Optimized single col tables

spacecowboy · May 28, 2024 · 29b259f · 29b259f
1 parent b276b08
commit 29b259f
Show file tree

Hide file tree

Showing 2 changed files with 48 additions and 3 deletions.
diff --git a/app/src/main/java/com/nononsenseapps/feeder/model/html/HtmlLinearizer.kt b/app/src/main/java/com/nononsenseapps/feeder/model/html/HtmlLinearizer.kt
@@ -465,6 +465,17 @@ class HtmlLinearizer {
                             }
                         }
 
+                        "td", "th" -> {
+                            // If we end up here, that means the table has been optimized out. Treat as a div.
+                            asElement(blockStyle) {
+                                linearizeChildren(
+                                    element.childNodes(),
+                                    blockStyle = blockStyle,
+                                    baseUrl = baseUrl,
+                                )
+                            }
+                        }
+
                         "table" -> {
                             finalizeAndAddCurrentElement(blockStyle)
 
@@ -493,8 +504,16 @@ class HtmlLinearizer {
                                         }
                                 }
 
-                            // If there is only a single row, then don't bother with a table
-                            if (rowSequence.count() == 1) {
+                            val colCount =
+                                rowSequence
+                                    .map { row ->
+                                        row.children().count { it.tagName() == "td" || it.tagName() == "th" }
+                                    }
+                                    .maxOrNull()
+                                    ?: 0
+
+                            // If there is only a single row, or a single column, then don't bother with a table
+                            if (colCount == 1 || rowSequence.count() == 1) {
                                 linearizeChildren(
                                     element.childNodes(),
                                     blockStyle = blockStyle,

diff --git a/app/src/test/java/com/nononsenseapps/feeder/model/html/HtmlLinearizerTest.kt b/app/src/test/java/com/nononsenseapps/feeder/model/html/HtmlLinearizerTest.kt
@@ -1160,10 +1160,36 @@ class HtmlLinearizerTest {
 
         val result = linearizer.linearize(html, baseUrl).elements
 
-        assertEquals(2, result.count { it is LinearTable }, "Expected two tables in result")
+        assertEquals(1, result.count { it is LinearTable }, "Expected one table in result")
         assertEquals(8, result.filterIsInstance<LinearTable>().first().rowCount, "Expected table with 8 rows")
     }
 
+    @Test
+    fun `table with single column is optimized out`() {
+        val html =
+            """
+            <table>
+                <tbody>
+                    <tr>
+                        <td>Single column table</td>
+                    </tr>
+                    <tr>
+                        <td>Second row</td>
+                    </tr>
+                </tbody>
+            </table>
+            """.trimIndent()
+
+        val baseUrl = "https://example.com"
+
+        val result = linearizer.linearize(html, baseUrl).elements
+
+        assertEquals(2, result.size, "Expected two text items: $result")
+        assertTrue("Expected all to be linear text items: $result") {
+            result.all { it is LinearText }
+        }
+    }
+
     @Test
     fun `insane nested table`() {
         // from kill-the-newsletter