theoj 0.0.3 → 1.0.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 86f78e023b1c26cfcd13eea306fcb1ee89a862a300e43e173437655fddcc53b9
-  data.tar.gz: 43c48ae2941a5f72076f048a58315f861aedddd1708a5886b437b1858739553d
+  metadata.gz: 984f9feb0bdcfeaf7d9533aa94c09dde7877ca2381e75e60b0f05d73cb390eb7
+  data.tar.gz: 749c5b08e99824f720c5ed4e820d45b7d2b772c345f820d4332393376b215512
 SHA512:
-  metadata.gz: 1cba25b3701a0f52f40317e74df287db75a4f79944183a13bd1ea23567e51b010cdaa2f6d707aaedaa6e5c2ac30119c1dd7eab3f83e0d107a4d048d247e9ee80
-  data.tar.gz: 592c523fb973d4f7c57044b451d17038b5a891aa0d534f4081180c4409f5da81ccebd31686c4f98151fbdeea465f61ea2e58df08aa4b1c2770d3af13a71d72d1
+  metadata.gz: 5fa3741bb127abcfb1c590c197fa8ea9631b510624b25b615ccfc60131cdf7ce68af2d400ab16f339cc6705a5aae1129403ac0c1d24e41e04209704bdfbcb076
+  data.tar.gz: 246621ddee2c0495eccad79723727205b724ec36b4b401b24ae60f22c1e1cb7a8ece320ab2ccefb6a5074e4423a4eabcbf401272d40e137b0eb831bf0dafd662

data/CHANGELOG.md

@@ -1,5 +1,16 @@
 # Changelog
 
+## 1.0.0 (2021-10-20)
+
+- Added method to Journal to create paper_id from issue_id
+- Added method to Journal to get a DOI based on a paper id
+- Added languages to Paper
+- Added authors info to Paper
+- Author object
+- Added ORCID validation
+- Added Submission object, grouping a paper, a review issue and a journal
+- Added paper depositing
+
 ## 0.0.3 (2021-10-08)
 
 - Added metadata methods to Paper

data/lib/theoj/author.rb

@@ -1,16 +1,94 @@
+require "nameable"
+
 module Theoj
   class Author
     attr_accessor :name
     attr_accessor :orcid
     attr_accessor :affiliation
 
-    def initialize(name, orcid, affiliation)
-      @name = name
-      @orcid = orcid
-      @affiliation = affiliation
+    AUTHOR_FOOTNOTE_REGEX = /^[^\^]*/
+
+    # Initialized with authors & affiliations block in the YAML header from an Open Journal paper
+    # e.g. https://joss.readthedocs.io/en/latest/submitting.html#example-paper-and-bibliography
+    def initialize(name, orcid, index, affiliations_hash)
+      parse_name name
+      @orcid = validate_orcid orcid
+      @affiliation = build_affiliation_string(index, affiliations_hash)
+    end
+
+    def given_name
+      @parsed_name.first
+    end
+
+    def middle_name
+      @parsed_name.middle
+    end
+
+    def last_name
+      @parsed_name.last
+    end
+
+    def initials
+      [@parsed_name.first, @parsed_name.middle].compact.map {|v| v[0] + "."} * ' '
+    end
+
+    def to_h
+      {
+        given_name: given_name,
+        middle_name: middle_name,
+        last_name: last_name,
+        orcid: orcid,
+        affiliation: affiliation
+      }
+    end
+
+    private
+
+    def parse_name(author_name)
+      @parsed_name = Nameable::Latin.new.parse(strip_footnotes(author_name))
+      @name = @parsed_name.to_nameable
+    end
+
+    # Input: Arfon Smith^[Corresponding author: arfon@example.com]
+    # Output: Arfon Smith
+    def strip_footnotes(author_name)
+      author_name.to_s[AUTHOR_FOOTNOTE_REGEX]
+    end
+
+    def validate_orcid(author_orcid)
+      return nil if author_orcid.to_s.strip.empty?
+
+      validator = Theoj::Orcid.new(author_orcid)
+      if validator.valid?
+        return author_orcid.strip
+      else
+        raise "Problem with ORCID (#{author_orcid}) for #{self.name}. #{validator.error}"
+      end
     end
 
