asciichem 0.28.0 → 0.28.2

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
checksums.yaml CHANGED
@@ -1,7 +1,7 @@
1
1
  ---
2
2
  SHA256:
3
- metadata.gz: 8fb92ff1cd35eeb546372b2ac9c3cf97b908971409e9a1d473b575479464e356
4
- data.tar.gz: 2cc3be9f85b7ca2d7e3fe773e981073b6a755344cc6b593b1940666653fb52e7
3
+ metadata.gz: 9e4646420eb8fdde22a218b2f65d6b4dce4661264124f98d1fd4716eb82a7faf
4
+ data.tar.gz: 1dd763180c8466a8c0245e0147b4f79201f8e40957586a8392827d8bf21fea11
5
5
  SHA512:
6
- metadata.gz: 6d2f065122017d83a4e0fdb5a9da9db50a9dbd321e335a9f49bce5c3d253e78ae11f7ae33899ad4ce60fa28b9053a085f4cac7acefbd22999c88287fc0d38d2d
7
- data.tar.gz: 369b73e26be9c6234fd36c6a2e5d14b53a8746a5fa44c7bc5949297cc17978e6a135e4f003a90641ad7e63a2ce4dd400e05107bb79157628028f14bc68104ac2
6
+ metadata.gz: 9783079bd787334d96149f8f7826bb80c0314aec5b9a4d45ba7be96b6b4a52ff43bce3f962c273c6bdbf2f6d78a4721a011f32a5ef2897a4fbcd99af5f074ef9
7
+ data.tar.gz: 2edf62206b51e7fa1cf97181976a19b8903f00bb4d5cd9cfbffc1c1a5750dfa0bb60d3f356f8b4031ce17cca1f142f147268f2aaa3ddb65ee8459c4c4130f2f8
@@ -16,6 +16,10 @@ jobs:
16
16
  - uses: actions/checkout@v4
17
17
  - name: Clone conformance corpus (asciichem-tests)
18
18
  run: git clone --depth 1 https://github.com/asciichem/asciichem-tests.git ../asciichem-tests
19
+ - name: Record corpus version for the conformance report
20
+ run: |
21
+ git -C ../asciichem-tests fetch --depth 1 --tags --quiet
22
+ echo "ASCIICHEM_CORPUS_VERSION=$(git -C ../asciichem-tests describe --tags --abbrev=0 2>/dev/null || echo main)" >> "$GITHUB_ENV"
19
23
  - uses: ruby/setup-ruby@v1
20
24
  with:
21
25
  ruby-version: ${{ matrix.ruby }}
data/CHANGELOG.md CHANGED
@@ -3,6 +3,22 @@
3
3
  All notable changes to AsciiChem are documented here.
4
4
  This project follows [Semantic Versioning](https://semver.org/).
5
5
 
6
+ ## [0.28.2] - 2026-09-14
7
+
8
+ ### Fixed
9
+ - Conformance reports now name the corpus version CI actually ran
10
+ (derived from the cloned asciichem-tests tag) instead of a stale
11
+ default. Pairs with asciichem-tests v0.4.0 (MathML golden suite).
12
+
13
+ ## [0.28.1] - 2026-09-14
14
+
15
+ ### Fixed
16
+ - Cascade transform is engine-agnostic: CascadeBuilder#canonicalise_hash
17
+ wraps scalar arrow/products values instead of Array(hash), which
18
+ enumerates a Hash. The parsanol 1.3.13 re-check (221/221 corpus-green
19
+ through the unmodified grammar) surfaced the shape; parslet is
20
+ unaffected. Spec'd for all three engine shapes.
21
+
6
22
  ## [0.28.0] - 2026-09-14
7
23
 
8
24
  ### Changed
data/benchmarks/README.md CHANGED
@@ -15,6 +15,7 @@ WORKLOAD = ["H_2O", "Ca^2+", "SO_4^2-", "(R)-CH_3CH(OH)COOH",
15
15
  |---|---|---|---|
16
16
  | Ruby (parslet), 3.4.8 arm64 | 29.1 ms | ~2.9 ms | `bundle exec ruby benchmarks/engines.rb` |
17
17
  | Ruby + parse+text | 35.3 ms | ~3.5 ms | round-trip adds the formatter |
18
+ | Ruby (parsanol compat, :ruby), 1.3.13 | 19.9 ms | ~2.0 ms | `benchmarks/parsanol_recheck.rb`, same session as the 11.8 ms parslet baseline below |
18
19
  | TypeScript (peggy), Node 24 | 0.19 ms | ~19 µs | `npm run bench` (asciichem-ts) |
19
20
  | Python (RD), 3.10 | 3.94 ms | ~394 µs | `python benchmarks/engines.py` (asciichem-py) |
20
21
 
