bolognese 0.4.3 → 0.5

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 6f368d772caba68f6a931b069036edd1fac3e15b
-  data.tar.gz: 85312771808c2971a7b10004f5b77cf6f6276952
+  metadata.gz: d3f81c7d2d95a4c66adb94672309f22ffdae3f7c
+  data.tar.gz: 8c00d86fb7ee25359bbe568638674140d99a6d1d
 SHA512:
-  metadata.gz: ed95b4b96238556e09a234defd26b402d72055204f31f4b9a0411774a5cca13079a91aeb80cccf3f812599a48b94c37fcffa1181ad324d730799894aafd37256
-  data.tar.gz: 2f33f803a3a5ef8b1e957db914225f55fc460a60f5706406f9b43ce3ec5a9542c99fd393267723aed7d5c400b2213f6fb48c334a23d8d8ab1044b1e158b712ca
+  metadata.gz: a64a2cc22e1be39d8e94aa94ef2cbc2ba35300d322c74043153a81c81589f99a6e0fe49922fa1dd493ba6bd9aeb7eee9e72f7edfc141bee13367b1359589035c
+  data.tar.gz: af27b299a2b95bf3250bc7dce7941898f69f17eb7c85dad303057926ec0e94245de4ee1f5b7c3b8ea902d388dcee706098b992ef94462feeb02bab7cf49aa16f

data/.travis.yml

@@ -2,13 +2,19 @@ language: ruby
 rvm:
 - 2.3.3
 
+addons:
+  code_climate:
+    repo_token: $CODECLIMATE_REPO_TOKEN
+
 before_install:
 - export TZ=Europe/Berlin
 
 install:
 - travis_retry bundle install
 
-script: bundle exec rspec
+script:
+- bundle exec rspec
+- bundle exec codeclimate-test-reporter
 
 notifications:
   email: false

data/Gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: .
   specs:
-    bolognese (0.4.3)
+    bolognese (0.5)
       activesupport (~> 4.2, >= 4.2.5)
       builder (~> 3.2, >= 3.2.2)
       dotenv (~> 2.1, >= 2.1.1)

data/README.md

@@ -1,8 +1,10 @@
 [![Build Status](https://travis-ci.org/datacite/bolognese.svg?branch=master)](https://travis-ci.org/datacite/bolognese)
+[![Code Climate](https://codeclimate.com/github/datacite/bolognese/badges/gpa.svg)](https://codeclimate.com/github/datacite/bolognese)
+[![Test Coverage](https://codeclimate.com/github/datacite/bolognese/badges/coverage.svg)](https://codeclimate.com/github/datacite/bolognese/coverage)
 
-# bolognese
+# Bolognese
 
-Command-line utility for conversion of DOI metadata from and to [schema.org](https://schema.org) in JSON-LD.
+Ruby gem and command-line utility for conversion of DOI metadata from and to [schema.org](https://schema.org) in JSON-LD.
 
 ## Features
 
@@ -33,29 +35,271 @@ gem install bolognese
 
 ## Commands
 
-The bolognese commands understand URLs and DOIs as arguments. The `--as` command
+The `bolognese` commands understand URLs and DOIs as arguments. The `--as` command
 line flag sets the format, either `crossref`, `datacite`, or `schema_org` (default).
 
 ## Examples
 
-Convert Crossref XML to schema.org/JSON-LD:
+Read Crossref XML:
 ```
-bolognese read https://doi.org/10.7554/elife.01567
+bolognese read https://doi.org/10.7554/elife.01567 --as crossref
 ```
 
