aboutsummaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* tasks: remove obsolete tasksMartin Czygan2021-08-022-71/+6
| | | | for Open Library we currenty use editions and authors only
* xio: mention lockMartin Czygan2021-07-311-1/+1
| | | | | there are other lock-free implementations, such as https://github.com/cloudfoundry/go-diodes
* tasks: include OL into brefMartin Czygan2021-07-311-7/+5
|
* wip: technical report draftMartin Czygan2021-07-303-9/+115
|
* wip: paperMartin Czygan2021-07-307-0/+638
|
* update notesMartin Czygan2021-07-301-0/+12
|
* tasks: update docsMartin Czygan2021-07-291-0/+8
|
* tasks: reuse tag for filename dateMartin Czygan2021-07-291-4/+14
|
* tasks: use 2021-07-28 as default dateMartin Czygan2021-07-281-2/+2
|
* tasks: read 'tag' from settings as wellMartin Czygan2021-07-281-1/+3
|
* v0.1.5Martin Czygan2021-07-281-1/+1
|
* tasks: remove non-ref sets from Bref taskMartin Czygan2021-07-281-4/+7
|
* v0.2.3Martin Czygan2021-07-281-1/+1
|
* reduce: remove log statementsMartin Czygan2021-07-281-4/+0
|
* reduce: adjust testMartin Czygan2021-07-281-3/+3
|
* reduce: fix test filesMartin Czygan2021-07-282-52/+104
|
* adjust reduce test casesMartin Czygan2021-07-284-1087/+1087
|
* v0.2.2Martin Czygan2021-07-281-1/+1
|
* leave ref.index unchangedMartin Czygan2021-07-281-6/+6
| | | | | | | | | previously, we started with 0-indexed input, but wanted 1-indexed values so we added increments at various points which probably lead to bug (missing refs), since at one point we would fuse the original ref data (w/o increments) with the matched data (w/ increments); with scholar:528804ad2e55983cf3e5e6659d8f46db0cab02b7 we can now leave indices as is
* reduce: add caseMartin Czygan2021-07-285-0/+93
|
* reduce: add more logging, temporarilyMartin Czygan2021-07-271-1/+6
|
* add bref augment scaffoldMartin Czygan2021-07-274-0/+43
|
* tests: use more sensible namesMartin Czygan2021-07-272-6/+6
|
* move test utilities into testutilMartin Czygan2021-07-272-32/+41
|
* annotate test casesMartin Czygan2021-07-271-5/+12
|
* update docsMartin Czygan2021-07-271-4/+1
|
* update dependenciesMartin Czygan2021-07-272-6/+0
|
* remove unused/partially implemented skate-dot for nowMartin Czygan2021-07-272-75/+1
|
* minor tweaks and doc improvementsMartin Czygan2021-07-272-49/+45
|
* update todo notesMartin Czygan2021-07-271-2/+22
|
* reuse timestampsMartin Czygan2021-07-271-6/+14
| | | | | | | | | | while time.Now is not really slow, thanks to vDSO (cf. https://git.io/J4SOH), it will be even faster to just call it once at the start of the processing; also: https://twitter.com/davidcrawshaw/status/1414243408936280073 > Turns out http://time.Now was taking its usual amount of time on linux, ~50 nanoseconds [...]
* reduce: explicitly name magic numbersMartin Czygan2021-07-271-3/+8
|
* schema: add note regarding field nameMartin Czygan2021-07-271-1/+1
|
* v0.2.1Martin Czygan2021-07-271-1/+1
|
* schema: tweaksMartin Czygan2021-07-272-5/+112
| | | | add String() to CSLDate; we only cover a few typical cases
* cleanup and docsMartin Czygan2021-07-271-25/+4
|
* reduce: use pascal caseMartin Czygan2021-07-261-2/+2
|
* v0.2.0Martin Czygan2021-07-261-1/+1
| | | | | | | | | | | | lots of tweaks * normalize (eg, lower-case) DOIs in all (or at least most?) situations, especially for equality comparisons * don't try to parse DOI from ref_key (which may contain a source DOI, but not a target DOI) * switch to using date-parts for year in target_csl output * switch from author.name to author.raw_name in target_csl output (neither are standard; raw_name indicates this better) * pass through match_provenance in unmatched case * in target_csl output, don't always include issued and accessed dates as * empty objects (could save significant ES index disk space?)
* switch to sligthly more performance string builderMartin Czygan2021-07-263-43/+41
|
* reduce: mention upcoming change to indexingMartin Czygan2021-07-261-1/+1
| | | | see: scholar:528804ad2e55983cf3e5e6659d8f46db0cab02b7
* Merge branch 'bnewbold-skate-tweaks' into 'master'Martin Czygan2021-07-269-54/+116
|\ | | | | | | | | proposed changes and fixes to skate matching See merge request martin/cgraph!3
| * skate: use SanitizeDOI in all inputsBryan Newbold2021-07-254-22/+9
| |
| * skate: fast SanitizeDOI helper for normalizing DOIsBryan Newbold2021-07-252-0/+71
| |
| * skate unstructured: don't parse DOI out of keyBryan Newbold2021-07-251-16/+0
| | | | | | | | | | | | DOIs in keys, usually from Crossref, are the DOI of the *source* of the reference, not the *target* of the reference. Thus, they should not be parsed and copied to the ref.biblio.doi field.
| * skate: pass-through match_provenance in more situationsBryan Newbold2021-07-251-0/+2
| |
| * schema: switch from '.name' to '.raw_name' for un-parsed CSL name fieldBryan Newbold2021-07-253-6/+6
| |
| * skate: use date-parts for year, not 'raw'Bryan Newbold2021-07-252-8/+9
| |
| * schema: have issued+accessed (CSLDate) actually omitemptyBryan Newbold2021-07-243-5/+5
| | | | | | | | | | Similar to TargetCSL, these should be pointer types so they don't get encoded as empty objects when not set.
| * add test for issued,accessed not being included in output JSONBryan Newbold2021-07-241-0/+17
| |
* | ci: show coverageMartin Czygan2021-07-261-2/+1
| |
generated by cgit v1.2.3 (git 2.39.1) at 2024-11-23 16:58:41 +0000