relaton 3.0.7 → 3.1.0

data/lib/relaton/iala.rb

@@ -1,7 +1,7 @@
 require "pubid"
-require "relaton/bib"
-require "relaton/index"
-require "relaton/version"
+require_relative "bib"
+require_relative "index"
+require_relative "version"
 
 # Relaton::Iala is the IALA (International Organization for Marine Aids
 # to Navigation) flavor of the unified relaton gem. Adds typed Ext fields

data/lib/relaton/iana/data_fetcher.rb

@@ -1,4 +1,4 @@
-require "relaton/core"
+require_relative "../core"
 require_relative "../iana"
 require_relative "parser"
 

data/lib/relaton/iana/processor.rb

@@ -1,4 +1,4 @@
-require "relaton/core/processor"
+require_relative "../core/processor"
 
 module Relaton
   module Iana

data/lib/relaton/iana.rb

@@ -1,9 +1,9 @@
 # frozen_string_literal: true
 
 require "faraday"
-require "relaton/index"
+require_relative "index"
 require "pubid"
-require "relaton/bib"
+require_relative "bib"
 require_relative "version"
 require_relative "iana/util"
 require_relative "iana/item_data"

data/lib/relaton/iec/data_fetcher.rb

@@ -1,5 +1,5 @@
 require "yeptris"
-require "relaton/core"
+require_relative "../core"
 require_relative "../iec"
 require_relative "data_parser"
 

data/lib/relaton/iec/processor.rb

@@ -1,4 +1,4 @@
-require "relaton/core/processor"
+require_relative "../core/processor"
 
 module Relaton
   module Iec

data/lib/relaton/iec.rb

@@ -3,8 +3,8 @@ require "net/http"
 require "moxml"
 require "pubid"
 require "zip"
-require "relaton/index"
-require "relaton/iso"
+require_relative "index"
+require_relative "iso"
 require_relative "version"
 require_relative "iec/util"
 require_relative "iec/bibliography"

data/lib/relaton/ieee/processor.rb

@@ -1,4 +1,4 @@
-require "relaton/core/processor"
+require_relative "../core/processor"
 
 module Relaton
   module Ieee

data/lib/relaton/ieee.rb

@@ -2,8 +2,8 @@ require "digest/md5"
 require "faraday"
 require "yaml"
 require "pubid"
-require "relaton/index"
-require "relaton/bib"
+require_relative "index"
+require_relative "bib"
 require_relative "version"
 require_relative "ieee/util"
 require_relative "ieee/bibliography"

data/lib/relaton/ietf/CLAUDE.md

@@ -375,7 +375,7 @@ and the affected files come through with their real characters —
 - **An oracle for `read_bibxml` must read binary.** The method is byte-verbatim (`File.binread`), so comparing it against `File.read(path, encoding: "UTF-8")` compares two different strings **on Windows only**: Ruby opens a file in text mode there and collapses CRLF to LF. The repo declares no `.gitattributes` and the GitHub Windows runner image defaults to `core.autocrlf=true`, so a fixture stored with LF is checked out CRLF — which failed all three `windows-latest` jobs while every Linux and macOS job passed. Use `File.read(path, mode: "rb", encoding: "UTF-8")`. macOS cannot reproduce this by checking the fixture out as CRLF, because its text-mode read does not translate; emulate Windows explicitly with `File.read(path, encoding: "UTF-8", universal_newline: true)`. The two `keeps CRLF line endings verbatim` examples pin the contract with `binwrite` temp files, independent of the checkout — one per branch, since the CP1252 repair rebuilds the string run by run and is where a line ending is likeliest to be lost. Note that a byte-equality oracle alone cannot catch a dropped `force_encoding`: `rfc.xml` is pure ASCII, and Ruby compares two ASCII-only strings as equal across encodings, so the example asserts `encoding` separately.
 - **Schema validation** via `ruby-jing` against the shared RNG grammars in the repo-root `grammar/` (`Jing.new "../../grammar/relaton-ietf-compile.rng"`, relative to the suite's CWD `spec/ietf/`)
 - **Shared examples** for org/person parsing in `bibxml_parser_spec.rb` (`parse_org`, `parse_person`)
-- **`equivalent-xml`** gem used for XML comparison (`be_equivalent_to` matcher)
+- **`canon`** gem used for XML comparison (`be_equivalent_to` matcher)
 