-Read Crossref XML:
+Convert Crossref XML to schema.org/JSON-LD:
 ```
-bolognese read https://doi.org/10.7554/elife.01567 --as crossref
+bolognese read https://doi.org/10.7554/elife.01567
+
+{
+    "@context": "http://schema.org",
+    "@type": "ScholarlyArticle",
+    "@id": "https://doi.org/10.7554/elife.01567",
+    "additionalType": "JournalArticle",
+    "name": "Automated quantitative histology reveals vascular morphodynamics during Arabidopsis hypocotyl secondary growth",
+    "author": [{
+        "@type": "Person",
+        "givenName": "Martial",
+        "familyName": "Sankar"
+    }, {
+        "@type": "Person",
+        "givenName": "Kaisa",
+        "familyName": "Nieminen"
+    }, {
+        "@type": "Person",
+        "givenName": "Laura",
+        "familyName": "Ragni"
+    }, {
+        "@type": "Person",
+        "givenName": "Ioannis",
+        "familyName": "Xenarios"
+    }, {
+        "@type": "Person",
+        "givenName": "Christian S",
+        "familyName": "Hardtke"
+    }],
+    "license": "http://creativecommons.org/licenses/by/3.0/",
+    "datePublished": "2014-02-11",
+    "dateModified": "2015-08-11T05:35:02Z",
+    "isPartOf": {
+        "@type": "Periodical",
+        "name": "eLife",
+        "issn": "2050-084X"
+    },
+    "citation": [{
+        "@type": "CreativeWork",
+        "@id": "https://doi.org/10.1038/nature02100",
+        "position": "1",
+        "datePublished": "2003"
+    }, {
+        "@type": "CreativeWork",
+        "@id": "https://doi.org/10.1534/genetics.109.104976",
+        "position": "2",
+        "datePublished": "2009"
+    }, {
+        "@type": "CreativeWork",
+        "@id": "https://doi.org/10.1034/j.1399-3054.2002.1140413.x",
+        "position": "3",
+        "datePublished": "2002"
+    }, {
+        "@type": "CreativeWork",
+        "@id": "https://doi.org/10.1162/089976601750399335",
+        "position": "4",
+        "datePublished": "2001"
+    }, {
+        "@type": "CreativeWork",
+        "position": "5",
+        "datePublished": "1995"
+    }, {
+        "@type": "CreativeWork",
+        "position": "6",
+        "datePublished": "1993"
+    }, {
+        "@type": "CreativeWork",
+        "@id": "https://doi.org/10.1016/j.semcdb.2009.09.009",
+        "position": "7",
+        "datePublished": "2009"
+    }, {
+        "@type": "CreativeWork",
+        "@id": "https://doi.org/10.1242/dev.091314",
+        "position": "8",
+        "datePublished": "2013"
+    }, {
+        "@type": "CreativeWork",
+        "@id": "https://doi.org/10.1371/journal.pgen.1002997",
+        "position": "9",
+        "datePublished": "2012"
+    }, {
+        "@type": "CreativeWork",
+        "@id": "https://doi.org/10.1038/msb.2010.25",
+        "position": "10",
+        "datePublished": "2010"
+    }, {
+        "@type": "CreativeWork",
+        "@id": "https://doi.org/10.1016/j.biosystems.2012.07.004",
+        "position": "11",
+        "datePublished": "2012"
+    }, {
+        "@type": "CreativeWork",
+        "@id": "https://doi.org/10.1016/j.pbi.2005.11.013",
+        "position": "12",
+        "datePublished": "2006"
+    }, {
+        "@type": "CreativeWork",
+        "@id": "https://doi.org/10.1105/tpc.110.076083",
+        "position": "13",
+        "datePublished": "2010"
+    }, {
+        "@type": "CreativeWork",
+        "@id": "https://doi.org/10.1073/pnas.0808444105",
+        "position": "14",
+        "datePublished": "2008"
+    }, {
+        "@type": "CreativeWork",
+        "@id": "https://doi.org/10.1016/0092-8674(89)90900-8",
+        "position": "15",
+        "datePublished": "1989"
+    }, {
+        "@type": "CreativeWork",
+        "@id": "https://doi.org/10.1126/science.1066609",
+        "position": "16",
+        "datePublished": "2002"
+    }, {
+        "@type": "CreativeWork",
+        "@id": "https://doi.org/10.1104/pp.104.040212",
+        "position": "17",
+        "datePublished": "2004"
+    }, {
+        "@type": "CreativeWork",
+        "@id": "https://doi.org/10.1038/nbt1206-1565",
+        "position": "18",
+        "datePublished": "2006"
+    }, {
+        "@type": "CreativeWork",
+        "@id": "https://doi.org/10.1073/pnas.77.3.1516",
+        "position": "19",
+        "datePublished": "1980"
+    }, {
+        "@type": "CreativeWork",
+        "@id": "https://doi.org/10.1093/bioinformatics/btq046",
+        "position": "20",
+        "datePublished": "2010"
+    }, {
+        "@type": "CreativeWork",
+        "@id": "https://doi.org/10.1105/tpc.111.084020",
+        "position": "21",
+        "datePublished": "2011"
+    }, {
+        "@type": "CreativeWork",
+        "@id": "https://doi.org/10.5061/dryad.b835k",
+        "position": "22",
+        "datePublished": "2014"
+    }, {
+        "@type": "CreativeWork",
+        "@id": "https://doi.org/10.1016/j.cub.2008.02.070",
+        "position": "23",
+        "datePublished": "2008"
+    }, {
+        "@type": "CreativeWork",
+        "@id": "https://doi.org/10.1111/j.1469-8137.2010.03236.x",
+        "position": "24",
+        "datePublished": "2010"
+    }, {
+        "@type": "CreativeWork",
+        "@id": "https://doi.org/10.1007/s00138-011-0345-9",
+        "position": "25",
+        "datePublished": "2012"
+    }, {
+        "@type": "CreativeWork",
+        "@id": "https://doi.org/10.1016/j.cell.2012.02.048",
+        "position": "26",
+        "datePublished": "2012"
+    }, {
+        "@type": "CreativeWork",
+        "@id": "https://doi.org/10.1038/ncb2764",
+        "position": "27",
+        "datePublished": "2013"
+    }],
+    "provider": {
+        "@type": "Organization",
+        "name": "Crossref"
+    }
+}
 ```
 
