schema-model 0.7.3 → 0.7.4

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: cd3db69d1a56251cda03284c2af23954d69d79e697fa74c47d564eb1ceb9ddb7
-  data.tar.gz: 79687f0746ef3ea15c78b9ca99baf650b361e31e7ef2f7612f54da6e6fc2ec2d
+  metadata.gz: f273128aa733d5737e486345a39b1c59f5af1bf3babea0917bc8d6e6a44707c8
+  data.tar.gz: 393e5e391cde21e5288da99d4321bb619f370d1a937052f427e32f570facd68e
 SHA512:
-  metadata.gz: 96fa7aa88bbc344fe24dbb245c7e2f3bc90a2dccda8386a41a0171eedb740ca15b3413bf6f73db4ef4f642e21142b83acf6a3de4f136e3b70e87a4e7752bb052
-  data.tar.gz: 7b0e3e1960f407ca135516910340b7f05679f1fd12ebed0efe66b4865b8f9d96ea73d64156bc60d45a3298a3be1aaa95cb2c2a34c82af1282438a69b0707ee45
+  metadata.gz: df5f93d0615667c6bb455bf920a08f44507f59eec3031cde8fb5e73eb99403738d0d3afe447d53a4a39d1dd67df7e8d3ff2f375810580f26a5804851289c8016
+  data.tar.gz: 01eb576246b0efa42c3935c0cb1c73038b350062053a767a815c8eaced1b4e1aec3b8c79fca3c32d2810fb26618b8eda397c4eb6b1ecc08e6baf79ff5467f138

data/.github/dependabot.yml

@@ -0,0 +1,18 @@
+version: 2
+updates:
+  # keeps the SHA-pinned actions in .github/workflows current
+  - package-ecosystem: github-actions
+    directory: /
+    schedule:
+      interval: weekly
+    commit-message:
+      prefix: ci
+      include: scope
+
+  - package-ecosystem: bundler
+    directory: /
+    schedule:
+      interval: weekly
+    commit-message:
+      prefix: build
+      include: scope

data/.release-please-manifest.json

@@ -1,3 +1,3 @@
 {
-  ".": "0.7.3"
+  ".": "0.7.4"
 }

data/.rubocop.yml

@@ -3,7 +3,6 @@ AllCops:
   NewCops: enable
   SuggestExtensions: false
   Exclude:
-    - 'bin/**/*'
     - 'vendor/**/*'
     - 'coverage/**/*'
 

data/ARCHITECTURE.md

@@ -99,7 +99,7 @@ end
 
 The `type_field` option tells SchemaCreator which key in the nested data determines the subclass (the type field attribute's aliases are checked too); `external_type_field` reads the type from an attribute on the parent instead. `type_ignorecase: true` compares type names case-insensitively. Each `add_type` creates a subclass of the association class, defined on the parent class as `<Name>AssociationType<Type>`.
 
-`Associations::Base` stores the parent class by name and resolves it with `Object.const_get`, so a parent class must be assigned to a constant before any association that uses dynamic types is declared.
+`Associations::Base` stores the parent class by name and resolves it with `Object.const_get` (so reloaded constants are picked up); a parent that was anonymous when the association was declared is kept by reference instead.
 
 ### Schema::Utils (`lib/schema/utils.rb`)
 
@@ -127,7 +127,7 @@ Parsing errors use these codes from `Schema::ParsingErrors`: `invalid`, `incompa
 
 ### CSV and Arrays
 
-- **`Schema::ArrayHeaders`** (`lib/schema/array_headers.rb`): `map_headers_to_attributes(headers)` returns a nested hash of `{ field: { index: n } }` entries. Has-many fields are matched as `<prefix><n><key>` (prefix is the association name or an alias, `n` counts from 1) and map to `{ indexes: [...] }`.
+- **`Schema::ArrayHeaders`** (`lib/schema/array_headers.rb`): `map_headers_to_attributes(headers)` returns a nested hash of `{ field: { index: n } }` entries. Has-many fields are matched as `<prefix><n><key>` (prefix is the association name or an alias, `n` counts from 1) and map to `{ indexes: [...] }`. Has-one fields are matched by their own key or alias; a column already claimed by the parent or an earlier has-one isn't reused. Associations never map to a column of their own.
 - **`Schema::Arrays`** (`lib/schema/arrays.rb`): `to_headers`, `to_empty_array`, `to_a` and `from_array(array, mapped_headers)`. Has-many associations need a `size:` option for the fixed-width methods.
 - **`Schema::CSVParser`** (`lib/schema/csv_parser.rb`): wraps a `CSV` object, maps its header row once, and yields a model per row (`each`, `shift`, `missing_fields`).
 