 ## RuboCop
 

data/lib/relaton/ietf/data_fetcher.rb

@@ -2,7 +2,7 @@ require "etc"
 require "parallel"
 require "pubid"
 require "pubid/ietf"
-require "relaton/core"
+require_relative "../core"
 require_relative "../ietf"
 require_relative "bibxml_parser"
 require_relative "rfc/index"

data/lib/relaton/ietf/processor.rb

@@ -1,4 +1,4 @@
-require "relaton/core/processor"
+require_relative "../core/processor"
 
 module Relaton
   module Ietf

data/lib/relaton/ietf.rb

@@ -1,8 +1,8 @@
 # frozen_string_literal: true
 
 require "net/http"
-require "relaton/index"
-require "relaton/bib"
+require_relative "index"
+require_relative "bib"
 require_relative "version"
 require_relative "ietf/util"
 require_relative "ietf/item_data"

data/lib/relaton/iho/CLAUDE.md

@@ -55,11 +55,11 @@ In `item.rb`, `require_relative "relation"` is placed **after** the `Item` class
 
 `Bibliography.search`/`.get` fetches YAML from the relaton-data-iho GitHub repo via `Relaton::Index`, then deserializes into model objects.
 
-**Editions and series.** The pubid `version` is the edition (`IHO B-12 3.0.0`; the grammar takes `a.b.c` only). A reference with no version answers with the **latest** edition: `recency_key` compares the segments as integers (`2.10.0` after `2.9.0`). It used to be `min_by` on the string, which returned the oldest.
+**Editions and series (the pubid 2.0.5 component model, relaton#440).** The pubid attribute is `edition` (dotted `a.b.c`, `IHO B-12 3.0.0`; the pre-440 name `version` is gone). A reference with no edition answers with the **latest**: `recency_key` compares the segments as integers (`2.10.0` after `2.9.0`). It used to be `min_by` on the string, which returned the oldest.
 
-**Series are identified by year or volume, not an edition (relaton#407).** B-7, P-5 and P-6 carry a `year` (`IHO P-5 2024`), the IHR (P-1) a `volume` with an optional `issue` (`IHO P-1 Vol. 30 (2)`) — plain scalars added by pubid#635, mutually exclusive with `version` (the grammar's tail picks exactly one). The same `recency_key` orders each kind numerically, so a bare series query answers with the latest year/volume/issue. Matching (`pubid_match?`) treats a component the query omits as a wildcard: `IHO P-1 Vol. 30` matches any issue of volume 30. The comparison runs through pubid's `#exclude` — excluding `:volume` nils the `issue` with it (the pair is one scalar cluster) — the way BIPM's `stem_keys` does. The subdivisions (part/appendix/annex/supplement) stay strict: this flavor cannot use the `===` subset match, because pubid does not declare `subset_strict` for IHO's subdivisions and a nil `part` would read as "any part" (`IHO S-100` would answer with a part row).
+**Serial locators are volumes (relaton#440 unified the year locators of relaton#407).** The yearbook year IS a volume: `IHO P-5 Vol. 2020` is canonical, the bare-year faces (`P-5 2020`, `P-5/2020`) parse as legacy aliases into the same `volume`, and the IHR is `P-1 Vol. 30` with `Issue 1` (no parens) or the short slash form `P-1/30`. A bare series query answers with the latest volume/issue. **Components can publish separately and carry their own edition:** `IHO S-66 Suppl 1 2.0.0` gives `supplement_edition`, `S-100 Part S-102 1.0.0` gives `part_edition`, annexes likewise — and the publication years in parens (`S-52 Annex A 4.0.3 (2021)`) are `*_year` metadata, not identity: a stated year pins the record, an omitted one wildcards. Matching (`pubid_match?`) treats a discriminator the query omits as a wildcard: `IHO P-1 Vol. 30` matches any issue of volume 30. The comparison runs through pubid's `#exclude` — excluding `:volume` nils the `issue`, excluding an `:appendix` nils its `appendix_subnumber` (scalar clusters) — the way BIPM's `stem_keys` does. The subdivisions (part/appendix/annex/supplement) stay strict: this flavor cannot use the `===` subset match, because pubid does not declare `subset_strict` for IHO's subdivisions and a nil `part` would read as "any part" (`IHO S-100` would answer with a part row).
 