-    def validate_orcid
+    # Takes the author affiliation index and a hash of all affiliations and
+    # associates them. Then builds the author affiliation string
+    def build_affiliation_string(index, affiliations_hash)
+      return nil if index.nil? # Some authors don't have an affiliation
+
+      # If multiple affiliations, parse each one and build the affiliation string
+      author_affiliations = []
+
+      # Turn YAML keys into strings so that mixed integer and string affiliations work
+      affiliations_hash.transform_keys!(&:to_s)
+
+      affiliations = index.to_s.split(',').map(&:strip)
+
+      # Raise if we can't parse the string, might be because of this bug :-(
+      # https://bugs.ruby-lang.org/issues/12451
+      affiliations.each do |a|
+        raise "Problem with affiliations for #{self.name}, perhaps the " +
+              "affiliations index need quoting?" unless affiliations_hash.has_key?(a)
+
+        author_affiliations << affiliations_hash[a].strip
+      end
+
+      author_affiliations.join(', ')
     end
 
   end

data/lib/theoj/git.rb

@@ -1,5 +1,5 @@
-require 'open3'
-require 'fileutils'
+require "open3"
+require "fileutils"
 
 module Theoj
   module Git

data/lib/theoj/journal.rb

@@ -25,6 +25,15 @@ module Theoj
       data[:current_issue] || (1 + ((Time.new.year * 12 + Time.new.month) - (launch_year * 12 + launch_month)))
     end
 
+    def paper_id_from_issue(review_issue_id)
+      id = "%05d" % review_issue_id
+      "#{@alias}.#{id}"
+    end
+
+    def paper_doi_for_id(paper_id)
+      "#@doi_prefix/#{paper_id}"
+    end
+
     private
 
     def set_data(custom_data)

data/lib/theoj/journals_data.rb

@@ -6,8 +6,9 @@ module Theoj
       name: "Journal of Open Source Software",
       alias: "joss",
       launch_date: "2016-05-05",
-      paper_repository: "openjournals/joss-papers",
-      reviews_repository: "openjournals/joss-reviews"
+      papers_repository: "openjournals/joss-papers",
+      reviews_repository: "openjournals/joss-reviews",
+      deposit_url: "https://joss.theoj.org/papers/api_deposit"
     },
     jose: {
       doi_prefix: "10.21105",
@@ -15,8 +16,9 @@ module Theoj
       name: "Journal of Open Source Education",
       alias: "jose",
       launch_date: "2018-03-07",
-      paper_repository: "openjournals/jose-papers",
-      reviews_repository: "openjournals/jose-reviews"
+      papers_repository: "openjournals/jose-papers",
+      reviews_repository: "openjournals/jose-reviews",
+      deposit_url: "https://joss.theoj.org/papers/api_deposit"
     }
   }
 end