data/CHANGELOG.md

@@ -1,5 +1,14 @@
 # Changelog
 
+## [0.7.4](https://github.com/dougyouch/schema/compare/v0.7.3...v0.7.4) (2026-10-04)
+
+
+### Bug Fixes
+
+* **array-headers:** fix nested field names, association columns and column collisions ([f0373e0](https://github.com/dougyouch/schema/commit/f0373e0450e9d474dcde5e02b1db76b9c7bbc7d4))
+* **associations:** support dynamic types in anonymous parent classes ([45d22da](https://github.com/dougyouch/schema/commit/45d22da233ec97703ba3c3bd90a91e577bac5390))
+* **cli:** write csv headers to stdout and lint schema-json2csv ([58ebb13](https://github.com/dougyouch/schema/commit/58ebb138d58cbddca60c2e7214ab980d3c0aae12))
+
 ## [0.7.3](https://github.com/dougyouch/schema/compare/v0.7.2...v0.7.3) (2026-10-04)
 
 

data/README.md

@@ -12,6 +12,12 @@ Add this line to your application's Gemfile:
 
 ```ruby
 gem 'schema-model'
+
+# needed for Schema::All and Schema::ActiveModelValidations (already present in Rails apps)
+gem 'activemodel'
+
+# needed for Schema::CSVParser and bin/schema-json2csv on Ruby 3.4+, where csv is no longer a default gem
+gem 'csv'
 ```
 
 And then execute:
@@ -20,6 +26,8 @@ And then execute:
 $ bundle install
 ```
 
+`Schema::Model` on its own has no dependencies beyond `inheritance-helper`.
+
 ## Quick Start
 