-**The edition reaches the index only through the committed row.** A record prints `B-12` and keeps its edition in `edition:`, so `relaton index` cannot read a version from the docid. `MachineIndex#refine` (relaton-cli) takes the committed `index-v3.yaml` row when it adds only a key in `REFINABLE_KEYS` to the parse — `version`/`edition`/`volume`/`year`/`issue` (relaton#407 added the year/issue pair for the series rows). Before that, the Pages index (generated 2026-10-29) had no `version` in any of its 325 rows, so every versioned reference answered not-found (relaton-data-iho#20). The suite did not see it: `spec/fixtures/index-v3.zip` is the committed index, which has versions. The suite's fixture currently carries the **re-keyed** series rows (year/volume) ahead of the data repo: the placeholder `1.0.0` rows for p-1/p-1-2x/p-5/p-6/b-7 are replaced by `year`/`volume` rows — refresh the fixture from the live published zip once relaton-data-iho re-keys, and reconcile any file-name drift then.
+**The edition reaches the index only through the committed row.** A record prints `B-12` and keeps its edition in `edition:`, so `relaton index` cannot read a version from the docid. `MachineIndex#refine` (relaton-cli) takes the committed `index-v3.yaml` row when it adds only a key in `REFINABLE_KEYS` to the parse — `version`/`edition`/`volume`/`year`/`issue` (relaton#407 added the year/issue pair for the series rows). Before that, the Pages index (generated 2026-10-29) had no `version` in any of its 325 rows, so every versioned reference answered not-found (relaton-data-iho#20). The suite did not see it: `spec/fixtures/index-v3.zip` is the committed index, which has versions. The suite's fixture carries the #48-era re-keyed series rows migrated to the #440 keys (`version`→`edition`/`<component>_edition`, year locators→`volume`; `support/migrate_index_fixture.rb` did the one-shot migration) — refresh the fixture from the live published zip once relaton-data-iho round 2 republishes, and reconcile any file-name drift then.
 
 ### Test Infrastructure
 

data/lib/relaton/iho/bibliography.rb

@@ -96,31 +96,37 @@ module Relaton
           )
         end
 
-        # The sort key for the latest record in a bucket (relaton#407). Each
-        # IHO identity carries exactly one trailing discriminator — a dotted
-        # `version` (an edition, `B-12 3.0.0`), a series `year` (`P-5 2024`)
-        # or a series `volume` with its `issue` (`P-1 Vol. 30 (2)`) — and the
-        # key orders by that component, numerically: version segments compare
-        # as integers (2.10.0 is later than 2.9.0), as do year, volume and
-        # issue. A row with no discriminator at all is the oldest. Rows in
-        # one bucket share the discriminator kind once the data is keyed;
-        # while kinds coexist the array compare is deterministic, and a
-        # year's bare [2024] beats a placeholder edition's [1, 0, 0].
+        # The sort key for the latest record in a bucket (relaton#407,
+        # reshaped by relaton#440's component model). Volume-locator series
+        # (P-1, P-5) carry a `volume` with an optional `issue` — the
+        # yearbook year IS a volume (`P-5 Vol. 2020`; the bare-year faces
+        # of the #48 re-key parse as legacy aliases), so the year locators
+        # fold into the volume ordering. Edition-keyed products carry a
+        # dotted `edition` (`B-12 3.0.0`), and a component published
+        # separately carries its own edition (`S-66 Suppl 1 2.0.0` →
+        # `supplement_edition`). Every component compares numerically
+        # segment by segment (2.10.0 after 2.9.0); a row with no
+        # discriminator at all is the oldest.
         def recency_key(id)
-          if id.year
-            [id.year.to_i]
-          elsif id.volume
+          if id.volume
             [id.volume.to_i, id.issue.to_i]
           else
-            id.version.to_s.split(".").map(&:to_i)
+            ed = id.edition || id.supplement_edition || id.part_edition ||
+              id.annex_edition || id.appendix_edition || id.volume_edition
+            ed.to_s.split(".").map(&:to_i)
           end
         end
 