@@ -54,3 +55,33 @@ upstream in parsanol-ruby#25):
54
55
  Revisit trigger unchanged: engage the native backend for full
55
56
  grammars, fix the repetition-termination bug, and beat parslet on
56
57
  this workload — then re-run the corpus against the port.
58
+
59
+ ### Re-check (2026-09-14, parsanol 1.3.13)
60
+
61
+ The upstream "native by default" + RepetitionTag work landed, so the
62
+ revisit trigger was tested (`benchmarks/parsanol_recheck.rb`):
63
+
64
+ 1. **The repetition-termination bug is fixed.** `SO_4^2-` parses and
65
+ round-trips, and the unmodified grammar passes the entire shared
66
+ corpus through the `Parsanol::Parslet` compat layer —
67
+ **221/221** parse/reject/round-trip cases identical to parslet.
68
+ One divergence surfaced on our side and is fixed in the gem:
69
+ parsanol's transform delivers cascade tail segments as scalar
70
+ hashes, and `CascadeBuilder#canonicalise_hash` used `Array(hash)`
71
+ (which enumerates a Hash instead of wrapping it) — now wraps
72
+ explicitly, engine-agnostic.
73
+ 2. **Perf (compat, :ruby forced):** 19.9 ms vs 11.8 ms per 10-input
74
+ pass in the same session — **~1.7x slower than parslet** (down
75
+ from ~6x in the first investigation). Correct but not a win.
76
+ 3. **Native still cannot serialize full parslet grammars** — two
77
+ upstream bugs (reported in parsanol-ruby#25):
78
+ `native.rb` never loads `native/dynamic` (NameError silently
79
+ falls back to :ruby), and `Dynamic.register` never increments
80
+ `@next_id`, so the second lazily-bound rule panics the Rust core
81
+ with "callback ID 1000000 is already registered".
82
+
83
+ **Verdict: still not adopted — but one small upstream fix away from
84
+ a meaningful re-measure.** Corpus correctness is already there; the
85
+ native path is the whole point and remains unmeasurable until
86
+ serialization survives a multi-rule grammar.
87
+
@@ -0,0 +1,79 @@
1
+ # frozen_string_literal: true
2
+
3
+ # Parsanol re-check (parsanol-ruby 1.3.13, post-issue-25): runs the
4
+ # UNMODIFIED AsciiChem grammar over the Parsanol engine via the
5
+ # Parslet compat shim, then (1) gates on the shared corpus, (2) gates
6
+ # on the issue-25 EOF repro, (3) measures against the parslet path.
7
+ #
8
+ # Native mode cannot serialize this grammar today (two upstream bugs:
9
+ # native.rb never loads native/dynamic; Dynamic.register never
10
+ # increments @next_id so the second callback panics the Rust core —
11
+ # parsanol-ruby#25). The measurement therefore forces :ruby, the
12
+ # only working mode for full parslet grammars via the shim.
13
+ #
14
+ # Run from asciichem-ruby/:
15
+ # ruby -I /tmp/parsanol_spike -I ../parsanol/parsanol-ruby/lib benchmarks/parsanol_recheck.rb
16
+ require "benchmark/ips"
17
+ require "asciichem"
18
+ require "json"
19
+
20
+ Parsanol::Native.singleton_class.define_method(:available?) { false } # spike: force :ruby
21
+
22
+ puts "parsanol #{Parsanol::VERSION} | parslet-compat Parser=#{Parsanol::Parslet::Parser}"
23
+ puts "AsciiChem::Grammar superclass: #{AsciiChem::Grammar.superclass}"
24
+
25
+ # -- 1. Issue-25 repro: repeat-of-maybe at end of input --------------
26
+ begin
27
+ formula = AsciiChem.parse("SO_4^2-")
28
+ puts "issue-25 repro SO_4^2-: PARSES -> #{formula.to_text.inspect}"
29
+ rescue AsciiChem::ParseError => e
30
+ puts "issue-25 repro SO_4^2-: FAILS -> #{e.message[0, 100]}"
31
+ end
32
+
33
+ # -- 2. Shared-corpus gate --------------------------------------------
34
+ corpus_dir = File.expand_path("../../asciichem-tests/corpus/fixtures", __dir__)
35
+ cases = Dir[File.join(corpus_dir, "*.json")].sort.flat_map { |p| JSON.parse(File.read(p)) }
36
+ parser_cases = cases.select { |c| c.key?("input") && !c.key?("lint") && !c.key?("convention") }
37
+
38
+ pass = fail_parse = fail_reject = fail_roundtrip = 0
39
+ parser_cases.each do |fixture|
40
+ input = fixture.fetch("input")
41
+ if fixture.fetch("parses")
42
+ begin
43
+ formula = AsciiChem.parse(input)
44
+ if fixture["roundTrip"] && formula.to_text != input
45
+ fail_roundtrip += 1
46
+ puts " ROUNDTRIP DIFF: #{input.inspect} -> #{formula.to_text.inspect}" if fail_roundtrip <= 5
47
+ end
48
+ pass += 1
49
+ rescue AsciiChem::ParseError, Parslet::ParseFailed => e
50
+ fail_parse += 1
51
+ puts " PARSE FAIL: #{input.inspect} -> #{e.message[0, 90]}" if fail_parse <= 8
52
+ end
53
+ else
54
+ begin
55
+ AsciiChem.parse(input)
56
+ fail_reject += 1
57
+ puts " SHOULD REJECT: #{input.inspect}" if fail_reject <= 8
58
+ rescue AsciiChem::ParseError, Parslet::ParseFailed
59
+ pass += 1
60
+ end
61
+ end
62
+ end
63
+ total = parser_cases.length
64
+ puts format("corpus gate: %d/%d ok (parse-fails %d, should-reject %d, roundtrip-diffs %d)",
65
+ pass, total, fail_parse, fail_reject, fail_roundtrip)
66
+
67
+ # -- 3. Performance ----------------------------------------------------
68
+ WORKLOAD = [
69
+ "H_2O", "Ca^2+", "SO_4^2-", "(R)-CH_3CH(OH)COOH",
70
+ "2H_2 + O_2 -> 2H_2O", "N_2 + 3H_2 <=>[Fe][400C] 2NH_3",
71
+ "C1-C-C-C-C-C1", "CH_3-CH_2-OH",
72
+ '^14C @name("carbon-14") @cas("14104-86-4")',
73
+ "A ->[heat] B ->[cool] C",
74
+ ].freeze
75
+
76
+ Benchmark.ips do |x|
77
+ x.report("parsanol parse x10") { WORKLOAD.each { |s| AsciiChem.parse(s) } }
78
+ x.compare!
79
+ end
@@ -830,11 +830,18 @@ module AsciiChem
830
830
 
