From 489e964c6d5b180848001201c5069db611f1d9aa Mon Sep 17 00:00:00 2001
From: Marcel Otto <marcelotto@gmx.de>
Date: Sat, 20 Apr 2019 23:33:09 +0200
Subject: [PATCH] Add RDF.Literal.matches?/3

---
 CHANGELOG.md               | 11 +++++++
 lib/rdf/literal.ex         | 57 ++++++++++++++++++++++++++++++++++++
 test/unit/literal_test.exs | 59 +++++++++++++++++++++++++++++++++++++-
 3 files changed, 126 insertions(+), 1 deletion(-)

diff --git a/CHANGELOG.md b/CHANGELOG.md
index 18d24d0..65f97d5 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -5,6 +5,17 @@ This project adheres to [Semantic Versioning](http://semver.org/) and
 [Keep a CHANGELOG](http://keepachangelog.com).
 
 
+## Unreleased
+
+### Added
+
+- `RDF.Literal.matches?/3` for XQuery regex pattern matching
+
+
+[Compare v0.6.0...HEAD](https://github.com/marcelotto/rdf-ex/compare/v0.6.0...HEAD)
+
+
+
 ## 0.6.0 - 2019-04-06
 
 ### Added
diff --git a/lib/rdf/literal.ex b/lib/rdf/literal.ex
index 77faae9..4c98ef2 100644
--- a/lib/rdf/literal.ex
+++ b/lib/rdf/literal.ex
@@ -319,6 +319,63 @@ defmodule RDF.Literal do
 
   def compare(_, _), do: nil
 
+
+  @doc """
+  Matches the string representation of the given value against a XPath and XQuery regular expression pattern.
+
+  The regular expression language is defined in _XQuery 1.0 and XPath 2.0 Functions and Operators_.
+
+  The `pattern` and the optional `flags` can be given as an Elixir string or as
+  `xsd:string` `RDF.Literal`s.
+
+  see <https://www.w3.org/TR/xpath-functions/#func-matches>
+  """
+  def matches?(value, pattern, flags \\ "")
+
+  def matches?(value, %RDF.Literal{datatype: @xsd_string} = pattern, flags),
+    do: matches?(value, pattern.value, flags)
+
+  def matches?(value, pattern, %RDF.Literal{datatype: @xsd_string} = flags),
+    do: matches?(value, pattern, flags.value)
+
+  def matches?(value, pattern, flags) when is_binary(pattern) and is_binary(flags) do
+    string = to_string(value)
+    case xpath_pattern(pattern, flags) do
+      {:regex, regex} ->
+        Regex.match?(regex, string)
+
+      {:q, pattern} ->
+        String.contains?(string, pattern)
+
+      {:qi, pattern} ->
+        string
+        |> String.downcase()
+        |> String.contains?(String.downcase(pattern))
+
+      _ ->
+        raise "Invalid XQuery regex pattern or flags"
+    end
+  end
+
+  defp xpath_pattern(pattern, flags) do
+    q_pattern(pattern, flags) || xpath_regex_pattern(pattern, flags)
+  end
+
+  defp q_pattern(pattern, flags) do
+    if String.contains?(flags, "q") and String.replace(flags, ~r/[qi]/, "") == "" do
+      {(if String.contains?(flags, "i"), do: :qi, else: :q), pattern}
+    end
+  end
+
+  defp xpath_regex_pattern(pattern, flags) do
+    with {:ok, regex} <- Regex.compile(pattern, xpath_regex_flags(flags)) do
+      {:regex, regex}
+    end
+  end
+
+  defp xpath_regex_flags(flags) do
+    String.replace(flags, "q", "") <> "u"
+  end
 end
 
 defimpl String.Chars, for: RDF.Literal do
diff --git a/test/unit/literal_test.exs b/test/unit/literal_test.exs
index e0ac470..61ed54c 100644
--- a/test/unit/literal_test.exs
+++ b/test/unit/literal_test.exs
@@ -232,6 +232,64 @@ defmodule RDF.LiteralTest do
   end
 
 
+  @poem RDF.string """
+    <poem author="Wilhelm Busch">
+    Kaum hat dies der Hahn gesehen,
+    Fängt er auch schon an zu krähen:
+    Kikeriki! Kikikerikih!!
+    Tak, tak, tak! - da kommen sie.
+    </poem>
+    """
+
+  describe "matches?" do
+    test "without flags" do
+      [
+        {~L"abracadabra", ~L"bra",    true},
+        {~L"abracadabra", ~L"^a.*a$", true},
+        {~L"abracadabra", ~L"^bra",   false},
+        {@poem, ~L"Kaum.*krähen",     false},
+        {@poem, ~L"^Kaum.*gesehen,$", false},
+
+        {~L"abracadabra"en, ~L"bra", true},
+        {"abracadabra", "bra",       true},
+        {RDF.integer("42"), ~L"4",   true},
+        {RDF.integer("42"), ~L"en",  false},
+      ]
+      |> Enum.each(fn {literal, pattern, expected_result} ->
+           result = Literal.matches?(literal, pattern)
+           assert result == expected_result,
+            "expected RDF.Literal.matches?(#{inspect literal}, #{inspect pattern}) to return #{inspect expected_result}, but got #{result}"
+         end)
+    end
+
+    test "with flags" do
+      [
+        {@poem, ~L"Kaum.*krähen",     ~L"s", true},
+        {@poem, ~L"^Kaum.*gesehen,$", ~L"m", true},
+        {@poem, ~L"kiki",             ~L"i", true},
+      ]
+      |> Enum.each(fn {literal, pattern, flags, result} ->
+           assert Literal.matches?(literal, pattern, flags) == result
+         end)
+    end
+
+    test "with q flag" do
+      [
+        {~L"abcd",         ~L".*",       ~L"q",  false},
+        {~L"Mr. B. Obama", ~L"B. OBAMA", ~L"iq", true},
+
+        # If the q flag is used together with the m, s, or x flag, that flag has no effect.
+        {~L"abcd",         ~L".*",       ~L"mq",   true},
+        {~L"abcd",         ~L".*",       ~L"qim",  true},
+        {~L"abcd",         ~L".*",       ~L"xqm",  true},
+      ]
+      |> Enum.each(fn {literal, pattern, flags, result} ->
+           assert Literal.matches?(literal, pattern, flags) == result
+         end)
+    end
+  end
+
+
   describe "String.Chars protocol implementation" do
     Enum.each values(:all_plain), fn value ->
       @tag value: value
@@ -258,7 +316,6 @@ defmodule RDF.LiteralTest do
            assert to_string(literal) == rep
          end
        end)
-
   end
 
 end