{
    "info": {
        "author": "John Kleint",
        "author_email": "mediantracker-general@lists.sourceforge.net",
        "bugtrack_url": null,
        "classifiers": [
            "Development Status :: 5 - Production/Stable",
            "Intended Audience :: Developers",
            "Intended Audience :: Science/Research",
            "License :: OSI Approved :: MIT License",
            "Natural Language :: English",
            "Operating System :: OS Independent",
            "Programming Language :: Python :: 2",
            "Topic :: Scientific/Engineering :: Mathematics"
        ],
        "description": "Track the median of a stream of values \"on-line\" in reasonably efficient \nfashion.\n\nUsage::\n\n    from mediantracker import MedianTracker\n    tracker = MedianTracker()\n    tracker.add(1)\n    tracker.add(2)\n    tracker.add(3)\n    tracker.add(4)\n    assert tracker.lower_median() == 2\n    assert tracker.upper_median() == 3\n    \n``MedianTracker`` supports efficient median queries on and dynamic\nadditions to a list of values. It provides both the lower and upper median\nof all values seen so far. Any ``__cmp__()``-able object can be tracked,\nin addition to numeric types. ``add()`` takes *log(n)* time for a tracker\nwith *n* items; ``lower_median()`` and ``upper_median()`` run in constant\ntime. Since all values must be stored, memory usage is proportional to the\nnumber of values added (*O(n)*).\n\nThe values can be accessed via iteration over the ``MedianTracker``,\nthough they are not ordered in any particular way::\n\n    sum = 0.0\n    for val in tracker:\n        sum += val\n    mean = sum / len(tracker) \n\nUse this module if you are processing values \"on-line,\" one at a time, as\nthey arrive, and need to know the new median after each new value (or\nbatch of values). If you just want the median of a whole list, there are\nmuch more efficient linear-time median (or more general k-th smallest\nselection) algorithms. Using this module will take *O(nlogn)* time and an\nextra *O(n)* space to compute the median of *n* items. On the other hand,\na ``MedianTracker`` will only take *O(nlogn + m)* time for any sequence of\n*n* adds and *m* median queries, whereas running a traditional\nnon-incremental median algorithm *m* times would take *O(n*m)*.\n\nFinally, some sources define the median of an even-length list to be the\naverage of the middle two values. This is easily and efficiently computed\n(in constant time)::\n\n    tracker = MedianTracker([1, 2, 3, 4])\n    median = (tracker.lower_median() + tracker.upper_median()) / 2.0\n    assert median == 2.5",
        "description_content_type": null,
        "docs_url": null,
        "download_url": "UNKNOWN",
        "downloads": {
            "last_day": -1,
            "last_month": -1,
            "last_week": -1
        },
        "home_page": "http://sourceforge.net/projects/mediantracker/",
        "keywords": "median statistics online incremental dynamic mathematics",
        "license": "MIT License",
        "maintainer": null,
        "maintainer_email": null,
        "name": "mediantracker",
        "package_url": "https://pypi.org/project/mediantracker/",
        "platform": "UNKNOWN",
        "project_url": "https://pypi.org/project/mediantracker/",
        "project_urls": {
            "Download": "UNKNOWN",
            "Homepage": "http://sourceforge.net/projects/mediantracker/"
        },
        "release_url": "https://pypi.org/project/mediantracker/1.0/",
        "requires_dist": null,
        "requires_python": null,
        "summary": "Tracks the overall median of a stream of values \"on-line\" in reasonably efficient fashion.",
        "version": "1.0"
    },
    "last_serial": 794633,
    "releases": {
        "1.0": [
            {
                "comment_text": "",
                "digests": {
                    "md5": "77666357b7340811a8b4d995c9ce1135",
                    "sha256": "c94504b2e82354e7f1241bfb521c88681beef159a8e95af0145262f77f5beacd"
                },
                "downloads": -1,
                "filename": "mediantracker-1.0.tar.gz",
                "has_sig": false,
                "md5_digest": "77666357b7340811a8b4d995c9ce1135",
                "packagetype": "sdist",
                "python_version": "source",
                "requires_python": null,
                "size": 6409,
                "upload_time": "2009-09-02T22:27:12",
                "url": "https://files.pythonhosted.org/packages/44/94/d67a68b8adb5c16ade6ff2a3054ef1c4a48f0197015cf08630b75d6d4ae5/mediantracker-1.0.tar.gz"
            }
        ]
    },
    "urls": [
        {
            "comment_text": "",
            "digests": {
                "md5": "77666357b7340811a8b4d995c9ce1135",
                "sha256": "c94504b2e82354e7f1241bfb521c88681beef159a8e95af0145262f77f5beacd"
            },
            "downloads": -1,
            "filename": "mediantracker-1.0.tar.gz",
            "has_sig": false,
            "md5_digest": "77666357b7340811a8b4d995c9ce1135",
            "packagetype": "sdist",
            "python_version": "source",
            "requires_python": null,
            "size": 6409,
            "upload_time": "2009-09-02T22:27:12",
            "url": "https://files.pythonhosted.org/packages/44/94/d67a68b8adb5c16ade6ff2a3054ef1c4a48f0197015cf08630b75d6d4ae5/mediantracker-1.0.tar.gz"
        }
    ]
}