data/lib/theoj/orcid.rb

@@ -0,0 +1,96 @@
+module Theoj
+  class Orcid
+    attr_reader :orcid, :error
+
+    def initialize(orcid)
+      @orcid = orcid.strip
+      @error = nil
+    end
+
+    def valid?
+      @error = nil
+      return false unless check_structure
+      return false unless check_length
+      return false unless check_chars
+
+      return false unless correct_checksum?
+
+      true
+    end
+
+    def packed_orcid
+      orcid.gsub('-', '')
+    end
+
+    private
+
+    # Returns the last character of the string
+    def checksum_char
+      packed_orcid[-1]
+    end
+
+    def first_11
+      packed_orcid.chop
+    end
+
+    def check_structure
+      groups = orcid.split('-')
+      if groups.size == 4
+        return true
+      else
+        @error = "ORCID looks malformed"
+        return false
+      end
+    end
+
+    def check_length
+      if packed_orcid.length == 16
+        return true
+      else
+        @error = "ORCID looks to be the wrong length"
+        return false
+      end
+    end
+
+    def check_chars
+      valid = true
+      first_11.each_char do |c|
+        if !numeric?(c)
+          @error = "Invalid ORCID digit (#{c})"
+          valid = false
+        end
+      end
+
+      return valid
+    end
+
+    def correct_checksum?
+      validate_against = checksum_char.to_i
+      validate_against = 10 if (checksum_char == "X" || checksum_char == "x")
+
+      if checksum == validate_against
+        return true
+      else
+        @error = "Invalid ORCID"
+        return false
+      end
+    end
+
+    # https://support.orcid.org/knowledgebase/articles/116780-structure-of-the-orcid-identifier
+    def checksum
+      total = 0
+      first_11.each_char do |c|
+        total = (total + c.to_i) * 2
+      end
+
+      remainder = total % 11
+      result = (12 - remainder) % 11
+    end
+
+
+    def numeric?(s)
+      Float(s) != nil rescue false
+    end
+
+  end
+end

data/lib/theoj/paper.rb

@@ -1,5 +1,7 @@
-require 'find'
-require 'yaml'
+require "find"
+require "yaml"
+require "rugged"
+require "linguist"
 
 module Theoj
   class Paper
@@ -19,22 +21,18 @@ module Theoj
     end
 
     def authors
-      []
+      @authors ||= parse_authors
     end
 
-    def self.find_paper_path(search_path)
-      paper_path = nil
+    def citation_author
+      surname = authors.first.last_name
+      initials = authors.first.initials
 
-      if Dir.exist? search_path
-        Find.find(search_path).each do |path|
-          if path =~ /paper\.tex$|paper\.md$/
-            paper_path = path
-            break
-          end
-        end
+      if authors.size > 1
+        return "#{surname} et al."
+      else
+        return "#{surname}, #{initials}"
       end
-
-      paper_path
     end
 
     def title
@@ -49,6 +47,10 @@ module Theoj
       @paper_metadata["date"]
     end
 
+    def languages
+      @languages ||= detect_languages
+    end
+
     def bibliography_path
       @paper_metadata["bibliography"]
     end
@@ -61,6 +63,21 @@ module Theoj
       FileUtils.rm_rf(local_path) if Dir.exist?(local_path)
     end
 
+    def self.find_paper_path(search_path)
+      paper_path = nil
+
+      if Dir.exist? search_path
+        Find.find(search_path).each do |path|
+          if path =~ /paper\.tex$|paper\.md$/
+            paper_path = path
+            break
+          end
+        end
+      end
+
+      paper_path
+    end
+
     def self.from_repo(repository_url, branch = "")
       Paper.new(repository_url, branch, nil)
     end
@@ -95,6 +112,44 @@ module Theoj
         end
       end
 
+      def parse_authors
+        parsed_authors = []
+        authors_metadata = @paper_metadata['authors']
+        affiliations_metadata = parse_affiliations(@paper_metadata['affiliations'])
+
+        # Loop through the authors block and build up the affiliation
+        authors_metadata.each do |author|
+          affiliation_index = author['affiliation']
+          failure "Author (#{author['name']}) is missing affiliation" if affiliation_index.nil?
+          begin
+            parsed_author = Author.new(author['name'], author['orcid'], affiliation_index, affiliations_metadata)
+          rescue Exception => e
+            failure(e.message)
+          end
+          parsed_authors << parsed_author
+        end
+
+        parsed_authors
+      end
+
+      def parse_affiliations(affliations_yaml)
+        affliations_metadata = {}
+
+        affliations_yaml.each do |affiliation|
+          affliations_metadata[affiliation['index']] = affiliation['name']
+        end
+
+        affliations_metadata
+      end
+
+      def detect_languages
+        repo = Rugged::Repository.discover(paper_path)
+        project = Linguist::Repository.new(repo, repo.head.target_id)
+
+        # Take top five languages from Linguist
+        project.languages.keys.take(5)
+      end
+
       def failure(msg)
         cleanup
         raise(msg)