 Convert Crossref XML to DataCite XML:
 ```
 bolognese read https://doi.org/10.7554/elife.01567 --as datacite
-```
 
-Convert DataCite XML to schema.org/JSON-LD:
-```
-bolognese read 10.5061/DRYAD.8515
+?xml version="1.0" encoding="UTF-8"?>
+<resource xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns="http://datacite.org/schema/kernel-4" xsi:schemaLocation="http://datacite.org/schema/kernel-4 http://schema.datacite.org/meta/kernel-4/metadata.xsd">
+  <identifier identifierType="DOI">10.7554/eLife.01567</identifier>
+  <creators>
+    <creator>
+      <creatorName>Sankar, Martial</creatorName>
+      <givenName>Martial</givenName>
+      <familyName>Sankar</familyName>
+    </creator>
+    <creator>
+      <creatorName>Nieminen, Kaisa</creatorName>
+      <givenName>Kaisa</givenName>
+      <familyName>Nieminen</familyName>
+    </creator>
+    <creator>
+      <creatorName>Ragni, Laura</creatorName>
+      <givenName>Laura</givenName>
+      <familyName>Ragni</familyName>
+    </creator>
+    <creator>
+      <creatorName>Xenarios, Ioannis</creatorName>
+      <givenName>Ioannis</givenName>
+      <familyName>Xenarios</familyName>
+    </creator>
+    <creator>
+      <creatorName>Hardtke, Christian S</creatorName>
+      <givenName>Christian S</givenName>
+      <familyName>Hardtke</familyName>
+    </creator>
+  </creators>
+  <titles>
+    <title>Automated quantitative histology reveals vascular morphodynamics during Arabidopsis hypocotyl secondary growth</title>
+  </titles>
+  <publisher>eLife</publisher>
+  <publicationYear>2014</publicationYear>
+  <resourceType resourceTypeGeneral="Text">JournalArticle</resourceType>
+  <dates>
+    <date dateType="Issued">2014-02-11</date>
+    <date dateType="Updated">2015-08-11T05:35:02Z</date>
+  </dates>
+  <relatedIdentifiers>
+    <relatedIdentifier relatedIdentifierType="DOI" relationType="References">https://doi.org/10.1038/nature02100</relatedIdentifier>
+    <relatedIdentifier relatedIdentifierType="DOI" relationType="References">https://doi.org/10.1534/genetics.109.104976</relatedIdentifier>
+    <relatedIdentifier relatedIdentifierType="DOI" relationType="References">https://doi.org/10.1034/j.1399-3054.2002.1140413.x</relatedIdentifier>
+    <relatedIdentifier relatedIdentifierType="DOI" relationType="References">https://doi.org/10.1162/089976601750399335</relatedIdentifier>
+    <relatedIdentifier relatedIdentifierType="DOI" relationType="References">https://doi.org/10.1016/j.semcdb.2009.09.009</relatedIdentifier>
+    <relatedIdentifier relatedIdentifierType="DOI" relationType="References">https://doi.org/10.1242/dev.091314</relatedIdentifier>
+    <relatedIdentifier relatedIdentifierType="DOI" relationType="References">https://doi.org/10.1371/journal.pgen.1002997</relatedIdentifier>
+    <relatedIdentifier relatedIdentifierType="DOI" relationType="References">https://doi.org/10.1038/msb.2010.25</relatedIdentifier>
+    <relatedIdentifier relatedIdentifierType="DOI" relationType="References">https://doi.org/10.1016/j.biosystems.2012.07.004</relatedIdentifier>
+    <relatedIdentifier relatedIdentifierType="DOI" relationType="References">https://doi.org/10.1016/j.pbi.2005.11.013</relatedIdentifier>
+    <relatedIdentifier relatedIdentifierType="DOI" relationType="References">https://doi.org/10.1105/tpc.110.076083</relatedIdentifier>
+    <relatedIdentifier relatedIdentifierType="DOI" relationType="References">https://doi.org/10.1073/pnas.0808444105</relatedIdentifier>
+    <relatedIdentifier relatedIdentifierType="DOI" relationType="References">https://doi.org/10.1016/0092-8674(89)90900-8</relatedIdentifier>
+    <relatedIdentifier relatedIdentifierType="DOI" relationType="References">https://doi.org/10.1126/science.1066609</relatedIdentifier>
+    <relatedIdentifier relatedIdentifierType="DOI" relationType="References">https://doi.org/10.1104/pp.104.040212</relatedIdentifier>
+    <relatedIdentifier relatedIdentifierType="DOI" relationType="References">https://doi.org/10.1038/nbt1206-1565</relatedIdentifier>
+    <relatedIdentifier relatedIdentifierType="DOI" relationType="References">https://doi.org/10.1073/pnas.77.3.1516</relatedIdentifier>
+    <relatedIdentifier relatedIdentifierType="DOI" relationType="References">https://doi.org/10.1093/bioinformatics/btq046</relatedIdentifier>
+    <relatedIdentifier relatedIdentifierType="DOI" relationType="References">https://doi.org/10.1105/tpc.111.084020</relatedIdentifier>
+    <relatedIdentifier relatedIdentifierType="DOI" relationType="References">https://doi.org/10.5061/dryad.b835k</relatedIdentifier>
+    <relatedIdentifier relatedIdentifierType="DOI" relationType="References">https://doi.org/10.1016/j.cub.2008.02.070</relatedIdentifier>
+    <relatedIdentifier relatedIdentifierType="DOI" relationType="References">https://doi.org/10.1111/j.1469-8137.2010.03236.x</relatedIdentifier>
+    <relatedIdentifier relatedIdentifierType="DOI" relationType="References">https://doi.org/10.1007/s00138-011-0345-9</relatedIdentifier>
+    <relatedIdentifier relatedIdentifierType="DOI" relationType="References">https://doi.org/10.1016/j.cell.2012.02.048</relatedIdentifier>
+    <relatedIdentifier relatedIdentifierType="DOI" relationType="References">https://doi.org/10.1038/ncb2764</relatedIdentifier>
+  </relatedIdentifiers>
+  <rightsList>
+    <rights rightsURI="http://creativecommons.org/licenses/by/3.0/">Creative Commons Attribution 3.0 (CC-BY 3.0)</rights>
+  </rightsList>
+</resource>
 ```
 
