aboutsummaryrefslogtreecommitdiffstats
path: root/python/fatcat_tools/harvest
diff options
context:
space:
mode:
authorBryan Newbold <bnewbold@robocracy.org>2018-11-21 10:13:24 -0800
committerBryan Newbold <bnewbold@robocracy.org>2018-11-21 10:13:24 -0800
commit75a0af495472bfddb15047d3ad15810d47aceb9a (patch)
tree95a1ddeee92b7a6a917553db7171b22e8dcce0b7 /python/fatcat_tools/harvest
parent5b0d7c1cd40a7a255886703c00d060eacfc27901 (diff)
downloadfatcat-75a0af495472bfddb15047d3ad15810d47aceb9a.tar.gz
fatcat-75a0af495472bfddb15047d3ad15810d47aceb9a.zip
fix OAI-PMH name/finished message
Diffstat (limited to 'python/fatcat_tools/harvest')
-rw-r--r--python/fatcat_tools/harvest/oaipmh.py7
1 files changed, 6 insertions, 1 deletions
diff --git a/python/fatcat_tools/harvest/oaipmh.py b/python/fatcat_tools/harvest/oaipmh.py
index 4044ff10..ffcd5e78 100644
--- a/python/fatcat_tools/harvest/oaipmh.py
+++ b/python/fatcat_tools/harvest/oaipmh.py
@@ -75,6 +75,7 @@ class HarvestOaiPmhWorker:
self.endpoint_url = None # needs override
self.metadata_prefix = None # needs override
+ self.name = "unnamed"
self.state = HarvestState(start_date, end_date)
self.state.initialize_from_kafka(self.kafka.topics[self.state_topic])
@@ -119,7 +120,7 @@ class HarvestOaiPmhWorker:
time.sleep(self.loop_sleep())
else:
break
- print("{} DOI ingest caught up".format(self.name))
+ print("{} OAI-PMH ingest caught up".format(self.name))
class HarvestArxivWorker(HarvestOaiPmhWorker):
@@ -128,6 +129,7 @@ class HarvestArxivWorker(HarvestOaiPmhWorker):
super().__init__(**kwargs)
self.endpoint_url = "https://export.arxiv.org/oai2"
self.metadata_prefix = "arXiv"
+ self.name = "arxiv"
class HarvestPubmedWorker(HarvestOaiPmhWorker):
@@ -136,6 +138,7 @@ class HarvestPubmedWorker(HarvestOaiPmhWorker):
super().__init__(**kwargs)
self.endpoint_url = "https://www.ncbi.nlm.nih.gov/pmc/oai/oai.cgi"
self.metadata_prefix = "pmc_fm"
+ self.name = "pubmed"
class HarvestDoajJournalWorker(HarvestOaiPmhWorker):
@@ -147,6 +150,7 @@ class HarvestDoajJournalWorker(HarvestOaiPmhWorker):
super().__init__(**kwargs)
self.endpoint_url = "https://www.doaj.org/oai"
self.metadata_prefix = "oai_dc"
+ self.name = "doaj-journal"
class HarvestDoajArticleWorker(HarvestOaiPmhWorker):
@@ -158,4 +162,5 @@ class HarvestDoajArticleWorker(HarvestOaiPmhWorker):
super().__init__(**kwargs)
self.endpoint_url = "https://www.doaj.org/oai.article"
self.metadata_prefix = "oai_doaj"
+ self.name = "doaj-article"