I have successfully made django-haystack with elasticsearch to work.
In the example below, I can search for any sales item and it would show up in the results.
I have created an Index:
class SalesItemIndex(SearchIndex, Indexable):
text = CharField(document=True, use_template=True)
item_description = CharField(model_attr='item_description')
def get_model(self):
return SalesItem
def index_queryset(self):
"""Used when the entire index for model is updated."""
return self.get_model().objects.all()
The model SalesItem:
class SalesItem(models.Model):
item_description = models.CharField(_(u"Item Description"), max_length=40)
company = models.ForeignKey(Company)
def __unicode__(self):
return self.item_description
The problem is though, the user can search for all sales items, even those that don't belong to his company.
Usually instead of returning all salesitems = SalesItem.objects.all() I would rather use this to make sure the user only sees the items that belons to his company:
profile = request.user.get_profile()
sales_items = profile.company.salesitem_set.all()
Would I be able to limit my search with this rule?
You could index the company id in the SalesItemIndex as well:
class SalesItemIndex(SearchIndex, Indexable):
...
company = IntegerField(model_attr='company_id')
And filter it directly in SearchQuerySet:
sales_items = salesitem_set.filter(company=profile.company_id)
Related
I am trying to delete one of the items of a field in the table.
For example, I want to remove the 'hadi', shown in the image, from the table.
I make query to filter and exclude instance but in final part, delete() query does not accept input such as id or variable to remove specific thing in one field.
image
models.py
class Expense(models.Model):
expenser = models.ForeignKey(
settings.AUTH_USER_MODEL,
on_delete=models.CASCADE,
)
debtors = models.ManyToManyField(
CustomUser,
related_name='debtors',
)
amount = models.PositiveIntegerField()
text = models.TextField()
date = models.DateField(null=True)
time = models.TimeField(null=True)
def get_debtors(self):
return ", ".join([str(d) for d in self.debtors.all()])
def who_has_no_paid(self):
return ", ".join([str(d) for d in self.debtors.all() if str(d) != str(self.expenser)])
def debt(self):
return int(float(self.amount/self.debtors.all().count()))
def get_absolute_url(self):
return reverse('expense_detail', args=[self.id])
Now what is the query that should be used in view?
Since it is a many to many field you need to use expense_obj.debtors.remove(hadi_obj) rather than using delete().
Documentation is really good for referencing
your view could be :
def remove_hadi(request):
expense_obj = Expense.objects.get(id=some_id) #get the expense obj from where u want to remove hadi
hadi_obj = User.objects.get(name="hadi") # get user obj u want to remove eg: hadi
# use m2m query
expense_obj.debtors.remove(hadi_obj)
return response
Why target_dialogue is always None?
Model:
class Dialogue(models.Model):
name = models.CharField(max_length=30, blank=True)
is_conference = models.BooleanField(default=False)
participants = models.ManyToManyField(
Person,
related_name='dialogues',
)
def __str__(self):
return self.name or str(self.pk)
And in view I want to get suitable dialogue which contain in participants field 2 objects - user and companion. And if this dialogue doesn't exist I create it:
target_dialogue = None
try:
target_dialogue = Dialogue.objects.get(is_conference=False,participants__in=[user, companion])
except ObjectDoesNotExist:
target_dialogue = Dialogue()
target_dialogue.save()
target_dialogue.participants.add(user)
target_dialogue.participants.add(companion)
finally:
return render(request, 'dialogues/dialogue.html', {
'dialogue': target_dialogue,
})
But target_dialogue is always None. What's a reason of it? I was supposed to solve only a trouble in getting a dialogue from db in order to bad filter parameters, but now I have doubts about it. Maybe something else?
request.user is not a object of Person model with which you have the relation in Dialogue.
You have to first fetch the person object:
user = Person.objecs.get(user=request.user). # According to your person model
Follow same for companion and then query:
target_dialogues = Dialogue.objects.filter(is_conference=False,participants__in=[user,companion]
What is the best way to limit only one record to be default in django
I have a model where i have a flag for default
class BOMVersion(models.Model):
name = models.CharField(max_length=200,null=True, blank=True)
material = models.ForeignKey(Material)
is_default = models.BooleanField(default=False)
I want to have only one value to be default for the same material but this material can have a lot of non default ones.
It was odd that this question was not addressed more often. If I have a default record, I want to record that in the class as a member variable. And to determine if an instance is the default, I want to compare the class default member variable with the instance id. Unfortunately I could not figure out how to access class variables and instance variables nicely in the same class function in Python (may be someone can comment), but this does not require to hit the database for a default or store a bunch of records pointing to a default. Just one member variable in the class.
