Commit message (Collapse) | Author | Age | Files | Lines | |
---|---|---|---|---|---|
* | langdetect: more text for 'zh' test case | Bryan Newbold | 2020-11-20 | 1 | -1/+1 |
| | | | | | | This is an attempt to fix spurious test failures, in which this text block was getting detected as 'kr' on occasion. Apparently there is non-determinism in the langdetect package. | ||||
* | crossref+datacite: remove confusing early update bail | Bryan Newbold | 2020-11-20 | 2 | -4/+0 |
| | | | | | Easy to miss that we skip updates *twice*, and with this early bailout were not updating counts correctly. | ||||
* | doaj: fix update code path (getattr not __dict__) | Bryan Newbold | 2020-11-20 | 1 | -4/+3 |
| | | | | Also add missing code coverage for update path (disabled by default). | ||||
* | DOAJ: handle empty identifier 'id' case | Bryan Newbold | 2020-11-20 | 1 | -0/+2 |
| | |||||
* | clean DOI: ban all non-ASCII characters | Bryan Newbold | 2020-11-19 | 1 | -1/+4 |
| | | | | | | | I believe this is safe and matches the regex filter in rust (fatcatd). Keep hitting one-off DOIs that were passing through python check, so being more strict from here forward. | ||||
* | normal: handle langdetect of 'zh-cn' (not len=2) | Bryan Newbold | 2020-11-19 | 1 | -0/+3 |
| | |||||
* | tweak DOAJ importer class args and default for do_updates | Bryan Newbold | 2020-11-19 | 1 | -2/+2 |
| | |||||
* | if a release has DOAJ article id, count as OA | Bryan Newbold | 2020-11-19 | 1 | -0/+3 |
| | |||||
* | implement remainder of DOAJ article importer | Bryan Newbold | 2020-11-19 | 1 | -57/+125 |
| | |||||
* | handle more non-ASCII DOI cases | Bryan Newbold | 2020-11-19 | 1 | -1/+3 |
| | |||||
* | more python normalizers, and move from importer common | Bryan Newbold | 2020-11-19 | 2 | -154/+326 |
| | | | | | | | | | | | | Moved several normalizer helpers out of fatcat_tools.importers.common to fatcat_tools.normal. Copied language name and country name parser helpers from chocula repository (built on existing pycountry helper library). Have not gone through and refactored other importers to point to these helpers yet; that should be a separate PR when this branch is merged. Current changes are backwards compatible via re-imports. | ||||
* | initial implementation of DOAJ importer | Bryan Newbold | 2020-11-19 | 2 | -0/+290 |
| | | | | Several things to finish implementing and polish. | ||||
* | html ingest: actual xhtml mimetype | Bryan Newbold | 2020-11-16 | 1 | -2/+2 |
| | |||||
* | ingest tool: support for setting ingest type | Bryan Newbold | 2020-11-06 | 1 | -6/+6 |
| | |||||
* | html ingest: remaining implementation | Bryan Newbold | 2020-11-06 | 1 | -22/+19 |
| | |||||
* | ingest: progress on HTML ingest | Bryan Newbold | 2020-11-05 | 1 | -14/+30 |
| | |||||
* | ingest: initial 'web' worker implementation | Bryan Newbold | 2020-11-05 | 2 | -67/+259 |
| | |||||
* | refactor: white/black -> allow/block | Bryan Newbold | 2020-11-05 | 1 | -4/+4 |
| | |||||
* | ingest: whitelist -> allowlist | Bryan Newbold | 2020-11-05 | 1 | -3/+3 |
| | |||||
* | ingest: basic checks for ingest_type | Bryan Newbold | 2020-11-05 | 1 | -3/+29 |
| | |||||
* | normalizer: filter out a specific non-ASCII character in DOI | Bryan Newbold | 2020-11-04 | 1 | -1/+3 |
| | |||||
* | entity updates: don't ingest JSTOR DOI prefixes | Bryan Newbold | 2020-10-23 | 1 | -0/+2 |
| | |||||
* | entity updater: new work update feed (ident and changelog metadata only) | Bryan Newbold | 2020-10-16 | 1 | -2/+24 |
| | |||||
* | chocula importer: small tweaks to update behavior | Bryan Newbold | 2020-10-08 | 1 | -8/+6 |
| | |||||
* | elastic transform: more preservation keepers | Bryan Newbold | 2020-10-08 | 1 | -1/+2 |
| | |||||
* | address spammy datacite titles | Martin Czygan | 2020-09-23 | 1 | -0/+19 |
