aboutsummaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
authorBryan Newbold <bnewbold@archive.org>2023-01-04 21:24:49 -0800
committerBryan Newbold <bnewbold@archive.org>2023-01-04 21:28:26 -0800
commit38ba3d6a9c5138afbd82a1a5025f43e08bbba6a2 (patch)
tree189cfc0845a2dd8c44b546ddaaf9c4e92a5a570e
parent011f309d5a0056f633b88de2b33e5714548eba56 (diff)
downloadfatcat-scholar-38ba3d6a9c5138afbd82a1a5025f43e08bbba6a2.tar.gz
fatcat-scholar-38ba3d6a9c5138afbd82a1a5025f43e08bbba6a2.zip
commit old notes
-rw-r--r--notes/2021_webpage_size.txt29
-rw-r--r--notes/2022-12_bulk_index.md72
2 files changed, 101 insertions, 0 deletions
diff --git a/notes/2021_webpage_size.txt b/notes/2021_webpage_size.txt
new file mode 100644
index 0000000..e4bcade
--- /dev/null
+++ b/notes/2021_webpage_size.txt
@@ -0,0 +1,29 @@
+
+comparing HTML size across changes (eg, adding citations and modals)
+
+query is "superglue" with defaults
+
+older prod: 61KB
+newer dev: 62KB (15 KB compressed)
+
+modal dev (buttons, modal skeleton): 104 KB
+
+dropdown dev (buttons, citations, drop-down fulltext, versions): 107 KB
+
+2020-01-20
+ bare (404 page): 15 kB
+ head CSS: 5.3 kB
+ translate svg: 1.7 kB
+ serp ("superglue"): 107 kB
+ citations: 10.4 kB
+ access drop-down: 10.2 kB
+ highlights: 7 kB
+ all serp content: 83 kB
+ all right column: 35 kB
+ serp no hits: 25.2 kB
+
+take-aways:
+- move CSS to separate file
+- citations not that bad, but some wins by linking instead of including (?)
+- links are heavy
+- don't sweat it too much, compresses well
diff --git a/notes/2022-12_bulk_index.md b/notes/2022-12_bulk_index.md
new file mode 100644
index 0000000..b2c5cac
--- /dev/null
+++ b/notes/2022-12_bulk_index.md
@@ -0,0 +1,72 @@
+
+Running through a full end-to-end re-indexing.
+
+
+## Fatcat Metadata Dumps
+
+Run following fatcat notes (elsewhere).
+
+Download to working machine:
+
+ export JOBDIR=/kubwa/fatcat/2022-11-24
+ mkdir -p $JOBDIR
+ cd $JOBDIR
+ wget -c https://archive.org/download/fatcat_bulk_exports_2022-11-24/release_export_expanded.json.gz
+
+## Microfilm
+
+Working directory: `aitio:/fast/fatcat-scholar`.
+
+Pulled latest git (`00d80752b7d83ae5a165540fbad641ddfc78b5f3`), and ran `make
+dep`.
+
+Run:
+
+ TODAY=2022-12-08 make issue-db
+
+Then, the SIM dump job, in parallel:
+
+ export JOBDIR=/kubwa/scholar/2022-12-08
+ mkdir -p $JOBDIR
+ pipenv shell
+ python -m fatcat_scholar.sim_pipeline run_print_issues \
+ | shuf \
+ | parallel -j16 --colsep "\t" python -m fatcat_scholar.sim_pipeline run_fetch_issue {1} {2} \
+ | pv -l \
+ | pigz \
+ > $JOBDIR/sim_intermediate.2022-12-08.json.gz
+ => 45.4M 42:09:42 [ 298 /s]
+
+TODO: there were some old publications that should not be included... gazetteer? registers?
+ "Daily Gazetteer" (sim_daily-gazetteer)
+
+## Works Bulk Fetch
+
+First split up the release dump into chunks:
+
+ export JOBDIR=/kubwa/scholar/2022-12-08
+ mkdir -p $JOBDIR
+ cd $JOBDIR
+ zcat /kubwa/fatcat/2022-11-24/release_export_expanded.json.gz | split --lines 8000000 - release_export_expanded.split_ -d --additional-suffix .json
+ => done
+
+Note: more shards this time around (up to 23, not 21).
+
+Starting the below commands on 2022-12-21.
+
+ export JOBDIR=/kubwa/scholar/2022-12-08
+ cd /fast/fatcat-scholar
+ pipenv shell
+ export TMPDIR=/sandcrawler-db/tmp
+ # possibly re-export JOBDIR from above?
+
+ # fetch
+ set -u -o pipefail
+ for SHARD in {00..23}; do
+ cat $JOBDIR/release_export_expanded.split_$SHARD.json \
+ | parallel -j8 --line-buffer --compress --tmpdir $TMPDIR --round-robin --pipe python -m fatcat_scholar.work_pipeline run_releases \
+ | pv -l \
+ | pigz \
+ > $JOBDIR/fatcat_scholar_work_fulltext.split_$SHARD.json.gz.WIP \
+ && mv $JOBDIR/fatcat_scholar_work_fulltext.split_$SHARD.json.gz.WIP $JOBDIR/fatcat_scholar_work_fulltext.split_$SHARD.json.gz
+ done