acangiani / django-tagging Goto Github PK

Methods which accept a tag currently require a Tag object.

Methods which accept lists of tags currently require a list of Tag objects
or a Tag QuerySet.

Make these methods more flexible and user friendly by also accepting tag
names or lists of tag names, loading the appropriate Tag objects as required.

I love django-tagging, but I had the need for tags with multiple words.  For 
example, if I wanted 
to tag an item as "stained glass", but didn't want it to break into two tags, I 
would have to use 
"stained-glass" or "stainedglass", neither of which I liked very much.

With just two tiny changes in utils.py, I was able to add the ability to 
designate a multiple word 
string as a single tag using quotes, without (as far as I can tell) affecting 
the previous 
functionality.  So now, the user can type 'window "stained glass" blue' 
(without single quotes) and 
the generated tags will be "window", "stained glass", and "blue".

Here's the changed section of my utils.py:
find_tag_re = re.compile('\"[-\w\s]+\"|[-\w]+', re.U)

def get_tag_name_list(tag_names):
    """
    Find tag names in the given string and return them as a list.
    """
    if not isinstance(tag_names, unicode) and tag_names is not None:
        tag_names = tag_names.decode(settings.DEFAULT_CHARSET)
        tag_names = re.sub('\s+', ' ', tag_names)
    results = find_tag_re.findall(tag_names or '')
    return [item.encode(settings.DEFAULT_CHARSET).replace('"', '') for item in results]

Note that I added the re.sub line to get rid of extra spaces that might crop up 
from this.

Hope this is useful to someone else.

Branton

>>> import re
>>> re.findall('\w+',',  , foo   ,   bar ,  ,baz, , ,')
['foo', 'bar', 'baz']
>>> 

So I have not started using this project, but I was curious does
django-tagging was able to keep track of the owners of the tags much like
del.icio.us?

`TagManager.update_tags` is using a call to `split` to break up the new tag 
list by spaces, but this 
can cause problems if a user inputs comma-separated tags. Compatibility with 
either type of 
separator could be achieved by using `re.split` instead:

{{{

import re
updated_tag_names = [t for t in re.split('[\s,]+', tag_list) if t]

}}}

This also has the advantage of automatically returning an empty list of tag 
names if tag_list was 
empty, removing the need for the 'if tag_list != None' check.

Right now it looks like it'd be tricky, if possible at all, to say "given this 
model, give me instances 
tagged with 'foo' AND 'bar'" -- which is a common use case -- because 
TaggedItemManager.get_for_model only accepts a single tag. Accepting a list of 
tags would make 
the DB query a bit more complicated (it would either turn into a pretty nasty 
join, or a 'tag_id IN (list 
of ids)', but would make this use case a whole lot simpler.

this issue was posted in django-users:
http://groups.google.com/group/django-users/browse_thread/thread/a28c1ada9c3f813
3/106c18b58e35f5c0#106c18b58e35f5c0

when fixing this i would suggest to add a is_disabled boolean to disable
unwanted tags posted by users and not have to delete them over and over again.

I get TypeError:

in django\template\__init__.py in resolve_variable, line 637

expected string or buffer.

I hunted it down, and it seems that there is some offending code in the
TagsForObjectNode class. You are setting the self.object property, and it
somehow (in my case) the resolve_variable is called again, trying to
resolve a model object, and fails. I have modified the render function to
resolve to a temp variable instead...

    def render(self, context):
        _object = resolve_variable(self.object, context)
        context[self.context_var] = Tag.objects.get_for_object(_object)
        return ''

Currently, `TagManager.update_tags` uses a set for the new tags, but doesn't 
check for the 
presence of built-in sets (which were only added in Python 2.4).

Compatibility to 2.3 could be achieved by testing for the presence of the 
built-in `set` type, and 
falling back to importing the `Set` class out of the `sets` module if it's not 
found, e.g., by 
adding this at the top of `models.py`:

{{{

try:
    set
except NameError:
    from sets import Set as set

}}}

Few months ago I posted a snippet on djangosnippets.org website:

http://www.djangosnippets.org/snippets/140/

It think that it could be useful to include it in the template tags

Regards,
Massimo

The last line of the model code currently reads;

tags = property(_get_tags, _set_tags)

which results in an undefined module error.