| | | | | | | | | | seemingly from zenodo: * https://fatcat.wiki/release/rzcpjwukobd4pj36ipla22cnoi * https://doi.org/10.5281/zenodo.4041777 About 3400 records with "FULL MOVIE" in title, currently. | ||||
* | ingest: default to crawl protocols.io DOIs | Bryan Newbold | 2020-09-10 | 1 | -0/+2 |
| | |||||
* | datacite: handle case of empty-string version | Bryan Newbold | 2020-09-10 | 1 | -1/+1 |
| | | | | | Includes a tiny tweak to the datacite import sample file to test this code path. | ||||
* | remove spurious print statement | Bryan Newbold | 2020-09-03 | 1 | -1/+0 |
| | |||||
* | generic file entity clean-ups as part of file_meta importer | Bryan Newbold | 2020-09-02 | 2 | -0/+50 |
| | |||||
* | fix comment typo (thanks martin) | Bryan Newbold | 2020-08-27 | 1 | -1/+1 |
| | |||||
* | fixes and test coverage for file_meta importer | Bryan Newbold | 2020-08-21 | 1 | -5/+10 |
| | |||||
* | initial implementation of file_meta importer | Bryan Newbold | 2020-08-21 | 2 | -0/+71 |
| | |||||
* | entity updater: handle doi=None case better | Bryan Newbold | 2020-08-14 | 1 | -1/+1 |
| | |||||
* | entity updater: es['publisher_type'] not always set | Bryan Newbold | 2020-08-14 | 1 | -1/+1 |
| | | | | This is a small bugfix for a production issue. | ||||
* | Merge branch 'bnewbold-ingest-improvements' into 'master' | Martin Czygan | 2020-08-13 | 2 | -33/+114 |
|\ | | | | | | | | | ingest behavior changes; some datacite metadata tweaks See merge request webgroup/fatcat!78 | ||||
| * | entity update: change big5 ingest behavior | Bryan Newbold | 2020-08-11 | 1 | -9/+15 |
| | | | | | | | | | | | | | | | | | | In addition to changing the OA default, this was the main intended behavior change in this group of commits: want to ingest fewer attempts that we *expect* to fail, but default to ingest/crawl attempt if we are uncertain. This is because there is a long tail of journals that register DOIs and are defacto OA (fulltext is available), but we don't have metadata indicating them as such. | ||||
| * | entity update: default to ingest non-OA works | Bryan Newbold | 2020-08-11 | 1 | -9/+10 |
| | | |||||
| * | entity update: skip ingest of figshare+zenodo 'group' DOIs | Bryan Newbold | 2020-08-11 | 1 | -0/+15 |
| | | |||||
| * | datacite import: figshare-specific hacks | Bryan Newbold | 2020-08-11 | 1 | -3/+3 |
| | | |||||
| * | datacite import: refactor release_type detection into static method | Bryan Newbold | 2020-08-11 | 1 | -14/+51 |
| | | |||||
| * | datacite import: refactor publisher-specific hacks into static method | Bryan Newbold | 2020-08-11 | 1 | -15/+29 |
| | | | | | | | | Also tweak title/publisher detection to use DOI prefixes | ||||
| * | update crawl blocklist for SPNv2 requests which mostly fail | Bryan Newbold | 2020-08-10 | 1 | -2/+10 |
| | | |||||
* | | harvest: datacite API yields HTTP 200 with broken JSON | Martin Czygan | 2020-08-10 | 1 | -1/+8 |
|/ | | | | As a first step: log response body for debugging. | ||||
* | release ES transform tweaks | Bryan Newbold | 2020-08-07 | 1 | -3/+5 |
| | | | | | | | | pass-through publisher_type from container extra metadata (ES field already existed; this is from newer chocula metadata) count arxiv and PMCID papers which haven't been crawled (by IA) as "dark", not "bright" | ||||
* | chocula import update tweaks | Bryan Newbold | 2020-08-04 | 1 | -10/+14 |
| | |||||
* | more update keys and cases for chocula importer | Bryan Newbold | 2020-08-04 | 1 | -5/+11 |
| | |||||
* | fix key name mismatch in chocula importer | Bryan Newbold | 2020-08-04 | 1 | -1/+1 |
| | | | | chocula 'export-fatcat' uses 'ident', not 'fatcat_ident' | ||||
* | basic toml transform helper | Bryan Newbold | 2020-07-30 | 2 | -4/+20 |
| | |||||
* | Merge branch 'bnewbold-more-lint-fixes' into 'master' | Martin Czygan | 2020-07-24 | 6 | -25/+18 |
|\ | | | | | | | | | more lint fixes See merge request webgroup/fatcat!69 |