data/lib/theoj/submission.rb

@@ -0,0 +1,70 @@
+require "json"
+require "base64"
+require "faraday"
+
+module Theoj
+  class Submission
+    attr_accessor :journal
+    attr_accessor :review_issue
+    attr_accessor :paper
+
+    def initialize(journal, review_issue, paper=nil)
+      @journal = journal
+      @review_issue = review_issue
+      @paper = paper || @review_issue.paper
+    end
+
+    # Create the payload to use to post for depositing with Open Journals
+    def deposit_payload
+      {
+        id: review_issue.issue_id,
+        metadata: Base64.encode64(metadata_payload),
+        doi: paper_doi,
+        archive_doi: review_issue.archive,
+        citation_string: citation_string,
+        title: paper.title
+      }
+    end
+
+    # Create a metadata json payload
+    def metadata_payload
+      metadata = {
+        paper: {
+          title: paper.title,
+          tags: paper.tags,
+          languages: paper.languages,
+          authors: paper.authors.collect { |a| a.to_h },
+          doi: paper_doi,
+          archive_doi: review_issue.archive,
+          repository_address: review_issue.target_repository,
+          editor: review_issue.editor,
+          reviewers: review_issue.reviewers.collect(&:strip),
+          volume: journal.current_volume,
+          issue: journal.current_issue,
+          year: journal.current_year,
+          page: review_issue.issue_id,
+        }
+      }
+
+      metadata.to_json
+    end
+
+    def deposit!(secret)
+      parameters = deposit_payload.merge(secret: secret)
+      Faraday.post(journal.data[:deposit_url], parameters.to_json, {"Content-Type" => "application/json"})
+    end
+
+    def citation_string
+      paper_year = Time.now.strftime('%Y')
+      "#{paper.citation_author}, (#{paper_year}). #{paper.title}. #{journal.name}, #{journal.current_volume}(#{journal.current_issue}), #{review_issue.issue_id}, https://doi.org/#{paper_doi}"
+    end
+
+    def paper_id
+      journal.paper_id_from_issue(review_issue.issue_id)
+    end
+
+    def paper_doi
+      journal.paper_doi_for_id(paper_id)
+    end
+  end
+end

data/lib/theoj/version.rb

@@ -1,3 +1,3 @@
 module Theoj
-  VERSION = "0.0.3"
+  VERSION = "1.0.0"
 end

data/lib/theoj.rb

@@ -1,6 +1,8 @@
 require_relative "theoj/version"
 require_relative "theoj/git"
 require_relative "theoj/github"
+require_relative "theoj/orcid"
+require_relative "theoj/submission"
 require_relative "theoj/journal"
 require_relative "theoj/review_issue"
 require_relative "theoj/paper"

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: theoj
 version: !ruby/object:Gem::Version
-  version: 0.0.3
+  version: 1.0.0
 platform: ruby
 authors:
 - Juanjo Bazán
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2021-10-08 00:00:00.000000000 Z
+date: 2021-10-20 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: octokit
@@ -24,6 +24,62 @@ dependencies:
     - - "~>"
       - !ruby/object:Gem::Version
         version: '4.21'
+- !ruby/object:Gem::Dependency
+  name: faraday
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
+- !ruby/object:Gem::Dependency
+  name: openjournals-nameable
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
+- !ruby/object:Gem::Dependency
+  name: github-linguist
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
+- !ruby/object:Gem::Dependency
+  name: rugged
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
 - !ruby/object:Gem::Dependency
   name: rake
   requirement: !ruby/object:Gem::Requirement
@@ -68,8 +124,10 @@ files:
 - lib/theoj/github.rb
 - lib/theoj/journal.rb
 - lib/theoj/journals_data.rb
+- lib/theoj/orcid.rb
 - lib/theoj/paper.rb
 - lib/theoj/review_issue.rb
+- lib/theoj/submission.rb
 - lib/theoj/version.rb
 homepage: http://github.com/xuanxu/theoj
 licenses: