theoj 0.0.1 → 1.1.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: '09237502c90ed400980274941c03683a2f5b33eaa911debc8b9d50d0747e5a91'
-  data.tar.gz: eb959c2aa0821a45246d3a7551f1fc265c0dc571c9ce791840e7465fc5bb6045
+  metadata.gz: 23579418555d490f75911e9894fd4c8bb8abaf3225fc6ee93eacd4d60de621fb
+  data.tar.gz: 8a05d5d50b4a6e22d5ea9dc420c11e4529f71c217727249178226fae5e14d4c1
 SHA512:
-  metadata.gz: 877313b88d2739e7076ec44b18e7efe50026f099f0e2a55e95919aa2af8ea8d72edbb35ee4882154c650b79134bbb72cdf2780f26e8daa64d06a7505ecd91e26
-  data.tar.gz: 54b6835727136c213a5339d303000bc6f2ee47e8ae24118a3577c2e70657a36613417397074bf7228ffd66a6b3949df0fd7b91649e06be4ba9bdb3b7664ec5b2
+  metadata.gz: faa8c6b8ada15280342a218da6485f762ed9535e3f4f9527d8378a66e852635254cbc7bae220827fa08162d1253992ccf0b25c9e8096345e37c502c7a4268bd8
+  data.tar.gz: 4ac1b3347391848f66ea6b1e3f0a648313c62ee1873d2d8d0e1d610732c860a2d656353f44ba8ce07b96ad0da08f13d3ded3f5dacee5e50a7d46a6abe6496b69

data/CHANGELOG.md

@@ -0,0 +1,36 @@
+# Changelog
+
+## 1.1.0 (2021-10-29)
+
+- Added Theoj::PublishedPaper object with metadata from Journal's API
+- Added custom Error class
+
+## 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
+- Added to ReviewIssue: editor, reviewers, archive, version
+- New method to read any value from review's issue body
+- Values read from issue boy will be empty if Pending or TBD
+- Added journal config data for OpenJournals: JOSS and JOSE
+
+
+## 0.0.2 (2021-09-22)
+
+- Available objects: Theoj::Journal, Theoj::ReviewIssue and Theoj::Paper
+
+
+## 0.0.1 (2021-09-22)
+
+- Gem created
+

data/README.md

@@ -2,3 +2,4 @@
 A library to manage editorial objects used by the open journals review process
 
 [![CI](https://github.com/xuanxu/theoj/actions/workflows/ci.yml/badge.svg)](https://github.com/xuanxu/theoj/actions/workflows/ci.yml)
+[![Gem Version](https://badge.fury.io/rb/theoj.svg)](https://badge.fury.io/rb/theoj)

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 Theoj::Error, "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 Theoj::Error, "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

@@ -0,0 +1,64 @@
+require_relative "journals_data"
+
+module Theoj
+  class Journal
+    attr_accessor :data
+    attr_accessor :doi_prefix
+    attr_accessor :url
+    attr_accessor :name
+    attr_accessor :alias
+    attr_accessor :launch_date
+
+    def initialize(custom_data = {})
+      set_data custom_data
+    end
+
+    def current_year
+      data[:current_year] || Time.new.year
+    end
+
+    def current_volume
+      data[:current_volume] || (Time.new.year - (launch_year - 1))
+    end
+
+    def current_issue
+      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 = default_data.merge(custom_data)
+      @doi_prefix = data[:doi_prefix]
+      @url = data[:url]
+      @name = data[:name]
+      @alias = data[:alias]
+      @launch_date = data[:launch_date]
+    end
+
+    def default_data
+      Theoj::JOURNALS_DATA[:joss]
+    end
+
+    def parsed_launch_date
+      @parsed_launch_date ||= Time.parse(data[:launch_date])
+    end
+
+    def launch_year
+      parsed_launch_date.year
+    end
+
+    def launch_month
+      parsed_launch_date.month
+    end
+  end
+end

data/lib/theoj/journals_data.rb

@@ -0,0 +1,25 @@
+module Theoj
+  JOURNALS_DATA = {
+    joss: {
+      doi_prefix: "10.21105",
+      url: "https://joss.theoj.org",
+      name: "Journal of Open Source Software",
+      alias: "joss",
+      launch_date: "2016-05-05",
+      papers_repository: "openjournals/joss-papers",
+      reviews_repository: "openjournals/joss-reviews",
+      deposit_url: "https://joss.theoj.org/papers/api_deposit"
+    },
+    jose: {
+      doi_prefix: "10.21105",
+      url: "https://jose.theoj.org",
+      name: "Journal of Open Source Education",
+      alias: "jose",
+      launch_date: "2018-03-07",
+      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

