diff --git a/CHANGELOG.md b/CHANGELOG.md
index 45f1142..425d2b2 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -4,6 +4,8 @@ All notable changes to this project will be documented in this file.
This project adheres to [Semantic Versioning](http://semver.org/).
## Unreleased
+- Fix concatenation of multiple matching tags in Title extractor
+- Fix exception when no matches are found in Title extractor
## [0.6.2] - 2015.11.22
- Scope the title tag selector to the head element
diff --git a/lib/readability/title_finder.ex b/lib/readability/title_finder.ex
index bf7243b..59ac9b3 100644
--- a/lib/readability/title_finder.ex
+++ b/lib/readability/title_finder.ex
@@ -33,7 +33,7 @@ defmodule Readability.TitleFinder do
@spec tag_title(html_tree) :: binary
def tag_title(html_tree) do
html_tree
- |> Floki.find("head title")
+ |> find_tag("head title")
|> clean_title()
|> String.split(@title_suffix)
|> hd()
@@ -45,7 +45,7 @@ defmodule Readability.TitleFinder do
@spec og_title(html_tree) :: binary
def og_title(html_tree) do
html_tree
- |> Floki.find("meta[property=og:title]")
+ |> find_tag("meta[property=og:title]")
|> Floki.attribute("content")
|> clean_title()
end
@@ -56,11 +56,22 @@ defmodule Readability.TitleFinder do
@spec h_tag_title(html_tree, String.t) :: binary
def h_tag_title(html_tree, selector \\ @h_tag_selector) do
html_tree
- |> Floki.find(selector)
- |> hd()
+ |> find_tag(selector)
|> clean_title()
end
+ defp find_tag(html_tree, selector) do
+ case Floki.find(html_tree, selector) do
+ [] ->
+ []
+ matches when is_list(matches) ->
+ hd(matches)
+ end
+ end
+
+ defp clean_title([]) do
+ ""
+ end
defp clean_title(html_tree) do
html_tree
|> Floki.text()
diff --git a/test/readability/title_finder_test.exs b/test/readability/title_finder_test.exs
index 301ce1f..dc08ea4 100644
--- a/test/readability/title_finder_test.exs
+++ b/test/readability/title_finder_test.exs
@@ -28,6 +28,19 @@ defmodule Readability.TitleFinderTest do
assert title == "og title"
end
+ test "does not merge multiple matching og:title tags" do
+ html = """
+
+
+
+
+
+
+ """
+ title = Readability.TitleFinder.og_title(html)
+ assert title == "og title 1"
+ end
+
test "extract tag title" do
title = Readability.TitleFinder.tag_title(@html)
assert title == "Tag title"
@@ -86,13 +99,51 @@ defmodule Readability.TitleFinderTest do
assert title == "Tag title"
end
+ test "does not merge multiple title tags" do
+ html = """
+
+
+ tag title 1
+ tag title 2
+
+
+ """
+ title = Readability.TitleFinder.tag_title(html)
+ assert title == "tag title 1"
+ end
+
test "extract h1 tag title" do
title = Readability.TitleFinder.h_tag_title(@html)
assert title == "h1 title"
end
- test "extrat h2 tag title" do
+ test "extract h2 tag title" do
title = Readability.TitleFinder.h_tag_title(@html, "h2")
assert title == "h2 title"
end
+
+ test "does not merge multile header tags" do
+ html = """
+
+
+ header 1
+ header 2
+
+
+ """
+ title = Readability.TitleFinder.h_tag_title(html)
+ assert title == "header 1"
+ end
+
+ test "returns an empty string when no title tag can be found" do
+ assert Readability.TitleFinder.tag_title("") == ""
+ end
+
+ test "returns an empty string when no og:title tag can be found" do
+ assert Readability.TitleFinder.og_title("") == ""
+ end
+
+ test "returns an empty string when no header tag can be found" do
+ assert Readability.TitleFinder.h_tag_title("") == ""
+ end
end