Whoosh indexing capabilities for Flask-SQLAlchemy
Latest commit 6977b74 Aug 19, 2016 @gyllstromk committed on GitHub Merge pull request #37 from huxuan/master
Update README for custom analyzer.


Welcome to Flask-WhooshAlchemy!

Flask-WhooshAlchemy is a Flask extension that integrates the text-search functionality of Whoosh with the ORM of SQLAlchemy for use in Flask applications.

Source code and issue tracking at GitHub.

View the official docs at http://packages.python.org/Flask-WhooshAlchemy/.


pip install flask_whooshalchemy


git clone https://github.com/gyllstromk/Flask-WhooshAlchemy.git


Let's set up the environment and create our model:

from whoosh.analysis import StemmingAnalyzer
import flask.ext.whooshalchemy

# set the location for the whoosh index
app.config['WHOOSH_BASE'] = 'path/to/whoosh/base'
# set the global analyzer, defaults to StemmingAnalyzer.
app.config['WHOOSH_ANALYZER'] = StemmingAnalyzer()

class BlogPost(db.Model):
  __tablename__ = 'blogpost'
  __searchable__ = ['title', 'content']  # these fields will be indexed by whoosh
  __analyzer__ = SimpleAnalyzer()        # configure analyzer; defaults to
                                         # StemmingAnalyzer if not specified

  id = app.db.Column(app.db.Integer, primary_key=True)
  title = app.db.Column(app.db.Unicode)  # Indexed fields are either String,
  content = app.db.Column(app.db.Text)   # Unicode, or Text
  created = db.Column(db.DateTime, default=datetime.datetime.utcnow)

Only four steps to get started:

(Actually, only the third one is required for using, others are all optional.)

  1. Set the WHOOSH_BASE to the path for the whoosh index. If not set, it will default to a directory called 'whoosh_index' in the directory from which the application is run.
  2. Set the WHOOSH_ANALYZER to the global analyzer. If not set, it will defalt to StemmingAnalyzer.
  3. Add a __searchable__ field to the model which specifies the fields (as str s) to be indexed .
  4. Add a __analyzer__ field to the model if you need a local custom analyzer for indexing.

Let's create a post:

    BlogPost(title='My cool title', content='This is the first post.')
); db.session.commit()

After the session is committed, our new BlogPost is indexed. Similarly, if the post is deleted, it will be removed from the Whoosh index.

Text Searching

To execute a simple search:

results = BlogPost.query.whoosh_search('cool')

This will return all BlogPost instances in which at least one indexed field (i.e., 'title' or 'content') is a text match to the query. Results are ranked according to their relevance score, with the best match appearing first when iterating. The result of this call is a (subclass of) :class:`sqlalchemy.orm.query.Query` object, so you can chain other SQL operations. For example:

two_days_ago = datetime.date.today() - datetime.timedelta(2)
recent_matches = BlogPost.query.whoosh_search('first').filter(
    BlogPost.created >= two_days_ago)

Or, in alternative (likely slower) order:

recent_matches = BlogPost.query.filter(
    BlogPost.created >= two_days_ago).whoosh_search('first')

We can limit results:

# get 2 best results:
results = BlogPost.query.whoosh_search('cool', limit=2)

By default, the search is executed on all of the indexed fields as an OR conjunction. For example, if a model has 'title' and 'content' indicated as __searchable__, a query will be checked against both fields, returning any instance whose title or content are a content match for the query. To specify particular fields to be checked, populate the fields parameter with the desired fields:

results = BlogPost.query.whoosh_search('cool', fields=('title',))

By default, results will only be returned if they contain all of the query terms (AND). To switch to an OR grouping, set the or_ parameter to True:

results = BlogPost.query.whoosh_search('cool', or_=True)