 Read DataCite XML:
@@ -63,6 +307,76 @@ Read DataCite XML:
 bolognese read 10.5061/DRYAD.8515 --as datacite
 ```
 
+Convert DataCite XML to schema.org/JSON-LD:
+```sh
+bolognese read 10.5061/DRYAD.8515
+
+{
+    "@context": "http://schema.org",
+    "@type": "Dataset",
+    "@id": "https://doi.org/10.5061/dryad.8515",
+    "additionalType": "DataPackage",
+    "name": "Data from: A new malaria agent in African hominids.",
+    "alternateName": "Ollomo B, Durand P, Prugnolle F, Douzery EJP, Arnathau C, Nkoghe D, Leroy E, Renaud F (2009) A new malaria agent in African hominids. PLoS Pathogens 5(5): e1000446.",
+    "author": [{
+        "@type": "Person",
+        "givenName": "Benjamin",
+        "familyName": "Ollomo"
+    }, {
+        "@type": "Person",
+        "givenName": "Patrick",
+        "familyName": "Durand"
+    }, {
+        "@type": "Person",
+        "givenName": "Franck",
+        "familyName": "Prugnolle"
+    }, {
+        "@type": "Person",
+        "givenName": "Emmanuel J. P.",
+        "familyName": "Douzery"
+    }, {
+        "@type": "Person",
+        "givenName": "Céline",
+        "familyName": "Arnathau"
+    }, {
+        "@type": "Person",
+        "givenName": "Dieudonné",
+        "familyName": "Nkoghe"
+    }, {
+        "@type": "Person",
+        "givenName": "Eric",
+        "familyName": "Leroy"
+    }, {
+        "@type": "Person",
+        "givenName": "François",
+        "familyName": "Renaud"
+    }],
+    "license": "http://creativecommons.org/publicdomain/zero/1.0/",
+    "version": "1",
+    "keywords": "Phylogeny, Malaria, Parasites, Taxonomy, Mitochondrial genome, Africa, Plasmodium",
+    "datePublished": "2011",
+    "hasPart": [{
+        "@type": "CreativeWork",
+        "@id": "https://doi.org/10.5061/dryad.8515/1"
+    }, {
+        "@type": "CreativeWork",
+        "@id": "https://doi.org/10.5061/dryad.8515/2"
+    }],
+    "citation": [{
+        "@type": "CreativeWork",
+        "@id": "https://doi.org/10.1371/journal.ppat.1000446"
+    }],
+    "publisher": {
+        "@type": "Organization",
+        "name": "Dryad Digital Repository"
+    },
+    "provider": {
+        "@type": "Organization",
+        "name": "DataCite"
+    }
+}
+```
+
 ## Development
 
 We use rspec for unit testing:
@@ -72,6 +386,7 @@ bundle exec rspec
 ```
 
