daidai 0.2.0 → 0.2.1.dev.20260725.d207d36

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: ccbe2659c02ddf1d88a941ba8c47a0d6cf621dd241ce7add8b7c2054f56f1baa
-  data.tar.gz: 8599d623feca52dfea34be00ff565b5c6b229a4e9a5204ee62485c41bd4e2a7d
+  metadata.gz: aa525465625e67b90786e617291fd3717d3eda5c0a16aa181dd54482b943bc25
+  data.tar.gz: e3f10ebd8fef9bea3c0d4083252611bfa21d634d3aa10751b319a42f52a39105
 SHA512:
-  metadata.gz: 7ab8e28aed4cc8d3c605a5dbfc8c3fbe5c063de3d33f39a59a293919e90733df98ab6449e7dc82ce084b36198ec02c69644d031225809c8904f5d911d8357a38
-  data.tar.gz: f6aa03749f8de5102010fb8eb6f653355b261a0a2ec86e1f063d7a46257389b6666850e118e8b7b68dca45629b00b8dcca0ff6ab96440ad59ccbdd6abf7c1188
+  metadata.gz: b7790b0b5725fb295f31461ee98558c1f2a7a3b1c08a104bc7903913d7a681ead6159d34c32a4a71803ca699312df6ba4cad986024bc5dcb2e08298404190c7e
+  data.tar.gz: 15f9c6594ba990b47917577133e2610d6e4594d6f196934d93131202d10ddaa1a2b7871f3d23de1ee6eb8ac0afb16e4554cfb00caf31ec7483c5937e31580340

data/CHANGELOG.md

@@ -6,6 +6,16 @@ follow [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
 ## [Unreleased]
 
+## [0.2.1] - 2026-07-25
+
+### Fixed
+
+- `Daidai.conjugate` now produces the correct godan-u negative conditional
+  (買う → 買わなかったら). The upstream JMdictDB `conjo.csv` stores this one cell
+  without the な (買わかったら); `Tables::CONJO_ERRATA` corrects it at load, so the
+  vendored tables stay byte-identical to upstream while every godan-u verb's
+  negative ~tara comes out right.
+
 ## [0.2.0] - 2026-06-27
 
 ### Added

data/lib/daidai/tables.rb

@@ -13,6 +13,21 @@ module Daidai
     # One okurigana rule: how to turn a dictionary form into one conjugation.
     Okurigana = Struct.new(:stem, :okuri, :euphr, :euphk, keyword_init: true)
 
+    # Corrections for known errors in the upstream JMdictDB tables, applied at
+    # load so the vendored CSVs stay byte-identical to upstream (keeping
+    # `rake daidai:check_resources` green) while daidai still emits the right
+    # form. Keyed by JMdict code + [conj_id, negative?, polite?, onum] so a
+    # pos-id renumber upstream can't silently misapply the fix; values override
+    # only the given Okurigana fields.
+    #
+    #   • v5u negative conditional (買う → 買わなかったら): upstream conjo.csv stores
+    #     "わかったら", dropping the な. Every other class is correct here (v5u-s
+    #     even has わなかったら), so this one cell is a typo. Without it every
+    #     godan-u verb's negative ~tara comes out as 買わかったら / 使わかったら / ….
+    CONJO_ERRATA = {
+      [ "v5u", 11, true, false, 1 ] => { okuri: "わなかったら" }
+    }.freeze
+
     class << self
       # conjugation id (Integer) => human name ("Past (~ta)", …)
       def conj
@@ -21,14 +36,17 @@ module Daidai
 
       # [pos_id, conj_id, negative?, polite?, onum] => Okurigana
       def conjo
-        @conjo ||= read("conjo.csv").each_with_object({}) do |r, table|
-          key = [ r["pos"].to_i, r["conj"].to_i, r["neg"] == "t", r["fml"] == "t", r["onum"].to_i ]
-          table[key] = Okurigana.new(
-            stem: r["stem"].to_i,
-            okuri: r["okuri"].to_s,
-            euphr: presence(r["euphr"]),
-            euphk: presence(r["euphk"])
-          )
+        @conjo ||= begin
+          table = read("conjo.csv").each_with_object({}) do |r, acc|
+            key = [ r["pos"].to_i, r["conj"].to_i, r["neg"] == "t", r["fml"] == "t", r["onum"].to_i ]
+            acc[key] = Okurigana.new(
+              stem: r["stem"].to_i,
+              okuri: r["okuri"].to_s,
+              euphr: presence(r["euphr"]),
+              euphk: presence(r["euphk"])
+            )
+          end
+          apply_conjo_errata(table)
         end
       end
 
@@ -43,6 +61,24 @@ module Daidai
 
       private
 
+      # Overlay CONJO_ERRATA onto the freshly-loaded table, resolving each JMdict
+      # code to its pos id. A correction that no longer matches a row (upstream
+      # fixed it, or restructured) is skipped silently — the errata is advisory.
+      def apply_conjo_errata(table)
+        CONJO_ERRATA.each do |(code, conj, neg, fml, onum), fix|
+          pos_id = pos_ids[code] or next
+          key = [ pos_id, conj, neg, fml, onum ]
+          row = table[key] or next
+          table[key] = Okurigana.new(
+            stem: fix.fetch(:stem, row.stem),
+            okuri: fix.fetch(:okuri, row.okuri),
+            euphr: fix.fetch(:euphr, row.euphr),
+            euphk: fix.fetch(:euphk, row.euphk)
+          )
+        end
+        table
+      end
+
       def read(file, headers: true)
         CSV.read(File.join(DIR, file), col_sep: "\t", headers: headers, quote_char: '"')
       end

data/lib/daidai/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module Daidai
-  VERSION = "0.2.0"
+  VERSION = "0.2.1.dev.20260725.d207d36"
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: daidai
 version: !ruby/object:Gem::Version
-  version: 0.2.0
+  version: 0.2.1.dev.20260725.d207d36
 platform: ruby
 authors:
 - davafons
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2026-06-27 00:00:00.000000000 Z
+date: 2026-07-25 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: csv