termium 0.3.1 → 0.3.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 86c7b7ab35fb63b5d7bc8a085125db1452284327970939abd89808bf6e628097
-  data.tar.gz: 2229305a3891ae26ec4b8ec82ad3ae55735912c7503c19ed5c478c52d3c1db68
+  metadata.gz: 38f41bac67c56e9aa6ec73f6ec79b729216cfdd677ed27d7317ea1bcd2363bc0
+  data.tar.gz: b69ed20d7a04e49f815a2714c4dc2d3c290af3ecce2b89f8eba9d63bedd85a09
 SHA512:
-  metadata.gz: 67d16aa7301c765ec97865d5eff33b98e02de93d8cce9332fb7d8953613fa638719941bc2e5b695a29ae4f0e8e4960864d4ad17a8d1ee8332bc66e877f81c099
-  data.tar.gz: 97d56933d5946f70209e0c3b054583e683e2e510ff235615d324dbb80c3f76dcdc1acea3fb1fd3003b9f1c417b193636df701685c85dfee0f2f33aee4173bc17
+  metadata.gz: f7cdfffe9d0e84f8af6a8724b82233646ac43351b9fa3107941de8a3754e5fe664c06bcf8841db0fe43aee606c7c66fef3c2dc29c0c2a137bba37125ed24dc83
+  data.tar.gz: 386cf6b5a8658f7a91686771f083558ef31483c3fd961cdb58860082505ba8d3fc066b684dc35157607438b36886d99689b36e245ccd71fb9d549de5673303f7

data/README.adoc

@@ -42,17 +42,37 @@ Results in:
 
 == Commands
 
-`termium convert`:: Convert a TERMIUM Plus export XML file to a Glossarist dataset
+`termium convert`:: Convert a TERMIUM Plus export XML file to a Paneron Glossarist dataset.
 
 
-=== Usage
+=== `termium convert`
+
+==== Purpose
+
+This command converts a TERMIUM Plus export XML (`<ns2:termium_extract>`) file
+to a Paneron Glossarist dataset.
+
+The resulting dataset will look like this:
+
+[source]
+----
+{OUTPUT_PATH}/
+├── concepts/
+│   ├── {CONCEPT_ID}.yaml
+│   ├── ...
+├── localized_concepts/
+    ├── {LOCALIZED_CONCEPT_ID}.yaml
+    ├── ...
+----
+
+==== Usage
 
 [source,sh]
 ----
-$ termium convert -i INPUT_XML_FILE [-o OUTPUT_PATH]
+$ termium convert -i INPUT_XML_FILE [-o OUTPUT_PATH] [-o DATE_ACCEPTED]
 ----
 
-=== Options
+==== Options
 
 [cols="a,a",options="header"]
 |===
@@ -61,7 +81,7 @@ $ termium convert -i INPUT_XML_FILE [-o OUTPUT_PATH]
 |`-i`, `--input-path`
 |
 Source path to TERMIUM Plus XML export file.
-The file needs to start with the `<extract>` tag.
+The file needs to start with the `<ns2:termium_extract>` element.
 
 |`-o`, `--output-path`
 |
@@ -69,9 +89,90 @@ Destination path to Glossarist dataset directory.
 If the directory doesn't exist it will be created.
 If not provided, defaults to the basename of the input file, e.g. `foo/bar.xml` will export to `foo/bar/`.
 
+|`--date-accepted`
+|
+Date of acceptance for the dataset. This fills in the `date_accepted` value of
+the universal concept (which is exported to a YAML file).
+
 |===
 