 Follow along via [Github Issues](https://github.com/datacite/bolognese/issues).
+Please open an issue if conversion fails or metadata are not properly supported.
 
 ### Note on Patches/Pull Requests
 

data/lib/bolognese/author_utils.rb

@@ -3,11 +3,11 @@ require 'namae'
 module Bolognese
   module AuthorUtils
     # only assume personal name when using sort-order: "Turing, Alan"
-    def get_one_author(author, options = {})
+    def get_one_author(author)
       orcid = get_name_identifier(author)
       author = author.fetch("creatorName", nil)
 
-      return { "Name" => "" } if author.strip.blank?
+      return { "Name" => "" } if author.to_s.strip.blank?
 
       author = cleanup_author(author)
       names = Namae.parse(author)
@@ -44,15 +44,17 @@ module Bolognese
     end
 
     # parse array of author strings into CSL format
-    def get_authors(authors, options={})
-      Array(authors).map { |author| get_one_author(author, options) }
+    def get_authors(authors)
+      Array(authors).map { |author| get_one_author(author) }
     end
 
     # pase nameIdentifier from DataCite
     def get_name_identifier(author)
       name_identifier = author.dig("nameIdentifier", "text")
+      name_identifier = validate_orcid(name_identifier)
+
       name_identifier_scheme = author.dig("nameIdentifier", "nameIdentifierScheme") || "ORCID"
-      if name_identifier_scheme.downcase == "orcid" && name_identifier = validate_orcid(name_identifier)
+      if name_identifier.present? && name_identifier_scheme.downcase == "orcid"
         "http://orcid.org/#{name_identifier}"
       else
         nil