ElectionAudits

Download project files

How do I verify a download?


0.9 (boulder) release from the trunk series released

Release information
Release notes:

See the README file for the recommended procedure, to use etc.

Changelog:

2008-12-04 <email address hidden>

 * setup.py: version 0.9

 * electionaudits/models.py (Contest.select_units):
 Add "selected" field to Contest
 Use selected contests, rather than hard-coding selections in results view.
 Factor out /reports/ code to provide a list of contest_batches for
 the contest.

 * electionaudits/views.py (results):
 Add /results/ view to show main results and statistics.
 Use selected contests, rather than hard-coding selections

 (Contest.stats):
 Add stats() parameters to override confidence or s (WPM) values.

 * doc/index.html:
 Bring project home page under source control

 * electionaudits/tests.py (TestCsv.test_csv):
 Add csv test
 Add tests for /selections/ and /results/, including selection and
 random seed

 * root/templates/base.html:
 Fix html validation error: commented-out </p>

2008-11-15 <email address hidden>

 This is the state of the software as it was used to publish
 the initial reports and selections on Thursday/Friday.

 * electionaudits/templates/electionaudits/reports.html:
 Add random selection of contests

 * electionaudits/templates/electionaudits/report.html:
 Add /selections/ url as a clone of /reports/ with random input

 * electionaudits/templates/electionaudits/index.html:
 Customize home page for Boulder
 Change Home page link on base.html, comment out other links

 Update README, ToDo (as usual....)

2008-11-13 <email address hidden>

 * electionaudits/models.py (Contest.priority):
 Add random selection, threshholds and priorities to /reports/ view
 and Contest class.

 * electionaudits/views.py (report):
 Use custom view for /audit_report/<contest>/
 Add threshhold and priority (threshhold/random), and sort by priority
 Return blanks in ssr and priority if seed is not 15 chars long

 * root/templates/base.html:
 display links across one line at the top

 * electionaudits/tests.py (SimpleTest.test_reports):
 resume tests of diffs against report.html

2008-11-12 <email address hidden>

 * root/templates/base.html:
 cleaner mirroring: style.css => /media/style.css
  and add trailing slash to urls like /report/1

 * electionaudits/parsers.py (parse_xml_crystal):
 Normalize contest names and candidate names: strip extra white space
 in case different databases vary them for subtle formatting reasons

 * electionaudits/erandom.py (ssr):
 Add ssr (Sum of Square Roots) random number generator (Rivest)

 * electionaudits/models.py (Batch.ssr):
 Add NEGEXP threshhold for ContestBatch to compare to ssr (commented out)

 * electionaudits/templates/electionaudits/report.html:
 Put contest name first in <title>
 Don't sort by batch object - makes order unstable

 * electionaudits/templates/electionaudits/reports.html:
 Comment out references to PPEBWR stats

2008-11-10 <email address hidden>

 * electionaudits/util.py (AuditUnit.update):
 Allow commas in vote counts, e.g. 1,345

 * electionaudits/models.py (ContestBatch):
 Add "proportion" to Contest for scaling selection numbers
 Add random_seed, for use of verifiably seeded RNGs like Rivest's SSR
 Add "selected" to ContestBatch
 Add "notes" to ContestBatch and Batch

2008-11-05 <email address hidden>

 * electionaudits/models.py (Contest):
 Add confidence and overall_margin to Contest and use them
 in calculating and displaying statistics.

 * electionaudits/admin.py (CountyElectionAdmin):
 Allow all contest margins, confidence levels to be audited inline by
 linking them to the appropriate CountyElection and using ContestInline

 * electionaudits/util.py (AuditUnit.__init__):
 Create election and contest database records when first mentioned,
 so they are created and listed in sequence.

 * electionaudits/templates/electionaudits/report.html:
 Clarify statistics more
 Improve README

2008-11-03 <email address hidden>

 * electionaudits/tests.py (SimpleTest.test_reports):
 Add some unit tests: run via `manage.py test electionaudits`

 * setup.py:
 Generate sdist from original trunk rather than a different bzr branch,
 to avoid bzr munging of mod times.
 See https://bugs.edge.launchpad.net/bzr/+bug/245170
 Drop use of find_packages, which was bringing in unneeded files

 * README (FAQ):
 Begin an FAQ

2008-11-02 <email address hidden>

 * electionaudits/varsize.py (plotprobs):
 Fix hidden dependency: Only import pylab when necessary
 drop unused import matplotlib.numerix

 * electionaudits/views.py (stats):
 make default batch size distribution more like boulder county 2008

 * root/settings.py:
 Configure CsrfMiddleware to avoid CSRF attacks with POSTs

2008-11-01 <email address hidden>

 * README (C):
 Improve installation and use instructions in README
 Version 0.9.dev-r47

2008-10-30 <email address hidden>

 * electionaudits/views.py (parse):
 New /parse/ form to parse incoming directory with -s -c

 * electionaudits/templates/electionaudits/stats.html:
 Call Ad-Hoc Contest Statistics a "Calculator" and fix title

