From e9d44c55a1fe87ec45071065c7b9b1625d828d2c Mon Sep 17 00:00:00 2001
From: Ralph Slooten <axllent@gmail.com>
Date: Sat, 14 Oct 2023 22:28:52 +1300
Subject: [PATCH] Tests: Add html2text tests

---
 .github/workflows/tests.yml                |  2 +-
 internal/tools/html2text/html2text_test.go | 56 ++++++++++++++++++++++
 2 files changed, 57 insertions(+), 1 deletion(-)
 create mode 100644 internal/tools/html2text/html2text_test.go

diff --git a/.github/workflows/tests.yml b/.github/workflows/tests.yml
index e26ce529f..c26e25830 100644
--- a/.github/workflows/tests.yml
+++ b/.github/workflows/tests.yml
@@ -24,7 +24,7 @@ jobs:
         key: ${{ runner.os }}-go-${{ hashFiles('**/go.sum') }}
         restore-keys: |
           ${{ runner.os }}-go-
-    - run: go test ./internal/storage ./server ./internal/tools -v
+    - run: go test ./internal/storage ./server ./internal/tools ./internal/tools/html2text -v
     - run: go test ./internal/storage -bench=.
     
     # build the assets
diff --git a/internal/tools/html2text/html2text_test.go b/internal/tools/html2text/html2text_test.go
new file mode 100644
index 000000000..dd1763918
--- /dev/null
+++ b/internal/tools/html2text/html2text_test.go
@@ -0,0 +1,56 @@
+package html2text
+
+import "testing"
+
+func TestPlain(t *testing.T) {
+	tests := map[string]string{}
+	tests["this is a  test"] = "this is a test"
+	tests["thiS IS a Test"] = "thiS IS a Test"
+	tests["thiS IS a Test :-)"] = "thiS IS a Test :-)"
+	tests["<h1>This is a test.</h1> "] = "This is a test."
+	tests["<p>Paragraph 1</p><p>Paragraph 2</p>"] = "Paragraph 1 Paragraph 2"
+	tests["<h1>Heading</h1><p>Paragraph</p>"] = "Heading Paragraph"
+	tests["<span>Alpha</span>bet <strong>chars</strong>"] = "Alphabet chars"
+	tests["<span><b>A</b>lpha</span>bet  chars."] = "Alphabet chars."
+	tests["<table><tr><td>First</td><td>Second</td></table>"] = "First Second"
+	tests[`<h1>Heading</h1>
+		<p>Paragraph</p>`] = "Heading Paragraph"
+	tests[`<h1>Heading</h1><p>   <a href="https://github.com">linked text</a></p>`] = "Heading linked text"
+	// broken html
+	tests[`<h1>Heading</h3><p>   <a href="https://github.com">linked text.`] = "Heading linked text."
+
+	for str, expected := range tests {
+		res := Strip(str, false)
+		if res != expected {
+			t.Log("error:", res, "!=", expected)
+			t.Fail()
+		}
+	}
+}
+
+func TestWithLinks(t *testing.T) {
+	tests := map[string]string{}
+	tests["this is a  test"] = "this is a test"
+	tests["thiS IS a Test"] = "thiS IS a Test"
+	tests["thiS IS a Test :-)"] = "thiS IS a Test :-)"
+	tests["<h1>This is a test.</h1> "] = "This is a test."
+	tests["<p>Paragraph 1</p><p>Paragraph 2</p>"] = "Paragraph 1 Paragraph 2"
+	tests["<h1>Heading</h1><p>Paragraph</p>"] = "Heading Paragraph"
+	tests["<span>Alpha</span>bet <strong>chars</strong>"] = "Alphabet chars"
+	tests["<span><b>A</b>lpha</span>bet  chars."] = "Alphabet chars."
+	tests["<table><tr><td>First</td><td>Second</td></table>"] = "First Second"
+	tests["<h1>Heading</h1><p>Paragraph</p>"] = "Heading Paragraph"
+	tests[`<h1>Heading</h1>
+		<p>Paragraph</p>`] = "Heading Paragraph"
+	tests[`<h1>Heading</h1><p>   <a href="https://github.com">linked text</a></p>`] = "Heading https://github.com linked text"
+	// broken html
+	tests[`<h1>Heading</h3><p>   <a href="https://github.com">linked text.`] = "Heading https://github.com linked text."
+
+	for str, expected := range tests {
+		res := Strip(str, true)
+		if res != expected {
+			t.Log("error:", res, "!=", expected)
+			t.Fail()
+		}
+	}
+}