Skip to content

Latest commit

 

History

History
133 lines (100 loc) · 4.74 KB

README.md

File metadata and controls

133 lines (100 loc) · 4.74 KB

Dead simple Python package for analyzing PYBOSSA project's results

Build Status Coverage Status PyPi Downloads Version PyPi Downloads Month

Makes it easy to statistically analyze PYBOSSA project results.

Install

You can install enki using pip, preferably while working in a virtualenv:

    $ pip install enki

Requirements

PYBOSSA server >= v1.2.0.

Usage

It is really simple:

    >>> import enki

    # setup the server connection
    >>> e = enki.Enki(api_key='your-key', endpoint='http://server',
                      project_short_name='your-project-short-name',
                      all=1)
    # Get all completed tasks and its associated task runs
    >>> e.get_all()

NOTE: You should use the all=1 if you are not the owner of the project and you have an admin API key, otherwise the PYBOSSA context will return a 404 as the project does not belong to you.

The previous command, loads all completed tasks and task runs into four variables:

  • e.tasks a list of tasks
  • e.task_runs a dictionary of task runs, where the keys are the project task IDs
  • e.tasks_df a Pandas list of data frames for the tasks
  • e.task_runs_df a Pandas dictionary of data frame for the task runs, where the keys are the project task IDs

Now that you have downloaded all the tasks and task runs, you can start analyzing them using Pandas_:

    # For example, for a given task of your project:
    >>> task = e.tasks[0]
    # Let's analyze it (note: if the answer is a simple string like 'Yes' or 'No'):
    >>> e.task_runs_df[task.id]['info'].describe()
    count       1
    unique      1
    top       Yes
    freq        1
    dtype: object

    # Otherwise, if the answer in info is a dict: info = {'x': 32, 'y': 24}
    # Enki explodes the info field, using its keys (x, y) for new data frames:
    >>> e.task_runs_df[task.id]['x'].describe()
    count    100.000000
    mean     265.640000
    std        4.358945
    min      235.000000
    25%      264.000000
    50%      266.000000
    75%      268.000000
    max      278.000000
    dtype: float64

Enki explodes the task_run info field if it is a dictionary (a JSON object). This will help you to analyze more easily for example, all the keys of the object via Pandas statistical solutions. All you have to do is to access the key and use Pandas methods.

WARNING: If a task or taskrun has inside the info field a key that is protected, it will be escaped. For example, if exists task.info.id == 13, then, enki will escape it to _id. Enki will do it with all the protected attributes of Tasks and TaskRuns.

NOTE: if you want to load partial results, you can do it. Instead of using e.get_all() method, use the following code:

e.get_tasks(state='ongoing')
e.get_task_runs()

That will get the partial results. Then you can proceed with the analysis as before.

Using PYBOSSA JSON files

PYBOSSA exports the tasks and task runs as ZIP files in JSON format. You can pass those files to Enki, and avoid using the API for a faster analysis. If that's the case, download both files (task and task runs) and import them:

    >>> import enki

    # setup the server connection
    >>> e = enki.Enki(api_key='your-key', endpoint='http://server',
                  project_short_name='your-project-short-name')
    # Get all completed tasks and its associated task runs
    e.get_tasks(json_file='path/to/your/tasks.json')
    e.get_task_runs(json_file='path/to/your/task_runs.json')

Then you can do the analysis as before.

NOTE: If you want to anlayze a project from another person in the same server, then you can use the all=1 parameter when creating the Enki object:

    >>> import enki

    # setup the server connection
    >>> e = enki.Enki(api_key='your-key', endpoint='http://server',
                  project_short_name='your-project-short-name', all=1)
    # Get all completed tasks and its associated task runs
    e.get_all()

This will ensure that Enki will search across all projects in PYBOSSA.

Contributing

Please, see CONTRIBUTING file

Copyright

2016 Copyrigth Scifabric LTD.

License

AGPLv3.0 see COPYING file.