It should read;

tags = property('_get_tags', '_set_tags')

which now works perfectly.

It would be nice if django-tagging supported machine tags (e.g.
http://www.flickr.com/groups/api/discuss/72157594497877875/) by including a
namespace field.

I am starting to use your tagging module in a series of projects I am working 
on (I ditched my 
nascent tagging app because yours does most of the stuff that mine did, and 
more stuff that 
mine had not yet gotten to.)

One of the big barriers that I am overcoming by adding code to my projects is 
dealing with cross 
model use of tags.

For example, I have forums which have discussions with have posts in one app. I 
have wiki topics 
in another app. I have user pages elsewhere. I also have blogs.

All of these have tags. As a matter of fact a lot of them will share tags. I 
need common views that 
let you see _all_ of the objects (or all of the models) that have a given tag.

Luckily with the generic relations and the simple object structure I can add 
this code to a 
tagging-support app in my project that lets people view lists of objects that 
match a tag so they 
can find all the relevant discussions, posts, wiki topics, and blog entries 
that cover a specific tag.

Do you have any ideas or plans in this direction? (I am not sure this is an 
"Issue" but I do not see 
any other way to submit comments on google code.)

I think it would be good to have the option of having case insensitive
tags.  E.g., convert all tags to lowercase before storing, so that
"Something" and "something" would show as the same tag, which they are.  

I understand there might be cases where you would want them to be
different, but for my apps I want them to be treated as the same tag, so
users don't get confused.  

If you attempt to use get_by_model and the tag doesn't exists, it creates
an exception in the cursor.  I fixed it w/:

226         tags = get_tag_list(tags)
227         if len(tags) == 1:
228             tag = tags[0] # Optimisation for single tag
229         elif len(tags) == 0:
230             return None
231         else:
232             return self.get_intersection_by_model(Model, tags)

What steps will reproduce the problem?

e = Entry.objects.get(pk=12)
e.tags = '中文 语言'

What is the expected output? What do you see instead?

e.tags # output nothing

But it should be "中文 语言".

What version of the product are you using? On what operating system?

svn version, ubuntu system, mysql database.

Please provide any additional information below.

I want to integrate tagging framework into my project. Unfortunately I met
a problem saving non-latin(Chinese) characters tags name in database. If
tags name look like "aa bb cc", they can be parsed perfectly and saved into
tag table respectively. But if tags name are non-latin(Chinese) characters,
there is nothing happend in tag table.

There is a TagField named tags in my Entry model.

I think something wrong with the regex when parsing tags name maybe? or
something about unicode?

What steps will reproduce the problem?

our Entry model uses the TagField like so:

tags = TagField()

Using the python interpreter:

1. from tagging.models import TaggedItem, Tag
2. from myapp.blog.models import Entry
3. post = Entry.objects.get(pk=3)
4. post.tags = 'last monkey marketing'
5. post.tags

'last marketing monkey'

then, when you use the templatetag tags_for_object , the tags seem to
disappear from the TaggedItem table.

What version of the product are you using? On what operating system?

latest SVN export on Centos 4, apache mod_python, mysql

Please provide any additional information below.

if you go to the django admin and view the TaggedItem list, that will also
delete all TaggedItems from the system.  The tags remain, but the associations.

The attached patch (which is really just an added file) adds a TagField
object (in ``tagging.fields``) that allows more natural access to tags on
tagged items. You define it thusly::

    from django.db import models
    from tagging.fields import TagField

    class MyModel(models.Model):
        ...

        tags = TagField()

And then use it pretty naturally::

    >>> from myapp.models import MyModel
    >>> m = MyModel.objects.get(...)

    >>> m.tags = "foo bar"
    >>> m.tags
    "bar foo"

    >>> m.tags = "foo bar baz"
    >>> m.tags
    "bar baz foo"

Hi Jonathan,

Congrats on the birth of your child.  Thanks for the django-tagging app, it
really is the greatest thing since sliced bread.

