diff options
author | Bryan Newbold <bnewbold@robocracy.org> | 2018-11-13 12:43:12 -0800 |
---|---|---|
committer | Bryan Newbold <bnewbold@robocracy.org> | 2018-11-13 12:43:12 -0800 |
commit | e8a2925394f4cce0b8b4514f58d2bd19f9d7490b (patch) | |
tree | c060f01ac5e3e63d08a28cf38d0ade55267fc893 /python/fatcat_tools/importers/common.py | |
parent | 572fdc7caf74d9539e642e97855d8c8ba94ff93a (diff) | |
download | fatcat-e8a2925394f4cce0b8b4514f58d2bd19f9d7490b.tar.gz fatcat-e8a2925394f4cce0b8b4514f58d2bd19f9d7490b.zip |
use Counter object instead of per-metric ints
Diffstat (limited to 'python/fatcat_tools/importers/common.py')
-rw-r--r-- | python/fatcat_tools/importers/common.py | 12 |
1 files changed, 6 insertions, 6 deletions
diff --git a/python/fatcat_tools/importers/common.py b/python/fatcat_tools/importers/common.py index 8dfee875..d289171d 100644 --- a/python/fatcat_tools/importers/common.py +++ b/python/fatcat_tools/importers/common.py @@ -4,6 +4,8 @@ import sys import csv import json import itertools +from collections import Counter + import fatcat_client from fatcat_client.rest import ApiException @@ -26,13 +28,11 @@ class FatcatImporter: self._orcid_regex = re.compile("^\\d{4}-\\d{4}-\\d{4}-\\d{3}[\\dX]$") if issn_map_file: self.read_issn_map_file(issn_map_file) - self.processed_lines = 0 - self.insert_count = 0 - self.update_count = 0 + self.counts = Counter({'insert': 0, 'update': 0, 'processed_lines': 0}) def describe_run(self): print("Processed {} lines, inserted {}, updated {}.".format( - self.processed_lines, self.insert_count, self.update_count)) + self.counts['processed_lines'], self.counts['insert'], self.counts['update'])) def process_source(self, source, group_size=100): """Creates and auto-accepts editgroup every group_size rows""" @@ -44,14 +44,14 @@ class FatcatImporter: self.api.accept_editgroup(eg.id) eg = self.api.create_editgroup( fatcat_client.Editgroup(editor_id='aaaaaaaaaaaabkvkaaaaaaaaae')) - self.processed_lines = self.processed_lines + 1 + self.counts['processed_lines'] += 1 if i == 0 or (i % group_size) != 0: self.api.accept_editgroup(eg.id) def process_batch(self, source, size=50): """Reads and processes in batches (not API-call-per-)""" for rows in grouper(source, size): - self.processed_lines = self.processed_lines + len(rows) + self.counts['processed_lines'] += len(rows) eg = self.api.create_editgroup( fatcat_client.Editgroup(editor_id='aaaaaaaaaaaabkvkaaaaaaaaae')) self.create_batch(rows, editgroup=eg.id) |