From 7ae08b1a0a7dc1a4ed7e1e9e1f9f6240dab7d9ae Mon Sep 17 00:00:00 2001
From: Tobias Becker <tobias.becker@methodpark.de>
Date: Fri, 28 Jan 2022 15:43:48 +0100
Subject: [PATCH] Add Regex extractor to problem parser

---
 ProblemParser/src/ProblemParser.jl | 10 ++++++++++
 ProblemParser/test/runtests.jl     | 16 ++++++++++++++++
 2 files changed, 26 insertions(+)

diff --git a/ProblemParser/src/ProblemParser.jl b/ProblemParser/src/ProblemParser.jl
index 7015fc6..ddec494 100644
--- a/ProblemParser/src/ProblemParser.jl
+++ b/ProblemParser/src/ProblemParser.jl
@@ -7,6 +7,7 @@ export FirstRest
 export LineMappings
 export Lines
 export Map
+export Extract
 export Mappings
 export Noop
 export Rectangular
@@ -59,6 +60,13 @@ struct Map <: GrammarElement
     func
 end
 
+struct Extract <: GrammarElement
+    regex::Regex
+    after::Union{GrammarElement,Nothing}
+end
+Extract() = Extract(r"\d+", Convert())
+Extract(regex::Regex) = Extract(regex, nothing)
+
 struct _Mappings <: GrammarElement
     splitter::Union{GrammarElement,Nothing}
     first_rest::FirstRest
@@ -103,6 +111,8 @@ Base.parse(apply::Apply, thing) = apply.func(thing)
 Base.parse(map::Map, list::AbstractArray) = Base.map(map.func, list)
 Base.parse(map::Map, text::AbstractString) = Base.map(map.func, collect(text))
 
+Base.parse(extract::Extract, text::AbstractString) = parse(extract.after, match(extract.regex, text).match)
+
 function Base.parse(mappings::_Mappings, text::AbstractString)
     kv_pairs = parse(mappings.splitter, text)
     parse(mappings.first_rest, kv_pairs) |> Dict
diff --git a/ProblemParser/test/runtests.jl b/ProblemParser/test/runtests.jl
index 10c8f76..9904655 100644
--- a/ProblemParser/test/runtests.jl
+++ b/ProblemParser/test/runtests.jl
@@ -198,4 +198,20 @@ end
         ([8, 10], "k") => "kklxkkkqkkkkk",
         ([3, 4], "h")  => "hrht",
     )
+
+    not_just_lines = """
+        --- scanner 0 ---
+        0,2
+        4,1
+        3,3
+
+        --- scanner 1 ---
+        -1,-1
+        -5,0
+        -2,1
+        """
+    @test parse(Blocks(FirstRest(Lines(), Extract(), Lines(Split(',', Convert())))), not_just_lines) == [
+        (0, [[0, 2], [4, 1], [3, 3]]),
+        (1, [[-1, -1], [-5, 0], [-2, 1]]),
+    ]
 end