2008-10-30 <email address hidden>

 * README (C):
 Document new -c and directory options in README

2008-10-29 <email address hidden>

 * electionaudits/parsers.py (parse):
 if a directory arg is given to parse, parse each file in the directory
 implement --chronological option and change -c to be that, not --contest

File Description Downloads
download icon README (md5) README 94
last downloaded 6 days ago
download icon ElectionAudits-0.9.tar.gz (md5, sig) Provides a results page, improved based on Boulder audit 294
last downloaded today
Total downloads: 388

0.8 (stats) release from the trunk series released

Release information
Release notes:

Features:
=========

It helps with several facets of the task of auditing elections:

 * Imports standard election report files
 * Protects voter anonymity by combining results for very small
   precincts or
   batches into larger audit units.
 * Enables audits of central-count systems and mail-in ballots
   without requiring that the paper ballots be sorted into piles by
   precinct. Works with election systems which can't produce batch
   results, by importing full election results for each batch and
   subtracting the result for the previous batch from the current batch.
 * Facilitates risk-limiting audits by calculating relevant statistics
   based on the margin, number, size, and results of the audit units.
   Support for PPEBWR, SAFE, and NEGEXP is provided, with more on the
   d...

Changelog:

2008-10-23 <email address hidden>

 * README (Features):
 Release Version 0.8
 Highlight the features of ElectionAudits
 Update ToDo

2008-10-22 <email address hidden>

 * electionaudits/views.py (StatsForm):
 New ad-hoc /stats/ form: statistics for a given margin, confidence

 * electionaudits/varsize.py (rule_of_thumb):
 Call rule_of_thumb from paper() for consistency
 Bug: needed "s" parameter to bmin() call for SAFE

 * electionaudits/parsers.py (parse_csv):
 introduce --batchid option for distinguishing data from different files

 * electionaudits/models.py (Contest.selection_stats):
 Cache selection_stats varsize results for speed
 Include prose form of stats in audit report

2008-10-21 <email address hidden>

 * electionaudits/varsize.py (bmin):
 Handle infinities with margin of 0 votes and problems with large margins
 Eliminate scipy requirement for electionaudits: only import where needed

 * electionaudits/parsers.py (parse_csv):
 use -e option as election name for csv also

2008-10-20 <email address hidden>

 * electionaudits/parsers.py (main):
 Rename makeauditunits to electionaudits/parsers.py
 Make it easier to use, via a new manage.py "parse" command:
    electionaudits/management/commands/parse.py
 Rename audittools directory to "root"

 * electionaudits/util.py (pushAuditUnit):
 change min_ballots from 5 to 25

 * electionaudits/models.py (Contest.selection_stats):
 Make varsize statistics for Contest available via selection_stats()
 First cut - not saving results - really inefficient....

 * electionaudits/varsize.py (read_precincts):
 Make it easier to use varsize logic from electionaudits
 Accept a list (iterable) of precincts as the input source as well as a
 filename.
 Return the most interesting results from paper() in a dict.
 Add parameters (with existing defaults) to paper(): alpha and s
 Don't generate plot or print full precinct list unless called
 as main program.
 Print various quantities to 4 decimal places.
 Add shorthand for specifying precinct sizes on the command-line

2008-10-20 <email address hidden>

 * electionaudits/varsize.py:
 Don't run anything when simply loading as a module, and call main() when run as command
 Resolve main() error: "global name 'V' is not defined"
 Resolve main() error: "print_precinct_stats() takes exactly 2 argument (0 given)"
 Add default for "s" argument to negexp_probs_for_confidence

2008-10-17 <email address hidden>

 * electionaudits/varsize.py:
 add varsize.py for auditing stats - original version from Ron Rivest
 http://people.csail.mit.edu/rivest/pps/varsize.py with MIT license

 * README (Acknowledgements):
 Add acknowledgements

2008-10-15 <email address hidden>
 rename electionaudit module to electionaudits
 improve README

2008-10-15 <email address hidden>

 * makeauditunits.py (main):
 fix DJANGO_SETTINGS_MODULE bug

2008-10-15 <email address hidden>

 * makeauditunits.py (main):
 print summary results of tally at end

2008-10-14 <email address hidden>

 * makeauditunits.py (parse_xml_crystal):
 convert parse_xml_crystal etc to use AuditUnits
 Add "--election" option to name the election.
 (parse_csv):
 add EarlyVotes

 * electionaudit/models.py (Batch):
 Change election day abbreviation to "ED", not "EL" (similar to Early)
 Avoid duplicate VoteCount, Batch, ContestBatch via Meta unique_together

2008-10-14 <email address hidden>

 * makeauditunits.py (parse_xml_crystal):
 finally, create EL and EV database records, now that I can test those
 if xml file is "cumulative.xml", use parent directory name

2008-10-13 <email address hidden>

 * makeauditunits.py (main):
 define settings module without environment var DJANGO_SETTINGS_MODULE