One thing that I would like to do with it is the following (and if it's
just something that I should put in my own templatetag, that's fine with
me, I'm only putting this out there in case it can be generalised and be of
use to others):  I'd like to be able to get a list of two different kinds
of tagged objects that relate to each other.  This is best illustrated by
an example.

If I have a weblog entry with the tag-list: "django tagging buchanan" I'd
like to display related links with one or more of those tags.  Ideally, I'd
like a list of links arranged by relevance (so that links that match more
tags are listed first).

I hope I'm being clear :(

Thanks!

Seemant

As of Django changeset 5172, the fields for generic relations now live in 
django.contrib.contenttypes.generic instead of django.db.models.

Currently, for a multi-user blogging app, the generated tag cloud comes
from the aggregation of everyone's blog entries.  Would it be possible to
somehow add the capability to restrict the tag cloud to a subset of blog
entries (in this example)?

What steps will reproduce the problem?
1. Have a model with a slug as a primary key, do not have an id...
2. Add tags = TagField() to the model
3. Witness breakage

What is the expected output? What do you see instead?
Expected to render the template.

Actual:
'MyModel' object has no attribute 'id'
/home/user/Projects/mc/../mc/tagging/models.py in update_tags, line 21


What version of the product are you using? On what operating system?
SVN trunk - 4/4/07

Please provide any additional information below.
If you need anymore information. I'm more than willing to help.

What steps will reproduce the problem?

enter 'tétéphone othertag' in a tag list.
return ['t','t','phone','othertag'] list of tags.

What is the expected output? What do you see instead?

return ['téléphone','othertag'] list of tags.

What version of the product are you using? On what operating system?

a svn version, but didn't update since a few week.

Please provide any additional information below.

my proposition for solve this problem :

change in utils.py :

#find_tag_re = re.compile('[-\w]+')
char_accepted_in_tag = u'œ@àéèùçâäîïôöêëûüÄÂÛÜÎÏÔÖÊË'
find_tag_re = re.compile(u'[-%s\w]+' % char_accepted_in_tag)

change in validators.py :

#tag_list_re = re.compile(r'^[-\w]+(?:(?:,\s|[,\s])[-\w]+)*$' )
char_accepted_in_tag = u'œ@àéèùçâäîïôöêëûüÄÂÛÜÎÏÔÖÊË'
tag_list_re = re.compile(r'^[-%s\w]+(?:(?:,\s|[,\s])[-\w]+)*$' %
char_accepted_in_tag)

change in model.py :

add before TagManager
def decodeall(txt):
    try:
        return txt.decode('utf-8')
    except:
        return txt

in TagManager:
            #if tag_name not in current_tag_names:
            ucurrent_tag_names=map(decodeall,current_tag_names)
            if tag_name not in ucurrent_tag_names:


  It's work...
but for searching... 
I will need to téléphone with entry téléphone or telephone. 
(because url didn't accept accent)
then I will peharps need to change the structure.

It would be nice if TagManager.usage_for_model returned a queryset instead
of a list so that code such as

Tag.objects.usage_for_model(Item, counts=True).order_by('-count') 

would still work.  This used to work in previous revisions.

When using tags on a model, it's fairly easy to create a simple interface for 
retrieving and setting 
them by putting a property on the model:

{{{

def _get_tags(self):
    return Tag.objects.get_for_object(self):

def _set_tags(self, tag_list):
    Tag.objects.update_tags(self, tag_list)

tags = property(_get_tags, _set_tags)

}}}

So suppose you have a blog entry model with tags, and you add the above to it; 
now you could 
do:

{{{

e = Entry.objects.get(pk=12)
e.tags # prints the tag list
e.tags = 'foo bar' # now the Entry is tagged with 'foo' and 'bar'.

}}}

(thanks for releasing this, btw -- I was just sitting down to write a tagging 
implementation for an 
app I'm working on when I saw this app scroll by, and it's already got 
everything I wanted!)

Attached is a new templatetag I've made. I find it useful, maybe it's
useful for anyone else. The patch is attached.

