spark.read.schema(int_schema).option('delimiter', '\t').option('header',True).csv(impulse_dir).repartition(num_partitions, 'id')
ram_df = spark.read.schema(float_schema).option('delimiter', '\t').option('header', True).csv(ram_dir).repartition(num_partitions, 'id')
refs_df = spark.read.schema(refs_schema).option('delimiter', '\t').option('header',True).csv(refs_dir).repartition(num_partitions, 'id')
# refs_df = spark.read.schema(refs_schema).option('delimiter', '\t').option('header',True).csv(refs_dir).repartition(num_partitions, 'id')
# ----------- TESTING CODE --------------- #
# pagerank_entries = pagerank_df.count()
@ -258,9 +261,10 @@ if mode == 'bip':
.select(results_df.id, 'pagerank', 'pagerank_normalized', 'attrank', 'attrank_normalized', 'cc', 'cc_normalized',\
'3-cc', '3-cc_normalized', F.col('score').alias('ram'))
# Add references
results_df = results_df.join(refs_df, ['id']).select(results_df.id, 'pagerank', 'pagerank_normalized', 'attrank', 'attrank_normalized', \
'cc', 'cc_normalized', '3-cc', '3-cc_normalized', 'ram', 'num_refs')
# Add references - THIS WAS REMOVED SINCE WE DON't GET DOI REFERENCES
# In case of emergency bring back
# results_df = results_df.join(refs_df, ['id']).select(results_df.id, 'pagerank', 'pagerank_normalized', 'attrank', 'attrank_normalized', \
# 'cc', 'cc_normalized', '3-cc', '3-cc_normalized', 'ram', 'num_refs')
# Write resulting dataframe to file
output_dir = "/".join(pagerank_dir.split('/')[:-1])