# Elasticmodels
Elasticmodels helps you index and query your Django models using elasticsearch.
It is designed to be an alternative to django-haystack when you need more control over
your index creation, and you are always going to use elasticsearch.
# Install
pip install elasticmodels
# Setup
## settings.py
In your Django settings file, define these variables:
```python
ELASTIC_SEARCH_CONNECTION = {
"urls": ["http://localhost:9200/"],
"index": "the_name_of_your_es_index",
# "http_auth": "username:password",
}
# these are used when your index is created
ELASTIC_SEARCH_SETTINGS = {
"settings": {
"analysis": {
"analyzer": {
"snowball": {
"type": "snowball",
"stopwords": "_none_"
}
}
}
}
}
```
Add elasticmodels to INSTALLED_APPS:
```python
INSTALLED_APPS = (
...
'elasticmodels',
)
```
## app/search_indexes.py
In a Django app, create a search_indexes.py file, like so:
```python
from elasticmodels import Indexable
from django.template.loader import render_to_string
from .models import File, FileTag
class FileIndex(Indexable):
# specify the model class this index is for
model = File
def mapping(self):
"""
Return the elasticsearch mapping for this model type
"""
return {
"properties": {
"pk": {"type": "integer", "index": "not_analyzed"},
"content": {"type": "string", "analyzer": "snowball"},
"tags": {"type": "string", "analyzer": "keyword"},
"org_id": {"type": "integer", "index": "not_analyzed"},
"type": {"type": "integer", "analyzer": "keyword"},
"uploaded_by_id": {"type": "integer", "analyzer": "keyword"},
}
}
def prepare(self, obj):
"""
Return obj transformed into a dict that corresponds to the mapping
you defined. This is what will be indexed by elasticsearch.
"""
return {
"pk": obj.pk,
"content": render_to_string("files/search.txt", {"object": obj}),
"tags": [ft.tag.name for ft in FileTag.objects.filter(file=obj).select_related("tag")],
"org_id": obj.org_id,
"type": obj.type,
"uploaded_by_id": obj.uploaded_by_id,
}
```
# Usage
## Deleting and recreating your index
./manage.py rebuild_index
**This will delete the entire elasticsearch index** and recreate it. All your
model objects will be re-indexed.
## Adding an individual object to the index
```python
from elasticmodels import make_searchable
f = File(name="Foo", type=1)
f.save()
make_searchable(f)
```
## Querying
Your subclass of elasticmodels.Indexable has a class attribute called `objects`
which returns an elasticutils `S` instance. You can then use whatever methods are
available in elasticutils on the S instance.
See:
http://elasticutils.readthedocs.org/en/latest/searching.html
http://elasticutils.readthedocs.org/en/latest/searching.html#filters-filter
http://elasticutils.readthedocs.org/en/latest/searching.html#queries-query
http://elasticutils.readthedocs.org/en/latest/searching.html#advanced-filters-f-and-filter-raw
```python
from elasticutils import F
from .search_indexes import FileIndex
results = FileIndex.objects.filter(F(type=1) | F(type=2)).query(content__match="foo")
for result in results:
print result.pk, result.content
```