    Retrieves a list of tags used for a given Model and stores them in 
    a context variable.

    The model is specified in ``[appname].[modelname]`` format.

    If specified it adds a count attribute to each tag. Just add an 
    extra parameter called ``count``

    Example usage::

        {% tags_for_model blog.comment as tag_list %}
        or 
        {% tags_for_model blog.comment as tag_list count %}

The attachment file is a patch for django-tagging to surpport non-ascii 
tag name. I've tested it with chinese.

in the file tagging.models,
in the function get_intersection_by_model :
            ids = cursor.fetchall()[0]
return only the first value for postgres database.

I use this code for patch postgres usage :
            resultall = cursor.fetchall()
            ids=[]
            for i in resultall:
                ids.append(i[0])

have a good day.

Xavier Manach. [email protected]

PS : thx for your application. 
It's help me a lot (i am noob in django).

What steps will reproduce the problem?
1. Try using get_by_model...

What is the expected output? What do you see instead?
If two tagged items (of the same model) both have two shared tags, and you
search for all objects of that model with the two tags, you'd expect the
two items to be returned, but you get only one.

After investigation, it seems that cursor.fetchall() returns:
    ((2L,), (3L,))  (where 2L and 3L are the pk ids)
And so cursor.fetchall()[0] dutifully returns
    (2L,)
which is wrong. The fix for my situation was
-   ids = cursor.fetchall()[0]
+   ids = [row[0] for row in cursor.fetchall()]

Perhaps this was changed in the django version? Perhaps I did something wrong?

Off the top of my head, it should be a simple wrapper around object_list
which sets up appropriate QuerySet.

Would be nice

Ivan Sagalaev's TagsField (http://softwaremaniacs.org/soft/tags/en/) uses a
little bit of JScript to yield Tag auto-complete.  Some form of
auto-complete might be handy to include in this tagging library.

Any model using django-tagging against the current development trunk of
django now fails with this error:
foo.app: 'module' object has no attribute 'GenericForeignKey'

GenericForeignKey has moved to the contenttypes module as of Django
revision 5172:
http://code.djangoproject.com/wiki/BackwardsIncompatibleChanges#Genericrelations
havemoved

Changing the import fixes it.  A try/except import block could handle both
the before and after case.

I just noticed that `models.py` imports `urllib`, but doesn't use it; it's 
probably safe to remove the 
import.

I'd like to display a tag cloud on a tag page with tags which have been
used with this tag. Hum, ok not really clear and you need an example:

post1.tags = 'foo bar'
post2.tags = 'bar baz'
post3.tags = 'foo ter'
post4.tags = 'bar ter'

Now, the foo tag page will contain a tag cloud with bar and ter tags. The
bar one with foo, baz and ter, and so one...

Let me know if it's easily feasible, maybe I can help.

I got the exception when i use the tagged_objects tag.then i found the error:
def do_tagged_objects(parser, token):
      if bits[2] != 'for':

is not 'for',is 'in'

In revision 57, line 166 of models.py reads:

        if len(tags) == 0:
            tag = tags[0] # Optimisation for single tag

Assuming the comment is correct, it should read:

if len(tags) == 1:

Because len(), like count/length in so many other languages starts with
one.  len(tags) == 0 should probably be an error condition.  Because this
"optimization" doesn't get called it doesn't look like it works either.

And release something for once!

i recently switched from sqlite to psql and got the following errors:

in related_for_model the second group by is missing the tag.name

          AND t.id NOT IN (%s)

        GROUP BY t.id, t.name

        ORDER BY 2 ASC""" % (


and the get_intersection_by_model doesnt work:

Exception Value:
ERROR: syntax error at or near ")" at character 354 SELECT
"blog_blogentry"."id","blog_blogentry"."title","blog_blogentry"."summary","blog_
blogentry"."tags","blog_blogentry"."created"
FROM "blog_blogentry" WHERE ("blog_blogentry"."id" IN ()) ORDER BY
"blog_blogentry"."created" DESC

its only throwing this error when there are no blogentries for the tags.

