Can't display new lines in variable

[jsykora]

First off, this may not be an issue, but more of a support question.

I have a model with a TextField and I am trying to use it in a .odt template converting to .pdf via Open Office.

The value of the field is u'Line one\r\nLine two'

I've tried just the plain variable as well as the following template tags: linebreaks, linebreaksbr, safe. But none of them seem to work to allow the carriage return or newline characters.

Am I missing the proper way to allow new lines?

[imankulov]

According to ODF specification, line breaks in ODF document must be represented with <text:line-break /> XML tag, so if you write a template filter converting u'Line one\r\nLine two' to u'Line one<text:line-break />Line two', it should work.

For more info, see "6.1.5 text:line-break" of the Specification (pdf), or simply perform a simple experiment as I did. I created a new document with a number of linebreaks and then took a look at the contents, as described here.

Please consider closing the ticket if it solves your problem. Otherwise, please let me know if something goes wrong.

[jsykora]

I figured it wasn't a bug but more of a feature that wasn't implemented.

It may be nice to have that filter included in the package. I assume wanting to put text with line breaks in documents would be fairly popular- blog posts, news articles, etc.

Here is a Gist of the code for a templatetag for xml line breaks. I borrowed/stole the code from Django's linebreaksbr filter and just switched it to the proper xml tag.

from django import template
from django.template.defaultfilters import stringfilter
from django.utils.safestring import mark_safe
from django.utils.text import normalize_newlines

register = template.Library()

@register.filter()
@stringfilter
def linebreaksodt(value, autoescape=None):
    """
    Converts all newlines in a piece of plain text to XML line breaks
    (``<text:line-break />``).
    """
    autoescape = autoescape and not isinstance(value, SafeData)
    value = normalize_newlines(value)
    if autoescape:
        value = escape(value)
    return mark_safe(value.replace('\n', '<text:line-break />'))

[imankulov]

Thank you. Indeed, it would be useful to include this filter in the webodt package. I re-open the issue as a reminder for myself.

[jsykora]

I don't know if this is any easier, or more practical, but maybe instead of the filter to convert \n to <text:line-break />, it'd be better to use the Django provided linebreaksbr tag and then convert all <br /> in a document to the xml line break.

Does that make sense? I guess the way I see it is that XML doesn't care about any <br />, so maybe if it's anywhere in a document, and already declared safe, it should be converted to the proper break without the programmer having to explicitly call a filter or load the custom template tags somewhere in the document.

[imankulov]

Regarding your last proposal. Frankly, It seems to me as too unobvious and incomplete. If we convert <br /> tags to <text:line-break />, why we wouldn't convert <em>, <strong>, <p>, <img>, etc the same way? At the end this functionality is a half-way to HTML to ODF converter which is not the goal of the project.

So I'd rather stay with template filter implementation.

[jsykora]

That makes perfect sense. I seemed to have entangled the actual scope of the project with how I use it or how I envision it being used.

In other news, the filter I posted above has been working well- I guess it helps that it is practically actual Django code. I just wish I didn't have to use the {% load %} tag in every template.

Thanks for your work on this project, it's awesome.

[jsykora]

I was having a problem with a couple pdfs I was trying to generate. After a lot of line by line debugging, I realized that the Objects which were causing problems were those which had '&' in the attributes that I was running through this line break filter.

The solution was to use escape option with the 'linebreaksodt' filter I defined above. I switched the autoescape kwarg to True which then made me realize I hadn't fully imported the necessary functions, so i had to add a few things to the imports.

The updated Gist is here: Gist 1676645

[vmassuchetto]

<text:line-break /> was printed after rendering instead of showing up as a line break. Any ideas?

[monkee-ch]

Am i right with my conclusion that this project is practically dead?

Many Thanks to @jsykora that line-break Gist works perfectly!

This helped me implementing it: https://docs.djangoproject.com/en/1.8/howto/custom-template-tags/

[jsykora]

@monkee-ch for what it's worth, I'm still using this in some internal apps, but migrating away from it. Mainly because the Open Office daemon is a pain. I run into a lot of crashes and issues with temp files pilling up. I have cron jobs to restart OO every hour.

[monkee-ch]

So far i use Abiword, and that works fine so far...

Only Problem i have now is that it somehow chokes on Special Characters like äöü - Any idea how to fix that?

Or do you recommend using something completely different than webodt?

[jsykora]

I don't really have to deal with special characters, but that may have been a limitation of abiword and googledocs processors. I needed to use Open Office for some more advanced tables anyway. (4 year ago, they might be better now).

I started using z3c.rml which is an open source implementation of Reportlab's RML.
Fairly high learning curve to get it to do what you want but it's been nice for me.

[monkee-ch]

I fixed my special Char Problem, just force the script to UTF8

# encoding=utf8
import sys
reload(sys)
sys.setdefaultencoding('utf8')`

Thanks for your Support and Tipps!