search

Knio / dominate

Dominate is a Python library for creating and manipulating HTML documents using an elegant DOM API. It allows you to write HTML pages in pure Python very concisely, which eliminate the need to learn another template language, and to take advantage of the more powerful features of Python.
GNU Lesser General Public License v3.0
1.72k stars 108 forks source link

support recursive with dominate.document() #157

Closed MarisaKirisame closed 2 years ago

MarisaKirisame commented 2 years ago 
import dominate
from dominate.tags import *

with dominate.document() as outer_doc:
    with dominate.document() as inner_doc:
        p('meow')
    print(inner_doc)
print(outer_doc)

output

<!DOCTYPE html>
<html>
  <head>
    <title>Dominate</title>
  </head>
  <body>
    <p>meow</p>
  </body>
</html>
<!DOCTYPE html>
<html>
  <head>
    <title>Dominate</title>
  </head>
  <body>
    <html>
      <head>
        <title>Dominate</title>
      </head>
      <body>
        <p>meow</p>
      </body>
    </html>
  </body>
</html>
Knio commented 2 years ago

No issue/problem/question stated and this looks working as intended

MarisaKirisame commented 2 years ago

@Knio The tag represents the root of an HTML document. (from https://www.w3schools.com/tags/tag_html.asp) so - i dont think it should nest, which is what is happening if you look at the output from the second occurence of <!DOCTYPE ...> instead, I think a better design choice is, when a html root is entered/exited, it does not register itself as a child of what is on the context, breaking the chain. roughly what happend is that with dominate.document() as xxx: ... should has no other effect then writing xxx.

This come up in real life, when I try to use dominate, as I try to create multiple webpage in a hirearchical order. I initially wrote a recursive function to do that, starting from the main entry, generating hyperlink on needed, but as dominate doesnt support that I have to rewrite it from a top-down to a bottom-up manner.

Knio commented 2 years ago

i dont think it should nest

This project is not meant to be an html validator - i.e. it doesn't check that <td> is inside a <tr>, or that some tag has all the required attributes set, etc. and I don't want to special case any nodes. Also, I think nesting is legal (ok maybe only in an <embed> but I'm not going to check the whole spec)

with dominate.document() as xxx: ... should has no other effect then writing xxx.

This is explicitly asking for it to be the context. Not doing so would break the expectations of the API. If you don't want the effect, why not just write xxx = dominate.document()?

If xxx = ... is still inside some outer context and causing problems:

but as dominate doesnt support that I have to rewrite it from a top-down to a bottom-up manner.

That seems like an excessive reaction, why don't you just subclass document to do what you want?

class mydoc(dominate.document):
    def _add_to_ctx(self): pass # don't add to contexts

With your example do what I think you want:

>>> with mydoc() as outer_doc:
...     with mydoc() as inner_doc:
...         p('meow')
...     print(inner_doc)
...
<dominate.tags.p at 1bb1a765160: 0 attributes, 1 child>
<!DOCTYPE html>
<html>
  <head>
    <title>Dominate</title>
  </head>
  <body>
    <p>meow</p>
  </body>
</html>
>>> print(outer_doc)
<!DOCTYPE html>
<html>
  <head>
    <title>Dominate</title>
  </head>
  <body></body>
</html>
MarisaKirisame commented 2 years ago

Thx. the subclass solution is excatly what I want.