Pagination issue with Haystack

Question

I am using alphabetical pagination with haystack that I borrowed from https://djangosnippets.org/snippets/1364/

But named pagination is hitting the database which it shouldn't. Line 15 is causing this, although I have same iteration code in a django template that that works fine.

UPDATED:

search_view.py

import json
from django.core import serializers

from django.core.paginator import InvalidPage, Paginator
from django.http import Http404, HttpResponse
from django.shortcuts import redirect, render_to_response

from haystack.query import SearchQuerySet
from haystack.views import FacetedSearchView

from directory.named_paginator import NamedPaginator


class FacetedSearchCustomView(FacetedSearchView):
"""
Overrides various default methods to allow for additional context, smoother UX for faceting
"""

def build_page(self):
    """
    Paginates the results appropriately.

    Overriden to redirect to page 1 if a page_no is not found
    """
    try:
        page_no = int(self.request.GET.get('page', 1))
    except (TypeError, ValueError):
        raise Http404("Not a valid number for page.")

    if page_no < 1:
        raise Http404("Pages should be 1 or greater.")

    paginator = NamedPaginator(self.results, on="brand", per_page=self.results_per_page)
    # import pdb; pdb.set_trace()
    try:
        page = paginator.page(page_no)
    except InvalidPage:
        # Redirect to page 1 of the
        path = self.request.path
        qs = self.request.GET.copy()
        qs['page'] = 1
        url = '%s?%s' % (path, qs.urlencode())
        return redirect(url)

    return paginator, page

def clean_filters(self):
    """Returns a list of tuples (filter, value) of applied facets"""
    filters = []
    # get distinct facets
    facets = list(set(self.form.selected_facets))
    for facet in facets:
        if ":" not in facet:
            continue
        field, value = facet.split(":", 1)
        field = field.replace('_', ' ').replace('exact', '').title()
        filters.append((field, value))
    return filters

def create_response(self):
    """
    Generates the actual HttpResponse to send back to the user.

    Overriding to allow the redirect to pass through from overriden build_page
    """
    try:
        (paginator, page) = self.build_page()
    except ValueError:
        return self.build_page()

    context = {
        'query': self.query,
        'form': self.form,
        'page': page,
        'paginator': paginator,
        'suggestion': None,
    }

    if self.results and hasattr(self.results, 'query') and self.results.query.backend.include_spelling:
        context['suggestion'] = self.form.get_suggestion()

    context.update(self.extra_context())
    return render_to_response(self.template, context, context_instance=self.context_class(self.request))

def extra_context(self):
    extra = super(FacetedSearchCustomView, self).extra_context()
    extra['filters'] = self.clean_filters()
    if not self.results:
        extra['facets'] = self.form.search().facet_counts()
    else:
        extra['facets'] = self.results.facet_counts()

    model_type = self.request.path.split('/')[1].rstrip('s')
    extra['model_type'] = None if model_type == "search" else model_type

Any help will be much appreciated.

Thanks.

UPDATED:

{  
    "store":{  
        "mappings":{  
            "modelresult":{  
                "_boost":{  
                    "name":"boost",
                    "null_value":1
                },
                "properties":{  
                    "brand":{  
                        "type":"string",
                        "analyzer":"synonym_analyzer"
                    },
                    "brand_auto":{  
                        "type":"string",
                        "analyzer":"edgengram_analyzer"
                    },
                    "brand_slug":{  
                        "type":"string",
                        "analyzer":"synonym_analyzer"
                    },
                    "brand_sort":{  
                        "type":"string",
                        "analyzer":"synonym_analyzer"
                    }
                }
            }
        }
    }
}

Make sure that you pass the SearchQuerySet and not model's queryset to the Paginator. It would be better if you shared your code. — Animesh Sharma, May 20 '15 at 09:40
@AnimeshSharma updated with search view code, named_paginator.py is exactly the same as in the snippet. — Chirdeep Tomar, May 20 '15 at 16:35

score 0 · Answer 1 · answered May 20 '15 at 18:15

0

I guess the issue is with the indexing. You are trying to perform the pagination on the field "brand" and it is most likely that this field is not indexed in Solr/Elastic search.

Solr performs a DB query only for the fields which are not indexed.

Hope this sorts out the issue.

answered May 20 '15 at 18:15

Animesh Sharma

3,258
1
17
33

Brand is indexed, added the mapping. Also I am accessing the brand property in the django template and it still works. – Chirdeep Tomar May 20 '15 at 19:53
Try putting this in build_page before calling the paginator. `start_offset = (page_no - 1) * self.results_per_page` `self.results[start_offset:start_offset + self.results_per_page]` – Animesh Sharma May 21 '15 at 06:14
Added those two lines before creating an instance of NamedPaginator and still hitting the database. – Chirdeep Tomar May 21 '15 at 06:48
Beats me. Do post the solution when you find it. – Animesh Sharma May 21 '15 at 06:55

Pagination issue with Haystack

1 Answers1