running_stats.py

'''Tool for a script to keep track changes performed on a large number
of objects.

StatsCount - when you are counting incidences of a small set of outcomes
StatsList - when you also want to remember an ID associated with each incidence

Examples:

from running_stats import StatsCount
package_stats = StatsCount()
for package in packages:
    if package.enabled:
        package.delete()
        package_stats.increment('deleted')
    else:
        package_stats.increment('not deleted')    
print package_stats.report()
> deleted: 30
> not deleted: 70
    
from running_stats import StatsList
package_stats = StatsList()
for package in packages:
    if package.enabled:
        package.delete()
        package_stats.add('deleted', package.name)
    else:
        package_stats.add('not deleted' package.name)
print package_stats.report()
> deleted: 30 pollution-uk, flood-regions, river-quality, ...
> not deleted: 70 spending-bristol, ... 

'''

import copy

class StatsCount(dict):
    # {category:count}
    _init_value = 0
    report_value_limit = 250
    
    def _init_category(self, category):
        if not self.has_key(category):
            self[category] = copy.deepcopy(self._init_value)
        
    def increment(self, category):
        self._init_category(category)
        self[category] += 1

    def report_value(self, category):
        '''Returns the value for a category and value to sort categories by.'''
        value = repr(self[category])
        if len(value) > self.report_value_limit:
            value = value[:self.report_value_limit] + '...'
        return (value, self[category])

    def report(self, indent=1, order_by_title=False):
        lines = []
        indent_str = '\t' * indent
        report_dict = dict()
        for category in self.keys():
            report_dict[category] = self.report_value(category)

        if order_by_title:
            items = sorted(report_dict.iteritems())
        else:
            items = sorted(report_dict.iteritems(),
                           key=lambda x: -x[1][1])

        for category, value_tuple in items:
            value = value_tuple[0]
            lines.append(indent_str + '%s: %s' % (category, value))
        if not self:
            lines = [indent_str + 'None']
        return '\n'.join(lines)

class StatsList(StatsCount):
    # {category:[values]}
    _init_value = []

    def add(self, category, value):
        self._init_category(category)
        self[category].append(value)
        return '%s: %s' % (category, value) # so you can log it too

    def report_value(self, category):
        value = self[category]
        number_of_values = len(value)
        value_str = '%i %r' % (number_of_values, value)
        if len(value_str) > self.report_value_limit:
            value_str = value_str[:self.report_value_limit] + '...'
        return (value_str, number_of_values)

class OpenSpendingStats(StatsList):
    def add_source(self, category, row, value='notspecified'):
        id = '%s %s %s' % (row['package_name'], row['resource_id'], row['url'])
        if value != 'notspecified':
            id += ' %r' % value
        return self.add(category, value=id)

    def add_spending(self, category, row, value='notspecified'):
        id = row['resource_id']
        if row['sheet_id']:
            id += ' Sheet=%s' % row['sheet_id']
        if 'row_number' in row:
            id += ' Row=%s' % row['row_number']
        id += ' RowID=%s' % row['row_id']
        if value != 'notspecified':
            id += ' %r' % value
        return self.add(category, value=id)

if __name__ == '__main__':
    package_stats = StatsList()
    package_stats.add('Success', 'good1')
    package_stats.add('Success', 'good2')
    package_stats.add('Success', 'good3')
    package_stats.add('Success', 'good4')
    package_stats.add('Failure', 'bad1')
    print package_stats.report()

    print StatsList().report()