-        # The components that make an IHO document the one it is beyond the
-        # series and subdivisions: the edition (`version`), or — for the
-        # series identified by year or volume instead (B-7, P-1, P-5, P-6;
-        # relaton#407) — the `year` or the `volume`/`issue` pair.
-        DISCRIMINATOR_KEYS = %i[version year volume issue].freeze
+        # The identity components beyond the series and subdivisions
+        # (relaton#440): the base `edition` — or the component-level
+        # edition when the component publishes separately — plus the
+        # volume/issue pair. The publication years (`year`, `*_year`) are
+        # metadata, not identity: a stated year pins the record, an
+        # omitted one wildcards it.
+        DISCRIMINATOR_KEYS = %i[
+          edition volume issue
+          supplement_edition part_edition annex_edition appendix_edition volume_edition
+        ].freeze
 
         # Both `row_id` and `query` are Pubid::Iho::Identifiers::* instances.
         # Class identity pins the series (S/P/M/B/C). Subdivision keys
@@ -132,21 +138,29 @@ module Relaton
         # so `IHO S-100` would answer with a part row.
         #
         # The discriminator keys stay nil-tolerant: an unqualified `IHO B-11`
-        # finds the latest edition, an unqualified `IHO P-5` the latest year,
-        # and `IHO P-1 Vol. 30` any issue of that volume — while a stated
-        # component must match exactly. The comparison runs through pubid's
-        # `#exclude` (the primitive `matches?` is built on), so the rules the
-        # flavor declares there ride along: excluding `:volume` nils the
-        # `issue` with it. `stem_keys` in the BIPM flavor is the same shape.
+        # finds the latest edition, an unqualified `IHO P-5` the latest
+        # volume (the yearbook year — relaton#440 folded the year
+        # locators in), and `IHO P-1 Vol. 30` any issue of that volume —
+        # while a stated component must match exactly. The comparison runs
+        # through pubid's `#exclude` (the primitive `matches?` is built
+        # on), so the rules the flavor declares there ride along: excluding
+        # `:volume` nils the `issue` with it, and excluding an appendix
+        # nils its subnumber — the scalar clusters are one unit. `stem_keys`
+        # in the BIPM flavor is the same shape.
         def pubid_match?(row_id, query)
           row_id.class == query.class &&
             row_id.exclude(*wildcards(query)) == query.exclude(*wildcards(query))
         end
 
         # The discriminators the query leaves unset: wildcards for the
-        # search, significant whenever the query states one.
+        # search, significant whenever the query states one — plus the
+        # publication years, which are stated-metadata: pinned when the
+        # query carries them, wildcarded when it does not.
         def wildcards(query)
-          DISCRIMINATOR_KEYS.select { |key| query.public_send(key).nil? }
+          keys = DISCRIMINATOR_KEYS + %i[
+            year supplement_year part_year annex_year appendix_year volume_year
+          ]
+          keys.select { |key| query.public_send(key).nil? }
         end
       end
     end

data/lib/relaton/iho/hash_parser_v1.rb

@@ -1,4 +1,4 @@
-require "relaton/bib/hash_parser_v1"
+require_relative "../bib/hash_parser_v1"
 require_relative "../iho"
 
 module Relaton

data/lib/relaton/iho/processor.rb

@@ -1,4 +1,4 @@
-require "relaton/core/processor"
+require_relative "../core/processor"
 
 module Relaton
   module Iho
@@ -26,9 +26,10 @@ module Relaton
       end
 
       # An IHO record prints `B-12` and keeps its edition in `edition`, so
-      # the edition goes on the key as the version (relaton#408).
+      # the edition goes on the key (relaton#408). The component model
+      # (relaton#440) renamed the pubid attribute `version` -> `edition`.
       def item_cache_key(pubid, item)
-        with_components pubid, version: item.edition&.content
+        with_components pubid, edition: item.edition&.content
       end
 