2008-10-10 <email address hidden>

 add setup.py and ez_setup.py
 * electionaudit/templates/electionaudit/report.html:
 don't use writing-mode: tb-rl on vote totals
 * audittools/settings.py:
 comment out 'lukeplant_me_uk.django.validator' and 'django_extensions'
 Improve documentation

2008-10-09 <email address hidden>

 * electionaudit/templates/electionaudit/report.html:
 sort columns alphabetically to deal with variable order, mismatches

2008-10-08 <email address hidden>

 * audittools/templates/base.html:
 only show /validator/ if in debug mode

2008-10-08 <email address hidden>

 Version 0.7.0
 Change from GPLv3 to MIT license for wider uptake, less complication

 * makeauditunits.py (parse_csv):
 Finish & refactor code to combine ballots for privacy for csv
 remove models.AuditUnit: util.AuditUnit does what we really want
 finally remove models.Batch.cumulative - we combine when we read in
 * electionaudit/urls.py:
 fix how settings are imported to use generic method so validator works
  with ./manage.py runserver_plus --settings settings_debug
 * electionaudit/templates/electionaudit/reports.html:
 fix some bad html

2008-10-08 <email address hidden>

 * electionaudit/templates/electionaudit/index.html:
 Add introduction to home page

2008-10-07 <email address hidden>

 * electionaudit/util.py (Pipe.push):
 Write util module with AuditUnit class to combine them for privacy
 Add pushAuditUnit and class Push, ready to refactor makeauditunits

2008-10-06 <email address hidden>

 * audittools/templates/404.html:
 add 404.html and 500.html
 Describe csv in README
 Better titles in templates

2008-10-05 <email address hidden>

 * electionaudit/urls.py:
 fix settings import in urls.py

 * makeauditunits.py (parse_csv):
 name csv election after filename

 * electionaudit/templates/electionaudit/reports.html:
 add headers and audit unit counts to contest list

2008-10-05 <email address hidden>

 * Version 0.6.0
 * electionaudit/models.py (Choice):
 (Contest.tally):
 add tally funtion to calculate margins
 add contest field to Choice model
 speed up via a single transaction for parsing each file:
  from 2:53 to 0:44 for testp.xml

 * makeauditunits.py (parse_csv):
 new function to parse csv files like that from San Mateo county

2008-10-04 <email address hidden>

 * electionaudit/templates/electionaudit/index.html:
 New home page via simple generic view by default (LP: #278330)
 * electionaudit/templates/electionaudit/contest_detail.html:
 Get /contests/*/ to list the relevant batches, FWIW

2008-10-03 <email address hidden>

 * makeauditunits.py:
 fixed documentation: must set $DJANGO_SETTINGS_MODULE "(LP: #277341)

 * audittools/settings.py:
 set full path for sqlite3 db so makeauditunits works in any dir

2008-10-01 <email address hidden>

 * Version 0.5.0
 * Improve README; Add GPL v3 COPYING file, doc/model_graph.png
 * ToDo: split out from makeauditunits.py

 * electionaudit/models.py (ContestBatch):
 add contest_ballots method, and add count to report

2008-09-29 <email address hidden>

 * makeauditunits.py (do_contests):
 Version"0.4.0"
 Introduce ContestBatch record and report of AuditUnits in a Contest

2008-09-27 <email address hidden>

 * electionaudit/views.py (report):
 added report view, template,
 added custom admin to display auditunits in a table
 trimmed batch name
 add htmlvalidator
 fix template errors
 add flatpages and home page

2008-09-25 <email address hidden>

 * electionaudit/models.py (Choice.__str__):
 add django models and get them working with django generic admin

2008-09-22 <email address hidden>
 * makeauditunits.py:
 Do single-row output for --subtract
 Partly move to printf() vs print
 Move files to testdata, test again
 Add more abbreviations
 Fix replace_dict to be ordered list: replacements
 Move data info to testnotes

2008-09-22 <email address hidden>

 * makeauditunits.py:
 Add --contest option
 add --subtract to report differences for incrementals
 sort reports

2008-09-22 <email address hidden>

 * makeauditunits.py:
 Separate "Early" from "Absentee", report both

2008-09-20 <email address hidden>

 * makeauditunits.py (make_audit_unit):
 divide into early and election results, without yet using Result class
 Version 0.3.0

2008-09-14 <email address hidden>

 * makeauditunits.py (make_audit_unit):
 start profiling
 general cleaning, todo additions
 fix parties to just get one
 work on better printing

File Description Downloads
download icon ElectionAudits-0.8.tar.gz (md5, sig) adds statistics, cleaned up 90
last downloaded 3 weeks ago
Total downloads: 90

0.7-r36 (preview) release from the trunk series released

File Description Downloads
download icon README (md5) README 24
last downloaded 18 weeks ago
download icon ElectionAudits-0.7-r36.tar.gz (md5, sig) a preview release 16
last downloaded 18 weeks ago
Total downloads: 40