 
+==== Examples
+
+The data structures of these files can be seen in the following examples.
+
+
+.Sample of `{CONCEPT_ID}.yaml`
+[example]
+====
+This is `88a7dd87-6199-3516-9cec-f4cd79ff09c6.yaml`.
+
+[source,yaml]
+----
+---
+data:
+  identifier: '2120638'
+  localized_concepts:
+    eng: e114ee44-e601-5623-9099-48cfc2be2224
+    fre: 9a7b88cb-4ee6-5d59-89bb-230425a3c96a
+related: []
+date_accepted: 2015-05-01
+status: valid
+id: 88a7dd87-6199-3516-9cec-f4cd79ff09c6
+----
+====
+
+.Sample of `{LOCALIZED_CONCEPT_ID}.yaml`
+[example]
+====
+This is `e114ee44-e601-5623-9099-48cfc2be2224.yaml`.
+
+[source,yaml]
+----
+---
+data:
+  dates: []
+  definition:
+  - content: layer whose nodes directly communicate with external systems
+  examples: []
+  id: '2120638'
+  notes:
+  - content: 'visible layer: term and definition standardized by ISO/IEC [ISO/IEC
+      2382-34:1999].'
+  - content: 34.02.09 (2382)
+  sources:
+  - origin:
+      ref: ISO/IEC 2382-34:1999
+    type: lineage
+    status: identical
+  - origin:
+      ref: Ranger, Natalie * 2006 * Bureau de la traduction / Translation Bureau *
+        Services linguistiques / Linguistic Services * Bur. dir. Centre de traduction
+        et de terminologie / Dir's Office Translation and Terminology Centre * Div.
+        Citoyenneté et Protection civile / Citizen. & Emergency preparedness Div.
+        * Normalisation terminologique / Terminology Standardization
+    type: lineage
+    status: identical
+  terms:
+  - type: expression
+    normative_status: preferred
+    designation: visible layer
+    grammar_info:
+    - preposition: false
+      participle: false
+      adj: false
+      verb: false
+      adverb: false
+      noun: false
+      gender: []
+      number:
+      - singular
+  language_code: eng
+----
+====
+
+
+
 == Library
 
 === Usage

data/lib/termium/cli.rb

@@ -53,7 +53,7 @@ module Termium
       puts "Converting to Glossarist..."
       convert_options = {}
       if options[:date_accepted]
-        convert_options[:date_accepted] = Date.parse(options[:date_accepted])
+        convert_options[:date_accepted] = Date.parse(options[:date_accepted]).iso8601
       end
       glossarist_col = termium_extract.to_concept(convert_options)
       # pp glossarist_col.first

data/lib/termium/core.rb

@@ -38,8 +38,8 @@ module Termium
     end
 
     # Deterministic v4 UUID by using the number string
-    def uuid
-      UUIDTools::UUID.md5_create(UUIDTools::UUID_DNS_NAMESPACE, identification_number).to_s
+    def uuid(str = identification_number)
+      UUIDTools::UUID.md5_create(UUIDTools::UUID_DNS_NAMESPACE, str).to_s
     end
 
     # TODO: Utilize "subject" in the Glossarist object:
@@ -60,11 +60,15 @@ module Termium
           concept.date_accepted = options[:date_accepted]
         end
 
-        language_module.map(&:to_concept).each do |localized_concept|
+        language_module.map do |lang_mod|
+          localized_concept = lang_mod.to_concept(options)
+
           # TODO: This is needed to skip the empty french entries of 10031781 and 10031778
           next if localized_concept.nil?
 
           localized_concept.id = identification_number
+          localized_concept.uuid = uuid("#{identification_number}-#{lang_mod.language}")
+
           universal_entry.each do |entry|
             localized_concept.notes << entry.value
           end

data/lib/termium/language_module.rb

@@ -60,6 +60,7 @@ module Termium
         "definition" => [{ content: definition }],
         "notes" => notes,
         "examples" => examples,
+        "entry_status" => "valid",
       }
 
       src["domain"] = domain if domain
@@ -67,11 +68,19 @@ module Termium
       src
     end
 
-    def to_concept
+    def to_concept(options = {})
       x = to_h
       return nil unless x
 
-      Glossarist::LocalizedConcept.new(x)
+      Glossarist::LocalizedConcept.new(x).tap do |concept|
+        # Fill in register parameters
+        if options[:date_accepted]
+          puts options[:date_accepted].inspect
+          concept.date_accepted = options[:date_accepted]
+        end
+
+        puts concept.inspect
+      end
     end
   end
 end

data/lib/termium/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Termium
-  VERSION = "0.3.1"
+  VERSION = "0.3.2"
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: termium
 version: !ruby/object:Gem::Version
-  version: 0.3.1
+  version: 0.3.2
 platform: ruby
 authors:
 - Ribose
 autorequire: 
 bindir: exe
 cert_chain: []
-date: 2024-11-18 00:00:00.000000000 Z
+date: 2024-11-19 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: glossarist