       # @param xml [String]

data/lib/relaton/iho.rb

@@ -1,5 +1,5 @@
-require "relaton/bib"
-require "relaton/index"
+require_relative "bib"
+require_relative "index"
 require "pubid"
 require_relative "version"
 require_relative "iho/util"

data/lib/relaton/index.rb

@@ -3,7 +3,7 @@
 require "yaml"
 require "zip"
 require "open-uri"
-require "relaton/logger"
+require_relative "logger"
 
 require_relative "version"
 require_relative "core/request_error"

data/lib/relaton/isbn/processor.rb

@@ -1,4 +1,4 @@
-require "relaton/core/processor"
+require_relative "../core/processor"
 
 module Relaton
   module Isbn

data/lib/relaton/isbn.rb

@@ -1,5 +1,5 @@
 require "net/http"
-require "relaton/bib"
+require_relative "bib"
 require_relative "version"
 require_relative "isbn/util"
 require_relative "isbn/isbn"

data/lib/relaton/iso/hash_parser_v1.rb

@@ -1,4 +1,4 @@
-require "relaton/bib/hash_parser_v1"
+require_relative "../bib/hash_parser_v1"
 require_relative "../iso"
 
 module Relaton

data/lib/relaton/iso/processor.rb

@@ -1,4 +1,4 @@
-require "relaton/core/processor"
+require_relative "../core/processor"
 
 module Relaton
   module Iso

data/lib/relaton/iso.rb

@@ -2,10 +2,10 @@
 
 require "net/http"
 require "pubid"
-require "relaton/index"
+require_relative "index"
 require "isoics"
-require "relaton/bib"
-require "relaton/core"
+require_relative "bib"
+require_relative "core"
 require_relative "version"
 require_relative "iso/util"
 require_relative "iso/item_data"

data/lib/relaton/itu/processor.rb

@@ -1,4 +1,4 @@
-require "relaton/core/processor"
+require_relative "../core/processor"
 
 module Relaton
   module Itu

data/lib/relaton/itu.rb

@@ -1,9 +1,9 @@
 require "mechanize"
 require "pubid"
 require "digest/md5"
-require "relaton/index"
-require "relaton/bib"
-require "relaton/core"
+require_relative "index"
+require_relative "bib"
+require_relative "core"
 require_relative "version"
 require_relative "itu/util"
 require_relative "itu/item_data"

data/lib/relaton/jcgm/processor.rb

@@ -1,4 +1,4 @@
-require "relaton/core/processor"
+require_relative "../core/processor"
 
 module Relaton
   module Jcgm

data/lib/relaton/jcgm.rb

@@ -1,5 +1,5 @@
-require "relaton/bib"
-require "relaton/index"
+require_relative "bib"
+require_relative "index"
 require "pubid"
 require_relative "version"
 require_relative "jcgm/util"

data/lib/relaton/jis/processor.rb

@@ -1,6 +1,6 @@
 # frozen_string_literal: true
 
-require "relaton/core/processor"
+require_relative "../core/processor"
 
 module Relaton
   module Jis

data/lib/relaton/jis.rb

@@ -3,8 +3,8 @@
 require "net/http"
 require "mechanize"
 require "pubid"
-require "relaton/iso"
-require "relaton/index"
+require_relative "iso"
+require_relative "index"
 require_relative "version"
 require_relative "jis/util"
 require_relative "jis/item"