After I wrote this, I realized every time the application is restarted, default is reset to None, so you will have to store this in a database. I have updated my answer accordingly. However, checking that the member variable is not null, and only hitting the database if it is would reduce hits here. The model I used was:
class RecordOfInterest(models.Model):
"""
Record Records of interest here. Stores ID, and identifying character
"""
# asume maximum 64 character limit for the model.
model_name = models.CharField(max_length=64, unique=True)
record_no = models.IntegerField()
human_ident = models.CharField(max_length=64, help_text='How is this of interest')
# Id it as default, deposit, ... Don't bother indexing, as there will only be a few
def __unicode__(self):
return u'Model %s record %d for %s' % (self.model_name, self.record_no, self.human_ident)
class Meta:
unique_together = ('model_name', 'human_ident')
class Product(models.Model):
"""
Allow one product record to be the default using "Product.default = prod_instance"
check default with "Product.is_default(prod_instance)"
"""
default = None # set this to id of the default record
cart_heading = models.CharField(max_length=64, blank=True)
country = CountryField()
pricing = models.ForeignKey(
'Price', blank=True, null=True, related_name='visas', on_delete=models.SET_NULL)
#classmethod
def is_default(cls, obj):
if cls.default_no == None:
try:
cls.default_no = RecordOfInterest.objects.get(model_name=cls.__name__, human_ident='default')
except RecordOfInterest.DoesNotExist:
default_no = None
return cls.default_no == obj.id
#classmethod
def set_default(cls, obj):
try:
default_rec = RecordOfInterest.objects.get(model_name=cls.__name__, human_ident='default')
except RecordOfInterest.DoesNotExist:
RecordOfInterest.objects.create(model_name=cls.__name__, record_no=obj.id, human_ident='default')
else:
if default_rec.record_no != obj.id:
default_rec.record_no = obj.id
default_rec.save()
cls.default_no = obj.id
return
Saving the ID in settings.py if it is static.
Save it into a separate "default" table with one record (or use the most recent) if it's dynamic.
Save the default in another table like this:
class BOMVersion(models.Model):
name = models.CharField(max_length=200,null=True, blank=True)
material = models.ForeignKey(Material)
class BOMVersionDefault(model.Models)
time_set= models.Datetime(auto_created=True)
default_material = models.ForiegnKey(Material)
To query:
default = BOMVerDefault.objects.latest(time_set).get().default_material
If you have several material types that each need a default then default_material would be a field in a material-type table.
Getting one record to be default in a table is most basic requirement we developer come face to face, after spending couple of hours over it, i think a neat and clean solution in django would be update all records to default false if current form instance has default value to be "true" and then save the record.
class FeeLevelRate(TimeStampedModel):
"""
Stores the all the fee rates depend on feelevel
"""
feelevel = models.ForeignKey(FeeLevel, on_delete= models.PROTECT)
firstconsultfee = models.DecimalField(_('First Consultation Charges'),max_digits=10,decimal_places=2,blank=True)
medcharges = models.DecimalField(_('Medicines Charges per Day'),max_digits=10,decimal_places=2,blank=True)
startdate = models.DateField(_("Start Date "), default=datetime.date.today)
default_level = models.BooleanField(_('Is Default Level?'),default=False)
class Meta:
constraints = [
models.UniqueConstraint(fields=["feelevel","startdate"], name='unique_level_per_date'),
]
def __str__(self):
return "%s applicable from (%s)" % ( self.feelevel, self.startdate.strftime("%d/%m/%Y"))
class FeeLevelRateCreate(CreateView):
model = FeeLevelRate
fields = ['feelevel', 'firstconsultfee', 'medcharges', 'startdate', 'default_level']
context_object_name = 'categories'
success_url = reverse_lazy('patadd:feerate_list')
def form_valid(self, form):
# Update all the default_level with false.
#UserAddress.objects.filter(sendcard=True).update(sendcard=False)
if form.instance.default_level:
FeeLevelRate.objects.filter(default_level=True).update(default_level=False)
return super().form_valid(form)
What I'm trying to do is a 2 tier search with drop down menus using Select widget, the results will be a listing of the fields from my Meta.model. the first Tier is a a State listing from State.model. Upon a select it is supposed to list out all of the cities with in the selected state, the problem I'm having (and I think its due to my lack of knowledge) is that the city listing is not filtered but a listing of all cities in my database regardless of state. I'm not sure where or how to pass my variable to be able invoke my .filter() statement.