@@ -1,4 +1,7 @@
-require 'find'
+require "find"
+require "yaml"
+require "rugged"
+require "linguist"
 
 module Theoj
   class Paper
@@ -7,25 +10,57 @@ module Theoj
     attr_accessor :review_issue
     attr_accessor :repository
     attr_accessor :paper_path
+    attr_accessor :branch
+    attr_accessor :paper_metadata
 
-    def initialize(repository, branch, path = nil)
-      @repository = repository
+    def initialize(repository_url, branch, path = nil)
+      @repository = repository_url
       @branch = branch
       find_paper path
+      load_metadata
     end
 
-    def find_paper(path)
-      if path.to_s.strip.empty?
-        setup_local_repo
-        @paper_path = Theoj::Paper.find_paper_path(local_path)
-        cleanup
+    def authors
+      @authors ||= parse_authors
+    end
+
+    def citation_author
+      surname = authors.first.last_name
+      initials = authors.first.initials
+
+      if authors.size > 1
+        return "#{surname} et al."
       else
-        @paper_path = path
+        return "#{surname}, #{initials}"
       end
     end
 
-    def authors
-      []
+    def title
+      @paper_metadata["title"]
+    end
+
+    def tags
+      @paper_metadata["tags"]
+    end
+
+    def date
+      @paper_metadata["date"]
+    end
+
+    def languages
+      @languages ||= detect_languages
+    end
+
+    def bibliography_path
+      @paper_metadata["bibliography"]
+    end
+
+    def local_path
+      @local_path ||= "tmp/#{SecureRandom.hex}"
+    end
+
+    def cleanup
+      FileUtils.rm_rf(local_path) if Dir.exist?(local_path)
     end
 
     def self.find_paper_path(search_path)
@@ -43,33 +78,81 @@ module Theoj
       paper_path
     end
 
-    def self.from_repo(repository, branch = "main")
-      Paper.new(repository, branch, nil)
+    def self.from_repo(repository_url, branch = "")
+      Paper.new(repository_url, branch, nil)
     end
 
     private
+
+      def find_paper(path)
+        if path.to_s.strip.empty?
+          setup_local_repo
+          @paper_path = Theoj::Paper.find_paper_path(local_path)
+        else
+          @paper_path = path
+        end
+      end
+
       def setup_local_repo
         msg_no_repo = "Downloading of the repository failed. Please make sure the URL is correct."
-        msg_no_branch = "Couldn't check the bibtex because branch name is incorrect: #{paper_branch.to_s}"
+        msg_no_branch = "Couldn't check the bibtex because branch name is incorrect: #{branch.to_s}"
 
-        error = clone_repo(target_repository, local_path) ? nil : msg_no_repo
-        (error = change_branch(paper_branch, local_path) ? nil : msg_no_branch) unless error
+        error = clone_repo(repository, local_path) ? nil : msg_no_repo
+        (error = change_branch(branch, local_path) ? nil : msg_no_branch) unless error
 
         failure(error) if error
         error.nil?
       end
 
-      def local_path
-        @local_path ||= "tmp/#{SecureRandom.hex}"
+      def load_metadata
+        @paper_metadata ||= if paper_path.include?('.tex')
+          YAML.load_file(paper_path.gsub('.tex', '.yml'))
+        else
+          YAML.load_file(paper_path)
+        end
       end
 
-      def cleanup
-        FileUtils.rm_rf(local_path) if Dir.exist?(local_path)
+      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)
+        raise Theoj::Error, msg
       end
   end
 end

