Merge pull request pqzx#16 from jongracecox/jongracecox/add-table-styles

Add support for styled tables

Author: pqzx

README.md

@@ -43,3 +43,23 @@ from htmldocx import HtmlToDocx
 new_parser = HtmlToDocx()
 docx = new_parser.parse_html_string(input_html_file_string)
 ```
+
+Change table styles
+
+Tables are not styled by default. Use the `table_style` attribute on the parser to set a table
+style. The style is used for all tables.
+
+```
+from htmldocx import HtmlToDocx
+
+new_parser = HtmlToDocx()
+new_parser.table_style = 'Light Shading Accent 4'
+```
+
+To add borders to tables, use the `TableGrid` style:
+
+```
+new_parser.table_style = 'TableGrid'
+```
+
+Default table styles can be found here: https://python-docx.readthedocs.io/en/latest/user/styles-understanding.html#table-styles-in-default-template

htmldocx/h2d.py

@@ -32,6 +32,9 @@
 LIST_INDENT = 0.5
 MAX_INDENT = 5.5 # To stop indents going off the page
 
+# Style to use with tables. By default no style is used.
+DEFAULT_TABLE_STYLE = None
+
 def get_filename_from_url(url):
     return os.path.basename(urlparse(url).path)
 
@@ -103,6 +106,7 @@ def __init__(self):
             'table > tbody > tr',
             'table > tfoot > tr'
         ]
+        self.table_style = DEFAULT_TABLE_STYLE
 
     def set_initial_attrs(self, document=None):
         self.tags = {
@@ -123,6 +127,10 @@ def set_initial_attrs(self, document=None):
         self.skip_tag = None
         self.instances_to_skip = 0
 
+    def copy_settings_from(self, other):
+        """Copy settings from another instance of HtmlToDocx"""
+        self.table_style = other.table_style
+
     def get_cell_html(self, soup):
         # Returns string of td element with opening and closing <td> tags removed
         # Cannot use find_all as it only finds element tags and does not find text which
@@ -244,6 +252,13 @@ def handle_table(self):
         table_soup = self.tables[self.table_no]
         rows, cols = self.get_table_dimensions(table_soup)
         self.table = self.doc.add_table(rows, cols)
+
+        if self.table_style:
+            try:
+                self.table.style = self.table_style
+            except KeyError as e:
+                raise ValueError(f"Unable to apply style {self.table_style}.") from e
+
         rows = self.get_table_rows(table_soup)
         cell_row = 0
         for row in rows:
@@ -255,6 +270,7 @@ def handle_table(self):
                     cell_html = "<b>%s</b>" % cell_html
                 docx_cell = self.table.cell(cell_row, cell_col)
                 child_parser = HtmlToDocx()
+                child_parser.copy_settings_from(self)
                 child_parser.add_html_to_cell(cell_html, docx_cell)
                 cell_col += 1
             cell_row += 1

tests/test.py

@@ -1,19 +1,24 @@
 import os
+from pathlib import Path
 import unittest
 from docx import Document
 from .context import HtmlToDocx, test_dir
 
 class OutputTest(unittest.TestCase):
 
+    @staticmethod
+    def get_html_from_file(filename: str):
+        file_path = Path(test_dir) / Path(filename)
+        with open(file_path, 'r') as f:
+            html = f.read()
+        return html
+
     @classmethod
     def setUpClass(cls):
         cls.document = Document()
-        textpath = os.path.join(test_dir, 'text1.html')
-        tablepath = os.path.join(test_dir, 'tables1.html')
-        with open(textpath, 'r') as tb:
-            cls.text1 = tb.read()
-        with open(tablepath, 'r') as tb:
-            cls.table_html = tb.read()
+        cls.text1 = cls.get_html_from_file('text1.html')
+        cls.table_html = cls.get_html_from_file('tables1.html')
+        cls.table2_html = cls.get_html_from_file('tables2.html')
 
     @classmethod
     def tearDownClass(cls):
@@ -54,6 +59,40 @@ def test_add_html_with_tables(self):
         )
         self.parser.add_html_to_document(self.table_html, self.document)
 
+    def test_add_html_with_tables_accent_style(self):
+        self.document.add_heading(
+            'Test: add html with tables with accent',
+        )
+        self.parser.table_style = 'Light Grid Accent 6'
+        self.parser.add_html_to_document(self.table_html, self.document)
+
+    def test_add_html_with_tables_basic_style(self):
+        self.document.add_heading(
+            'Test: add html with tables with basic style',
+        )
+        self.parser.table_style = 'TableGrid'
+        self.parser.add_html_to_document(self.table_html, self.document)
+
+    def test_add_nested_tables(self):
+        self.document.add_heading(
+            'Test: add nested tables',
+        )
+        self.parser.add_html_to_document(self.table2_html, self.document)
+
+    def test_add_nested_tables_basic_style(self):
+        self.document.add_heading(
+            'Test: add nested tables with basic style',
+        )
+        self.parser.table_style = 'TableGrid'
+        self.parser.add_html_to_document(self.table2_html, self.document)
+
+    def test_add_nested_tables_accent_style(self):
+        self.document.add_heading(
+            'Test: add nested tables with accent style',
+        )
+        self.parser.table_style = 'Light Grid Accent 6'
+        self.parser.add_html_to_document(self.table2_html, self.document)
+
     def test_add_html_skip_tables(self):
         # broken until feature readded
         self.document.add_heading(