annotator 0.13.2

Database backend for Annotator (http://annotatorjs.org)

Annotator Store

This is a backend store for Annotator.

The functionality can roughly be separated in two parts:

1. An abstraction layer wrapping Elasticsearch, to easily manage annotation storage. It features authorization to filter search results according to their permission settings.

2. A Flask blueprint for a web server that exposes an HTTP API to the annotation storage. To use this functionality, build this package with the [flask] option.

Getting going

You’ll need a recent version of Python (Python 2 >=2.6 or Python 3 >=3.3) and ElasticSearch (>=1.0.0) installed.

The quickest way to get going requires the pip and virtualenv tools (easy_install virtualenv will get them both). Run the following in the repository root:

virtualenv pyenv
source pyenv/bin/activate
pip install -e .[flask]
cp annotator.cfg.example annotator.cfg
python run.py

You should see something like:

* Running on http://127.0.0.1:5000/
* Restarting with reloader...

If you wish to customize the configuration of the Annotator Store, make your changes to annotator.cfg or dive into run.py.

Additionally, the HOST and PORT environment variables override the default socket binding of address 127.0.0.1 and port 5000.

Store API

The Store API is designed to be compatible with the Annotator. The annotation store, a JSON-speaking REST API, will be mounted at /api by default. See the Annotator documentation for details.

Running tests

We use nosetests to run tests. You can just pip install -e .[testing], ensure ElasticSearch is running, and then:

$ nosetests
......................................................................................
----------------------------------------------------------------------
Ran 86 tests in 19.171s

OK

Alternatively (and preferably), you should install Tox, and then run tox. This will run the tests against multiple versions of Python (if you have them installed).

Please open an issue if you find that the tests don’t all pass on your machine, making sure to include the output of pip freeze.

0.13.2

• Avoid a confusing error about reindexing when annotator is used as a library and not a standalone application (#107).

0.13.1

• Reindexer can run even when target exists.

0.13

• Slight changes to reindex.py to ease subclassing it.

0.12

• A tool for migrating/reindexing elasticsearch (reindex.py) was added (#103).

• The store returns more appropriate HTTP response codes (#96).

• Dropped support for ElasticSearch versions before 1.0.0 (#92).

• The default search query has been changed from a term-filtered “match all” to a set of “match queries”, resulting in more liberal interpretations of queries (#89).

• The default elasticsearch analyzer for annotation fields has been changed to “keyword” in order to provide more consistent case-sensitivity behaviours (#73, #88).

• Made Flask an optional dependency: it is now possible to use the persistence components of the project without needing Flask (#76).

• Python 3 compatibility (#72).

0.11.2

• SECURITY: Fixed bug that allowed authenticated users to overwrite annotations on which they did not have permissions (#82).

0.11.1

• Fixed support for using ElasticSearch instances behind HTTP Basic auth

0.11.0

• Add support for ElasticSearch 1.0

• Create changelog