textcortex · basheb · Jan 17, 2025
diff --git a/examples/html/html_all_elements.html b/examples/html/html_all_elements.html
@@ -445,14 +445,7 @@ <h1>Tables</h1>
         <caption>
           This is a caption for a table
         </caption>
-        <thead>
-          <tr>
-            <th>ID</th>
-            <th>Name</th>
-            <th>Date</th>
-            <th>Address</th>
-          </tr>
-        </thead>
+
         <tfoot>
           <tr>
             <td>Table footer info</td>
@@ -472,6 +465,14 @@ <h1>Tables</h1>
             <td>999 Spruce Lane, Somewhere, CA 94101</td>
           </tr>
         </tbody>
+        <thead>
+          <tr>
+            <th>ID</th>
+            <th>Name</th>
+            <th>Date</th>
+            <th>Address</th>
+          </tr>
+        </thead>
       </table>
     </section>
     <section>

diff --git a/examples/html/html_just_table.html b/examples/html/html_just_table.html
@@ -0,0 +1,94 @@
+<!DOCTYPE html>
+<html lang="en">
+  <head>
+    <meta charset="utf-8" />
+    <meta http-equiv="X-UA-Compatible" content="IE=edge,chrome=1" />
+
+    <!-- Set the title of your site here -->
+    <title>HTML Patterns</title>
+
+    <meta name="author" content="mrmrs" />
+    <meta name="description" content="Common HTML patterns" />
+    <meta name="viewport" content="width=device-width, initial-scale=1" />
+
+    <!-- Link to your stylesheet here
+    <link rel="stylesheet" href="css/i.css">
+
+    Favicons and Touch Device Icons
+    <link rel="shortcut icon" href="favicon.ico">
+    <link rel="apple-touch-icon" href="touch-icon-iphone-precomposed.png">
+    <link rel="apple-touch-icon" sizes="72x72"   href="touch-icon-ipad-precomposed.png">
+    <link rel="apple-touch-icon" sizes="114x114" href="touch-icon-iphone-retina-precomposed.png">
+    <link rel="apple-touch-icon" sizes="144x144" href="touch-icon-ipad-retina-precomposed.png">
+    -->
+  </head>
+
+  <body>
+    <header>
+      <h1>HTML</h1>
+      <h2>Every html element in one place. Just waiting to be styled.</h2>
+    </header>
+
+    <section>
+      <h1>Tables</h1>
+      <!--
+        From the HTML spec (http://www.w3.org/TR/html401/struct/tables.html)
+
+        TFOOT must appear before TBODY within a TABLE definition so that user agents can
+        render the foot before receiving all of the (potentially numerous) rows of data.
+        The following summarizes which tags are required and which may be omitted:
+
+        The TBODY start tag is always required except when the table contains only one
+        table body and no table head or foot sections. The TBODY end tag may always be
+        safely omitted.
+
+        The start tags for THEAD and TFOOT are required when the table head and foot sections
+        are present respectively, but the corresponding end tags may always be safely
+        omitted.
+
+        Conforming user agent parsers must obey these rules for reasons of backward
+        compatibility.
+        -->
+      <table>
+        <caption>
+          This is a caption for a table
+        </caption>
+
+        <tfoot>
+          <tr>
+            <td>Table footer info</td>
+          </tr>
+        </tfoot>
+        <tbody>
+          <tr>
+            <td>#999-32ac</td>
+            <td>First Name</td>
+            <td>13 May, 2013</td>
+            <td>999 Spruce Lane, Somewhere, CA 94101</td>
+          </tr>
+          <tr>
+            <td>#888-32dd</td>
+            <td>Sample Name</td>
+            <td>17 May, 1984</td>
+            <td>999 Spruce Lane, Somewhere, CA 94101</td>
+          </tr>
+        </tbody>
+        <thead>
+          <tr>
+            <th>ID</th>
+            <th>Name</th>
+            <th>Date</th>
+            <th>Address</th>
+          </tr>
+        </thead>
+      </table>
+    </section>
+    <section>
+      <h1>Footer</h1>
+      <footer>
+        <small>© 2014 Some company name</small>
+        <address>email@email.com</address>
+      </footer>
+    </section>
+  </body>
+</html>
diff --git a/jsondoc/convert/html.py b/jsondoc/convert/html.py
@@ -834,7 +834,15 @@ def convert_tr(self, el, convert_as_inline):
         """
         Table row
         """
-        return ConvertOutput(main_object=create_table_row_block())
+        # Check if this is a header row
+        is_header = False
+        is_footer = False
+        if el.find_parent("thead") or el.find_parent("th"):
+            is_header = True
+        # Check if this is a footer row
+        if el.find_parent("tfoot"):
+            is_footer = True
+        return ConvertOutput(main_object=create_table_row_block(isHeader=is_header, isFooter=is_footer))
 
 
 def html_to_jsondoc(html: str | bytes, **options) -> Page | BlockBase | List[BlockBase]:

diff --git a/jsondoc/convert/markdown.py b/jsondoc/convert/markdown.py
@@ -349,7 +349,7 @@ def convert_quote_block(self, block: QuoteBlock, convert_as_inline: bool) -> str
         )
 
     def _convert_table_row_block(
-        self, block: TableRowBlock, convert_as_inline: bool, is_headrow: bool
+        self, block: TableRowBlock, convert_as_inline: bool, is_headrow: bool, is_footrow: bool
     ) -> str:
         # cells = block.table_row.cells
 