data/lib/relaton/nist/CLAUDE.md

@@ -69,7 +69,8 @@ All data models inherit from `Lutaml::Model::Serializable` and use declarative a
   - **`#write_file` serializes before it indexes**, so a record that fails to serialize leaves no index row naming a file that was never written. That is how a `Bib::Date` in `Scraper.fetch_dates` hid: `Nist::ItemData` takes `Nist::Date`, and YAML serialization rejects the parent class (the same strictness as the docidentifier note below).
 - `ModsParser` — maps MODS XML (via `loc_mods`) to `ItemData`
 - `Scraper` — fetches items from GitHub YAML or CSRC JSON
-- `PubsExport` — singleton; caches CSRC pubs-export zip with thread-safe daily updates
+- `PubsExport` — singleton; caches CSRC pubs-export zip with thread-safe daily updates. The cache file is `#datafile`, under `Relaton::Index.config.storage_dir` (default `Dir.home`; the `DATAFILE` constant is only that default).
+  - **A read-only file system must not stop a search (relaton-nist#57).** When the cache write fails, `#fetch_data` keeps the downloaded zip in `@buffer`, and `#unzip` reads it with `Zip::File.open_buffer`. In rubyzip 3, `open_buffer` returns the buffer, not the block value, so collect the rows in a local. The cache write goes to a temp file and is renamed into place (`#write_datafile`), so a failed write never leaves a truncated zip that later processes read as fresh. `HitCollection#search` also rescues the local-cache errors in `LOCAL_DATA_ERRORS` from `from_json` (`#from_json_or_empty`), logs a warning, and continues with `from_ga`. Do not widen that list to `SystemCallError`: `Errno::ECONNREFUSED` is one, and a CSRC outage would then answer "not found" from the index, which `Relaton::Db` can cache. Specs assert the warning with `output(...).to_stderr_from_any_process`. A `receive(:warn)` stub on `Util` is never hit: `Kernel#warn` is private, the stub stays private, and the explicit-receiver call goes to `method_missing`.
 
 ### Pubid-backed docidentifier
 
@@ -115,6 +116,15 @@ NIST-specific gotchas (why this isn't a straight copy of IEC/CCSDS):
   numbered editions/revisions (`e2`, `r5`). Dates that ride the `update`
   component (some NBS supplements) are left intact — `update` also encodes
   non-date update codes (`/Upd2`), so stripping it would corrupt identity.
+- **The item's `edition` comes from the DOI (relaton-nist#56).**
+  `ModsParser#parse_edition` sets `<edition number="N">N</edition>` only for a
+  numbered `e<N>` edition in the DOI (`NBS.CIRC.111e2` → 2, `11e2.1915` → 2):
+  250 of the 20,392 live MODS records. An `e<year>` edition is a date, and an
+  `r<N>` revision is not an edition, so both stay without one. The MODS
+  `<edition>` element is prose (`New version`, `Erratum`), never a number. The
+  issue's "edition 0 → 1" rule came from the old Crossref feed; it is kept for
+  a DOI `e0` (none in the live data). A DOI pubid cannot parse gives no
+  edition, not an error.
 - **No `(all parts)` marker in `content`.** `content` is live-derived from
   `@pubid`, so `to_all_parts!` freezes the already-stripped rendering into
   `@raw_content` before wrapping `@pubid` in pubid's `AllParts` — matching

data/lib/relaton/nist/README.adoc

@@ -311,7 +311,13 @@ Relaton::Nist::Item.from_yaml File.read("spec/fixtures/item.yaml")
 
 === Fetching data
 
-On a search, the CSRC data sources are fetched and stored in the user's cache.
+On a search, the CSRC data sources are fetched and stored in the user's cache,
+`~/.relaton/nist/pubs-export.zip`. The cache follows
+`Relaton::Index.config.storage_dir`, so one setting moves it together with the
+index cache. When the cache cannot be written (for example on a read-only file
+system such as AWS Lambda), the data stays in memory for the process. When the
+CSRC data cannot be used at all, the search continues with the
+`relaton-data-nist` index.
 
 This gem uses the https://github.com/usnistgov/NIST-Tech-Pubs/releases/download/May2024/allrecords-MODS.xml dataset as NIST-Tech-Pubs data source.
 

data/lib/relaton/nist/hit_collection.rb

@@ -12,6 +12,12 @@ module Relaton
 
       GHNISTDATA = "https://raw.githubusercontent.com/relaton/relaton-data-nist/v2/"
 
+      # Errors of the local CSRC cache (read-only or full disk, a bad zip).
+      LOCAL_DATA_ERRORS = [
+        Errno::EROFS, Errno::EACCES, Errno::EPERM, Errno::ENOSPC,
+        Errno::ENOENT, Zip::Error
+      ].freeze
+
       attr_reader :reference
       attr_accessor :array
 