 ```ruby
@@ -521,6 +529,8 @@ schema-json2csv --require ./my_schema.rb --schema MySchema --json data.json --cs
 cat data.json | schema-json2csv --require ./my_schema.rb --schema MySchema - --csv output.csv
 ```
 
+The JSON can be one object or an array of objects. Output goes to stdout when `--csv` is omitted. A header row (from `to_headers`) is written to stdout or to a new file; an existing `--csv` file is appended to without repeating it. Has-many associations need a `size:` so every row has the same columns.
+
 ## Contributing
 
 1. Fork the repository

data/bin/schema-json2csv

@@ -1,4 +1,5 @@
 #!/usr/bin/env ruby
+# frozen_string_literal: true
 
 require 'schema-model'
 require 'csv'
@@ -13,7 +14,7 @@ options = {
 }
 
 OptionParser.new do |opts|
-  opts.banner = "Usage: " + File.basename(__FILE__)
+  opts.banner = "Usage: #{File.basename(__FILE__)} --require FILE --schema CLASS (--json FILE | -) [--csv FILE]"
 
   opts.on('--json JSON_FILE', 'JSON data') do |v|
     options[:json_file] = v
@@ -23,7 +24,7 @@ OptionParser.new do |opts|
     options[:require_file] = v
   end
 
-  opts.on('--csv CSV_FILE', 'File to output csv data') do |v|
+  opts.on('--csv CSV_FILE', 'File to output csv data (appended to if it exists)') do |v|
     options[:csv_file] = v
   end
 
@@ -32,54 +33,39 @@ OptionParser.new do |opts|
   end
 end.parse!
 
-raise('no filed required') unless options[:require_file]
-raise('no schema specified') unless options[:schema_class_name]
+abort('no file required') unless options[:require_file]
+abort('no schema specified') unless options[:schema_class_name]
+abort("file #{options[:require_file]} not found") unless File.exist?(options[:require_file])
 
-$LOAD_PATH << '.'
-raise("file #{options[:require_file]} not found") unless require(options[:require_file])
+require File.expand_path(options[:require_file])
 
-raise("schema #{options[:schema_class_name]} not found") unless Object.const_defined?(options[:schema_class_name])
+abort("schema #{options[:schema_class_name]} not found") unless Object.const_defined?(options[:schema_class_name])
 
 schema_class = Object.const_get(options[:schema_class_name])
 schema_class.schema_include Schema::Arrays
 
 json_data =
   if options[:json_file]
-    raise("json file #{options[:json_file]} not found") unless File.exist?(options[:json_file])
+    abort("json file #{options[:json_file]} not found") unless File.exist?(options[:json_file])
     File.read(options[:json_file])
   elsif ARGV.last == '-'
-    ARGF.read
+    $stdin.read
   else
-    raise('no json data specified')
+    abort('no json data specified')
   end
 
-def output_csv(schema, io)
-  str = CSV.generate do |csv|
-    csv << schema.to_a.flatten
-  end
-
-  io.print str
-end
+# headers are written to stdout and to new files; an existing file is appended to as is
+csv_file = options[:csv_file]
+write_headers = csv_file.nil? || !File.exist?(csv_file)
+io = csv_file ? File.open(csv_file, 'ab') : $stdout
 
-io =
-  if file = options[:csv_file]
-    if File.exist?(file)
-      File.open(options[:csv_file], 'ab')
-    else
-      headers = CSV.generate do |csv|
-        csv << schema_class.to_headers
-      end
-      f = File.open(options[:csv_file], 'wb')
-      f.print headers
-      f
-    end
-  else
-    $stdout
-  end
+io.print CSV.generate_line(schema_class.to_headers) if write_headers
 
 json_data = JSON.parse(json_data)
 json_data = [json_data] unless json_data.is_a?(Array)
 
 json_data.each do |data|
-  output_csv(schema_class.from_hash(data), io)
+  io.print CSV.generate_line(schema_class.from_hash(data).to_a.flatten)
 end
+
+io.close if csv_file

data/lib/schema/array_headers.rb

@@ -45,13 +45,11 @@ module Schema
 
       private
 
+      # has_many columns are reported as <prefix>X<field>; has_one fields share the parent's prefix
       def get_model_field_names(field_name, field_options, mapped_headers, header_prefix, mapped)
         mapped_model = mapped_headers[field_name] || {}
-        const_get(field_options[:class_name]).get_field_names(
-          mapped_model,
-          header_prefix || field_options[:aliases]&.first,
-          mapped
-        )
+        header_prefix ||= (field_options[:aliases]&.first || field_name).to_s if field_options[:type] == :has_many
+        const_get(field_options[:class_name]).get_field_names(mapped_model, header_prefix, mapped)
       end
 
       def skip_field?(field_name, mapped_headers, mapped)
@@ -72,20 +70,36 @@ module Schema
         const_get(field_options[:class_name]).map_headers_to_attributes(headers, header_prefix)
       end
 
+      # a column already claimed by the parent (or an earlier has_one) isn't reused for a nested field
       def map_headers_to_has_one_associations(headers, mapped_headers, header_prefix)
+        claimed_indexes = mapped_header_indexes(mapped_headers)
         schema.each do |field_name, field_options|
           next unless field_options[:type] == :has_one
           # aliases are matched through the association's own entry
           next if field_options[:alias_of]
 
-          mapped_model = get_mapped_model(field_options, headers, header_prefix)
+          available_headers = headers.each_with_index.map { |header, idx| claimed_indexes.include?(idx) ? nil : header }
+          mapped_model = get_mapped_model(field_options, available_headers, header_prefix)
           next if mapped_model.empty?
 
           mapped_headers[field_name] = mapped_model
+          claimed_indexes.concat(mapped_header_indexes(mapped_model))
         end
         mapped_headers
       end
 
+      def mapped_header_indexes(mapped_headers)
+        mapped_headers.each_value.flat_map do |info|
+          if info.key?(:index)
+            [info[:index]]
+          elsif info.key?(:indexes)
+            info[:indexes]
+          else
+            mapped_header_indexes(info)
+          end
+        end
+      end
+
       def map_headers_to_has_many_associations(headers, mapped_headers)
         schema.each do |field_name, field_options|
           next unless field_options[:type] == :has_many
@@ -104,6 +118,9 @@ module Schema
 
       def map_headers_to_fields(headers, mapped_headers, header_prefix)
         schema.each do |field_name, field_options|
+          # associations are mapped through their nested fields, not a column of their own
+          next if field_options[:association]
+
           if header_prefix
             unless (indexes = find_indexes_for_field(headers, field_options, header_prefix)).empty?
               mapped_headers[field_options[:alias_of] || field_name] = { indexes: indexes }

data/lib/schema/associations/base.rb

@@ -12,12 +12,15 @@ module Schema
       module ClassMethods
         attr_accessor :schema_name
 
+        # named classes are looked up by name so reloaded constants are picked up;
+        # anonymous classes have no name, so they are kept by reference
         def base_schema_class=(kls)
+          @base_schema_class = kls
           @base_schema_class_name = kls.name
         end
 
         def base_schema_class
-          Object.const_get(@base_schema_class_name)
+          @base_schema_class_name ? Object.const_get(@base_schema_class_name) : @base_schema_class
         end
 
         def schema_options

data/lib/schema/version.rb

@@ -2,5 +2,5 @@
 
 module Schema
   # Gem version, bumped by release-please
-  VERSION = '0.7.3'
+  VERSION = '0.7.4'
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: schema-model
 version: !ruby/object:Gem::Version
-  version: 0.7.3
+  version: 0.7.4
 platform: ruby
 authors:
 - Doug Youch
@@ -34,6 +34,7 @@ executables:
 extensions: []
 extra_rdoc_files: []
 files:
+- ".github/dependabot.yml"
 - ".github/workflows/ci.yml"
 - ".github/workflows/release.yml"
 - ".gitignore"