@@ -383,18 +383,37 @@ def _convert_table_row_block(
         return overline + "|" + text + "\n" + underline
 
     def convert_table_block(self, block: TableBlock, convert_as_inline: bool) -> str:
+        # TODO: Caption is not at the very bottom after the footer. Fix it.
         text = ""
-
         for n, row in enumerate(block.children):
-            is_headrow = block.table.has_column_header and n == 0
-
-            text += self._convert_table_row_block(
+            is_headrow = block.table.has_column_header and row.isHeader
+            is_footrow = row.isFooter
+
+            textHeader = "\n\n"
+            textFooter = ""
+
+            if is_headrow:
+                textHeader = self._convert_table_row_block(
                 row,
                 convert_as_inline,
                 is_headrow=is_headrow,
+                is_footrow=is_footrow,
             )
-
-        return "\n\n" + text + "\n"
+            elif is_footrow:
+                textFooter = self._convert_table_row_block(
+                row,
+                convert_as_inline,
+                is_headrow=is_headrow,
+                is_footrow=is_footrow,
+            )
+            else:
+                text += self._convert_table_row_block(
+                    row,
+                    convert_as_inline,
+                    is_headrow=is_headrow,
+                    is_footrow=is_footrow,
+                )
+        return textHeader + text + textFooter
 
 
 def jsondoc_to_markdown(jsondoc, **options):

diff --git a/jsondoc/convert/utils.py b/jsondoc/convert/utils.py
@@ -408,6 +408,8 @@ def create_table_row_block(
     cells: List[List[RichTextBase]] = [],
     id: str | None = None,
     created_time=None,
+    isHeader: bool = False,
+    isFooter: bool = False,
 ) -> TableRowBlock:
     if id is None:
         id = generate_id()
@@ -420,6 +422,8 @@ def create_table_row_block(
         created_time=created_time,
         table_row=TableRow(cells=cells),
         has_children=False,
+        isHeader=isHeader,
+        isFooter=isFooter,
     )
 
 
@@ -430,6 +434,7 @@ def create_table_block(
     table_width: int | None = None,
     has_column_header: bool = False,
     has_row_header: bool = False,
+    caption: str | None = None,
 ) -> TableBlock:
     if id is None:
         id = generate_id()

diff --git a/jsondoc/models/block/types/table_row/__init__.py b/jsondoc/models/block/types/table_row/__init__.py
@@ -23,3 +23,5 @@ class TableRow(BaseModel):
 class TableRowBlock(BlockBase):
     type: Literal['table_row'] = 'table_row'
     table_row: TableRow
+    isHeader: bool
+    isFooter: bool
diff --git a/schema/block/types/table_row/table_row_schema.json b/schema/block/types/table_row/table_row_schema.json
@@ -26,6 +26,12 @@
       },
       "required": ["cells"],
       "additionalProperties": false
+    },
+    "isHeader": {
+      "type": "boolean"
+    },
+    "isFooter": {
+      "type": "boolean"
     }
   },
   "required": ["type", "table_row"],

diff --git a/tests/html_jsondoc_pairs/test_table_basic.json b/tests/html_jsondoc_pairs/test_table_basic.json
@@ -49,7 +49,9 @@
               }
             ]
           ]
-        }
+        },
+        "isHeader": true,
+        "isFooter": false
       },
       {
         "object": "block",
@@ -90,7 +92,9 @@
               }
             ]
           ]
-        }
+        },
+        "isHeader": false,
+        "isFooter": false
       },
       {
         "object": "block",
@@ -131,7 +135,9 @@
               }
             ]
           ]
-        }
+        },
+        "isHeader": false,
+        "isFooter": false
       },
       {
         "object": "block",
@@ -172,7 +178,9 @@
               }
             ]
           ]
-        }
+        },
+        "isHeader": false,
+        "isFooter": false
       }
     ]
   }

diff --git a/tests/html_jsondoc_pairs/test_table_with_caption.json b/tests/html_jsondoc_pairs/test_table_with_caption.json
@@ -61,7 +61,9 @@
                 }
               ]
             ]
-          }
+          },
+          "isHeader": true,
+          "isFooter": false
         },
         {
           "object": "block",
@@ -112,7 +114,9 @@
                 }
               ]
             ]
-          }
+          },
+          "isHeader": false,
+          "isFooter": false
         },
         {
           "object": "block",
@@ -163,7 +167,9 @@
                 }
               ]
             ]
-          }
+          },
+          "isHeader": false,
+          "isFooter": false
         },
         {
           "object": "block",
@@ -187,7 +193,9 @@
               [],
               []
             ]
-          }
+          },
+          "isHeader": false,
+          "isFooter": true
         }
       ]
     },

diff --git a/tests/test_html_to_jsondoc.py b/tests/test_html_to_jsondoc.py
@@ -87,6 +87,18 @@ def test_examples():
         print("PASS")
 
 
+def test():
+    path = "examples/html/html_just_table.html" 
+    content = open(path, "r").read()
+    ret = html_to_jsondoc(content)
+    # print(jsondoc_dump_json(ret, indent=2))
+
+    # print("\n\nConverted to markdown:\n\n")
+    # print(jsondoc_to_markdown(ret[0]))
+    print(jsondoc_to_markdown(ret))
+
 if __name__ == "__main__":
     test_examples()
     test_convert_html_all_elements()
+    # Test for working specifically on tables with caption
+    # test()