@@ -47,7 +53,7 @@ module Relaton
       # @return [Relaton::Nist::HitCollection] hits collection
       #
       def search
-        @array = from_json
+        @array = from_json_or_empty
         @array = from_ga unless @array.any?
         sort_hits!
       end
@@ -400,6 +406,24 @@ module Relaton
         select_data.map { |h| Hit.new(hit_data(h), self) }
       end
 
+      #
+      # The CSRC hits, or none when the local CSRC data cannot be used: a
+      # file-system error must not stop the search, because the
+      # relaton-data-nist index can still answer (relaton-nist#57). Network
+      # errors are not rescued: they still reach Bibliography.search, which
+      # raises Relaton::RequestError, so a CSRC outage is not cached as
+      # "not found".
+      #
+      # @return [Array<Relaton::Nist::Hit>] hits
+      #
+      def from_json_or_empty
+        from_json
+      rescue *LOCAL_DATA_ERRORS => e
+        Util.warn "CSRC pubs-export data is not available (#{e.message}); " \
+                  "searching the Relaton repository only", key: @reference
+        []
+      end
+
       def pubs_export_id(json)
         self.class.pubs_export_id json
       end

data/lib/relaton/nist/mods_parser.rb

@@ -13,7 +13,7 @@ module Relaton
       }.freeze
 
       ATTRS = %i[type docidentifier title source abstract date contributor
-                  relation place series].freeze
+                  relation place series edition].freeze
 
       def initialize(doc, series, errors = {})
         @doc = doc
@@ -51,17 +51,41 @@ module Relaton
       def get_id_from_str(str)
         return if str.nil? || str.empty?
 
-        # DOIs arrive dotted ("NIST.HB.135e2022-upd1"); that MR-style form
-        # parses through a path that skips UpdateCodes date lookup and drops
-        # letter suffixes. Space the leading "PUB.SERIES." prefix so it parses
-        # like the canonical short form (only the separator dots are touched,
-        # not any inside the number). Then force :human rendering.
-        spaced = str.sub(/\A(NIST|NBS)\.([A-Z][A-Za-z]*)\./, '\1 \2 ')
-        ::Pubid::Nist::Identifier.parse(spaced).to_s(format: :human)
+        parse_pubid(str).to_s(format: :human)
       rescue ::Pubid::Errors::ParseError
         str.gsub(".", " ").sub(/^[\D]+/, &:upcase)
       end
 
+      # DOIs arrive dotted ("NIST.HB.135e2022-upd1"); that MR-style form
+      # parses through a path that skips UpdateCodes date lookup and drops
+      # letter suffixes. Space the leading "PUB.SERIES." prefix so it parses
+      # like the canonical short form (only the separator dots are touched,
+      # not any inside the number).
+      def parse_pubid(str)
+        spaced = str.sub(/\A(NIST|NBS)\.([A-Z][A-Za-z]*)\./, '\1 \2 ')
+        ::Pubid::Nist::Identifier.parse(spaced)
+      end
+
+      # The edition number a DOI states ("NBS.CIRC.111e2" -> 2), relaton-nist#56.
+      # The MODS carries no edition number of its own (its <edition> is prose:
+      # "New version", "Erratum"), so the DOI is the only source. Only a
+      # numbered "e<N>" edition counts: an "e<year>" edition is a date, and an
+      # "r<N>" revision is not an edition. An edition 0 counts as 1 (the issue's
+      # second rule; no live DOI has one).
+      # @return [Bib::Edition, nil]
+      def parse_edition
+        doi = parse_doi
+        return if doi.nil? || doi.empty?
+
+        edition = parse_pubid(doi).edition
+        return unless edition&.type == "e" && edition.id.to_s.match?(/\A\d{1,3}\z/)
+
+        number = edition.id.to_i.zero? ? "1" : edition.id
+        Bib::Edition.new(number: number, content: number)
+      rescue ::Pubid::Errors::Error, ArgumentError
+        nil
+      end
+
       # @return [String]
       def replace_wrong_doi(id)
         case id

data/lib/relaton/nist/processor.rb

@@ -1,4 +1,4 @@
-require "relaton/core/processor"
+require_relative "../core/processor"
 
 module Relaton
   module Nist