models.py
class Meta(models.Model):
rcabbr = models.CharField(max_length = 15)
slug = models.SlugField(unique=False)
state = models.ForeignKey('State')
rc_state = models.CharField(max_length = 3)
oerp = models.CharField(max_length=18)
subgrp = models.SlugField()
sonus_pic = models.CharField(max_length=8)
ems = models.CharField(max_length=14)
agc = models.CharField(max_length=14)
def __unicode__(self):
return self.rcabbr
class State(models.Model):
name = models.CharField(max_length=2)
slug = models.SlugField(unique=True)
state_long = models.CharField(max_length=15)
owning_site = models.CharField(max_length=12)
def __unicode__(self):
return self.name
return self.state_long
forms.py
class states(forms.Form):
invent = [(k.name,k.state_long) for k in State.objects.all()]
rclist = forms.ChoiceField(widget=forms.Select, choices=invent)
class rateCenter(forms.Form):
invention = [(k.id,k.rcabbr,k.rc_state) for k in Meta.objects.all()]
rcviews = forms.ChoiceField(widget=forms.Select, choices=invention)
views.py
def StateAll(request):
""" This lists out all of the states within the database that
are served"""
rclist = states()
return render(request, 'statelist.html',{'rclist': rclist})
def RcView(request):
""" this should list all the rateCenters within
the state that was selected in StateAll() """
rclist = request.GET['rclist']
forms = rateCenter()
return render(request, 'rclist.html',{'forms': forms})
Logic tells me I should to do my .filter() statement in the forms.py but unsure how to pass the result from the request.GET in StateAll() view. I do have the debug_toolbar installed so I can see the variable u'rclist' and the value u'LA' (my test state). I had this working 100% using hyperlinks however the size of my test database is miniscule in comparison to what is going to be in the production version and HREF's are just not possible.
my understanding is:
ChainedForeignKey(LinkedModel, LinkedModel.field = "field in first Tier", chained_model_field = "current model_field")
so simple model should I think be something like this?
def State(model.models):
name = models.CharField(max_length=20) #this is the state abbreviation
state_long = models.CharFeild(max_length=20)#this is state long form
def Meta(model.models):
state = models.CharField(max_length=20)
slug = models.SlugField(unique = False) #same values as rcabbr
rcabbr = ChainedForeignKey(State, chained_field = "state_long",
chained_model_field = "slug")
.....
Does that look about right........so the First Field in the drop down should be the State_long, once selected the next should be the slug?. at which time the slug should be passed to my urls and the views for the that final page.
I am going to try this however I'm not 100% sure how to do my views and if I need to do something with forms page or does this cover it? The documentation is not user friendly for someone new to this so any input would be most appreciated!
There are many third party libraries django-smart-selects and dajax come to mind - that will automate this for you along with provide you the necessary javascript to filter the form fields on the fly.
If you are investigating those, here is how you would do it with just the django forms:
class RateCenterForm(forms.Form):
rate_center = forms.ModelChoiceField(queryset=Meta.objects.all())
def __init__(self, *args, **kwargs):
state = kwargs.pop('state')
super(RaterCenterForm, self).__init__(*args, **kwargs)
self.fields['rate_center'].queryset = Meta.objects.filter(state=state)
A ModelChoiceField is a select drop down that takes its values from a model.
Now in your view, you would call it like this:
def show_rate_centers(request):
form = RateCenterForm(state='SomeState')
# .. your normal logic here
How to create simple search engine?
I have something like this:
def search(request):
if 'search' in request.GET:
term = request.GET['search']
if len(term) > 3:
d = Data.objects.filter(Q(content__contains=term) | Q(
desc__contains=term))
count = d.count()
return render_to_response('search_result.html', {'d': d, 'count': count}, context_instance=RequestContext(request))
return render_to_response('search_result.html', context_instance=RequestContext(request))
This is ok if I search in model but I need search in html content (I use django-chunks)
Django chunks data is stored in a model too. You can import it and filter it like any other model
class Chunk(models.Model):
"""
A Chunk is a piece of content associated
with a unique key that can be inserted into
any template with the use of a special template
tag
"""
key = models.CharField(_(u'Key'), help_text=_(u"A unique name for this chunk of content"), blank=False, max_length=255, unique=True)
content = models.TextField(_(u'Content'), blank=True)
description = models.CharField(_(u'Description'), blank=True, max_length=64, help_text=_(u"Short Description"))
class Meta:
verbose_name = _(u'chunk')
verbose_name_plural = _(u'chunks')
def __unicode__(self):
return u"%s" % (self.key,)
contains is an ok way to query if you have an extremely small dataset. It just issues a LIKE query, so it will not scale very well. If you are going to have a significant amount of data it will be a good idea to look for an engine that was created specifically for full text searching.