data/lib/theoj/published_paper.rb

@@ -0,0 +1,46 @@
+require "faraday"
+require "yaml"
+require "json"
+
+module Theoj
+  class PublishedPaper
+    include Theoj::Git
+
+    attr_accessor :metadata
+
+    def initialize(doi)
+      doi_url = "https://doi.org/#{doi}"
+      doi_response = Faraday.get(doi_url)
+      if doi_response.status == 302
+        paper_url = doi_response.headers[:location]
+      else
+        raise Theoj::Error, "The DOI is invalid, url does not resolve #{doi_url}"
+      end
+
+      paper_data = Faraday.get(paper_url + ".json")
+      if paper_data.status == 200
+        @metadata = JSON.parse(paper_data.body, symbolize_names: true)
+      else
+        raise Theoj::Error, "Could not find the paper data at #{paper_url + ".json"}"
+      end
+    end
+
+    [:title, :state, :submitted_at, :doi, :published_at,
+     :volume, :issue, :year, :page, :authors, :editor,
+     :editor_name, :editor_url, :editor_orcid, :reviewers,
+     :languages, :tags, :software_repository, :paper_review,
+     :pdf_url, :software_archive].each do |method_name|
+      define_method(method_name) { metadata[method_name] }
+      define_method("#{method_name}=") {|value| @metadata[method_name] = value }
+    end
+
+    def yaml_metadata
+      metadata.to_yaml
+    end
+
+    def json_metadata
+      metadata.to_json
+    end
+
+  end
+end

data/lib/theoj/review_issue.rb

@@ -16,20 +16,41 @@ module Theoj
     end
 
     def issue_body
-      issue(repository, issue_id).body
+      @issue_body ||= issue(repository, issue_id).body
     end
 
     def paper
+      @paper ||= Theoj::Paper.new(target_repository, paper_branch)
     end
 
     def target_repository
       @target_repository ||= read_value_from_body("target-repository")
     end
 
+    def reviewers
+      @reviewers ||= read_value_from_body("reviewers").split(",").map{|r| r.strip} - ["Pending", "TBD"]
+    end
+
+    def editor
+      @editor ||= read_value_from_body("editor")
+    end
+
+    def archive
+      @archive ||= read_value_from_body("archive")
+    end
+
+    def version
+      @version ||= read_value_from_body("version")
+    end
+
     def paper_branch
       @paper_branch ||= read_value_from_body("branch")
     end
 
+    def value_for(value_name)
+      read_value_from_body(value_name)
+    end
+
     private
 
       def read_from_body(start_mark, end_mark)
@@ -37,6 +58,8 @@ module Theoj
         issue_body.match(/#{start_mark}(.*)#{end_mark}/im) do |m|
           text = m[1]
         end
+
+        text = "" if ["Pending", "TBD"].include?(text.strip)
         text.strip
       end
 

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 = "0.0.1"
+  VERSION = "1.1.0"
 end

data/lib/theoj.rb

@@ -1,9 +1,14 @@
 require_relative "theoj/version"
 require_relative "theoj/git"
 require_relative "theoj/github"
+require_relative "theoj/orcid"
+require_relative "theoj/published_paper"
+require_relative "theoj/submission"
+require_relative "theoj/journal"
 require_relative "theoj/review_issue"
 require_relative "theoj/paper"
 require_relative "theoj/author"
 
 module Theoj
+  class Error < StandardError; end
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: theoj
 version: !ruby/object:Gem::Version
-  version: 0.0.1
+  version: 1.1.0
 platform: ruby
 authors:
 - Juanjo Bazán
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2021-09-22 00:00:00.000000000 Z
+date: 2021-10-29 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
@@ -59,21 +115,27 @@ executables: []
 extensions: []
 extra_rdoc_files: []
 files:
+- CHANGELOG.md
 - LICENSE
 - README.md
 - lib/theoj.rb
 - lib/theoj/author.rb
 - lib/theoj/git.rb
 - lib/theoj/github.rb
+- lib/theoj/journal.rb
+- lib/theoj/journals_data.rb
+- lib/theoj/orcid.rb
 - lib/theoj/paper.rb
+- lib/theoj/published_paper.rb
 - lib/theoj/review_issue.rb
+- lib/theoj/submission.rb
 - lib/theoj/version.rb
 homepage: http://github.com/xuanxu/theoj
 licenses:
 - MIT
 metadata:
   bug_tracker_uri: https://github.com/xuanxu/theoj/issues
-  changelog_uri: https://github.com/xuanxu/theoj/blob/master/CHANGELOG.md
+  changelog_uri: https://github.com/xuanxu/theoj/blob/main/CHANGELOG.md
   documentation_uri: https://www.rubydoc.info/gems/theoj
   homepage_uri: http://github.com/xuanxu/theoj
   source_code_uri: http://github.com/xuanxu/theoj