831
831
  def canonicalise_hash(hash)
832
832
  first = hash[:first]
833
- arrows = Array(hash[:arrow])
834
- products = Array(hash[:products])
835
- tail = arrows.zip(products)
833
+ # Wrap, never Array(): Array(hash) on a scalar Hash value
834
+ # (the shape parsanol's transform delivers) would enumerate
835
+ # key/value pairs instead of wrapping it.
836
+ tail = wrap(hash[:arrow]).zip(wrap(hash[:products]))
836
837
  [first, tail]
837
838
  end
839
+
840
+ def wrap(value)
841
+ return [] if value.nil?
842
+
843
+ value.is_a?(Array) ? value : [value]
844
+ end
838
845
  end
839
846
 
840
847
  # Maps the captured stereo letter to the model's stereo symbol.
@@ -1,5 +1,5 @@
1
1
  # frozen_string_literal: true
2
2
 
3
3
  module AsciiChem
4
- VERSION = "0.28.0"
4
+ VERSION = "0.28.2"
5
5
  end
metadata CHANGED
@@ -1,7 +1,7 @@
1
1
  --- !ruby/object:Gem::Specification
2
2
  name: asciichem
3
3
  version: !ruby/object:Gem::Version
4
- version: 0.28.0
4
+ version: 0.28.2
5
5
  platform: ruby
6
6
  authors:
7
7
  - Ribose Inc.
@@ -190,6 +190,7 @@ files:
190
190
  - benchmarks/RESULTS.md
191
191
  - benchmarks/benchmark.rb
192
192
  - benchmarks/engines.rb
193
+ - benchmarks/parsanol_recheck.rb
193
194
  - exe/asciichem
194
195
  - lib/asciichem.rb
195
196
  - lib/asciichem/citation.rb