I needed to access in a template to the "cloud_for_model" function, so i
added a new template tag in the tagging_tags file.

Isn't this a good thing to add in the next version of the django-tagging app?

Future work is give the list of options in the template, and the template
tag return the correspondent value.


In the attach file is the code i used.

Thanks

What steps will reproduce the problem?
1. Use tagged_objects template tag in a template per README.txt.
   eg. {% tagged_objects foo_tag in blog.Entry as entries %}
2. Load template.
3. TemplateSyntaxError: second argument to tagged_objects tag must be 'in'

What is the expected output? What do you see instead?
Expected to see a rendered template, instead raises TemplateSyntaxError.

What version of the product are you using? On what operating system?
Checked out from svn Apr 9, 2007. Cygwin, Windows XP Pro.

Please provide any additional information below.
Line 61 of tagging_tags.py checks bits[2] != 'for'. Probably should be
bits[2] != 'in'. README.txt (and exception output!) defines syntax using 'in'.

hiho

say i have 1000 tags and limit the result to 
tag_cloud = Tag.objects.cloud_for_model(MyModel, steps=4)[:100]

the tagcloud shows tags from letter a to m for example, but it should show
up the 100 most used tags and sort them by name.

i don´t see how i can do this.

It would be useful provide update_tags functionality for additional tag
formats when importing data from other services.  Off the top of my head:

Flickr style: tag "tag with a space" othertag
CSV (LibraryThing, among others) style: tag?,tag with a space,othertag!

The Flickr double-quotes style could be added to the existing regex and by
doing a quick transform from spaces to slug-style hypens (-).  The regex
would then be something like:

find_tag_re = re.compile('(?P[-\w]+)|\"(?P[-\w ]+)\"', re.U)
(I added capturing groups to keep the "" quotes out of the returned tag.)

and then with the tags just do a quick tag_name.replace(' ','-')

For the CSV style it would be useful to provide some quick flag to switch
to "real CSV style" and run django's slugify() to clean gunk like
non-alphanumeric characters for the final tag name.

I can flesh this out into a full patch in a few days if there is interest.
 I'll probably be writing it for my own use in the near future anyway.

It would also be nice to support "verbose names" for tags where spaces
might look cleaner/nicer or be more preferred as the non-URL display name.

1. Implement an equal assignment algorithm in addition to the current
log-based algorithm - allow the user to choose which is used when
calculating a cloud.

2. Move the cloud calculation function out of the TagManager class - it now
operates on a given lists of tags, so it doesn't really fit there any more.

Reported on django-users:

http://groups.google.com/group/django-users/browse_thread/thread/79d079db754bfa6
3

Form control for editing tags with an autocomplete

P.S.: I found another project of tagging for Django (for if you need more
ideas to tagging)
http://softwaremaniacs.org/soft/tags/en/

What steps will reproduce the problem?
1. Install django-tagging as explained in README.txt
2. add line to import Tag in myproj.myapp.models: from
myproj.tagging.models import Tag
3. myapp disappears from Django admin.

What is the expected output? What do you see instead?

As soon as I add the line "from myproj.tagging.models import Tag" to my
model, that model disappears from the Admin.  The funny thing is that
everything works fine from the python shell.


What version of the product are you using? On what operating system?
Dreamhost.  Python 2.3.5.  Current svn of both Django and django-tagging.


Please provide any additional information below.

I tried to be far to clever when I first wrote the tag field. 

The end result is that it's currently *very* chatty with the database: doing 
something like 
Model.objects.all() (on an object with a TagField) causes O(N) database hits 
instead of the supposed 
O(1).

Attached is a patch that's a lot less clever, but also much more economical 
with database hits. It 
works just as well, but the minor difference is that now tag changes are only 
saved back to the 
database when object.save() gets called. I think this is more correct (it makes 
TagField behave much 
more like a CharField), but it is a difference from the previous behavior.

So I have not started using this project, but I was curious does
django-tagging was able to keep track of the owners of the tags much like
del.icio.us?

What would have to happen for django-tagging to support Unicode tags?