update bulk edit CHANGELOG and orcid notes

1 files changed, 43 insertions, 0 deletions

diff --git a/notes/bulk_edits/2019-12-20_orcid.md b/notes/bulk_edits/2019-12-20_orcid.md
new file mode 100644
index 00000000..33dde32f
--- /dev/null
+++ b/notes/bulk_edits/2019-12-20_orcid.md
@@ -0,0 +1,43 @@
+
+Newer ORCID dumps are XML, not JSON. But there is a conversion tool!
+
+    https://github.com/ORCID/orcid-conversion-lib
+
+Commands:
+
+    wget https://github.com/ORCID/orcid-conversion-lib/raw/master/target/orcid-conversion-lib-0.0.2-full.jar
+    java -jar orcid-conversion-lib-0.0.2-full.jar OPTIONS
+
+    java -jar orcid-conversion-lib-0.0.2-full.jar --tarball -i ORCID_2019_summaries.tar.gz -v v3_0rc1 -o ORCID_2019_summaries_json.tar.gz
+
+    # [...]
+    # Sat Dec 21 04:43:50 UTC 2019 done 7300000
+    # Sat Dec 21 04:44:08 UTC 2019 done 7310000
+    # Sat Dec 21 04:44:17 UTC 2019 finished  errors 0
+
+Importing in QA, ran in to some lines like:
+
+    {"response-code":409,"developer-message":"409 Conflict: The ORCID record is locked and cannot be edited. ORCID https://orcid.org/0000-0003-0014-6598","user-message":"The ORCID record is locked.","error-code":9018,"more-info":"https://members.orcid.org/api/resources/troubleshooting"}
+    {"response-code":409,"developer-message":"409 Conflict: The ORCID record is locked and cannot be edited. ORCID https://orcid.org/0000-0003-3750-5654","user-message":"The ORCID record is locked.","error-code":9018,"more-info":"https://members.orcid.org/api/resources/troubleshooting"}
+    {"response-code":409,"developer-message":"409 Conflict: The ORCID record is locked and cannot be edited. ORCID https://orcid.org/0000-0003-1424-4826","user-message":"The ORCID record is locked.","error-code":9018,"more-info":"https://members.orcid.org/api/resources/troubleshooting"}
+    {"response-code":409,"developer-message":"409 Conflict: The ORCID record is locked and cannot be edited. ORCID https://orcid.org/0000-0002-5340-9665","user-message":"The ORCID record is locked.","error-code":9018,"more-info":"https://members.orcid.org/api/resources/troubleshooting"}
+
+Needed to patch to filter those out. Then ran ok like:
+
+    zcat /srv/fatcat/datasets/ORCID_2019_summaries.sample_10k.json.gz | ./fatcat_import.py orcid -
+    Counter({'total': 10000, 'exists': 5323, 'insert': 4493, 'skip': 184, 'skip-no-person': 160, 'update': 0})
+
+New dump is about 7.3 million rows, so expecting about 3.2 million new
+entities, 250k skips.
+
+Doing bulk run like:
+
+    time zcat /srv/fatcat/datasets/ORCID_2019_summaries.json.gz | parallel -j8 --round-robin --pipe ./fatcat_import.py orcid -
+
+Prod timing:
+
+    Counter({'total': 910643, 'exists': 476812, 'insert': 416583, 'skip': 17248, 'update': 0})
+
+    real    47m27.658s
+    user    245m44.272s
+    sys     14m50.836s