Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Add user interface to support Dominion conversion process #408

Closed
moldover opened this issue Feb 25, 2020 · 10 comments
Closed

Add user interface to support Dominion conversion process #408

moldover opened this issue Feb 25, 2020 · 10 comments

Comments

@moldover
Copy link
Contributor

moldover commented Feb 25, 2020

After talking with @CalebKleppner we arrived at the following:

The tabulator needs to ingest and process Dominion json data for various elections (Wyoming, Alaska, Kansas Democratic primaries, and eventually NYC local).

To allow users to input, convert (verification / auditing - #404) and tabulate, we need a way for users to specify Dominion as a data provider. For Dominion data providers, the user needs to specify an input Dominion folder location. (First Vote Column Index, etc... are not needed)

CLI support for conversion this can be run from CLI: user supplies input Dominion data folder location. Tabulator will input Dominion cvr json data from this folder.
Tabulator will export generic cvr csv to the same folder (#404) with a file per contest.

GUI support for conversion

User can click a "convert dominion cvr json to generic csv" button on the main panel. This opens a file chooser. User can select a file to convert. Exported files will be written to the same folder as the input file. Per @CalebKleppner no need to select an export folder.

Moved the Tabulation support tasks to a new ticket: #418

@tarheel tarheel changed the title Add command-line options to suppot Dominion conversion process Add command-line options to support Dominion conversion process Feb 25, 2020
@tarheel
Copy link
Contributor

tarheel commented Feb 25, 2020

Should there also be an arg for the output file path?

@moldover
Copy link
Contributor Author

moldover commented Feb 29, 2020

@CalebKleppner what do you think about Louis' question above? A couple options I see of the top of my head:

  1. automatically export all contests' cvrs into the same input folder as CvrExport-[contestId].csv
  2. specify contest to export + output path

@CalebKleppner
Copy link
Collaborator

CalebKleppner commented Feb 29, 2020

A few comments:

  1. The format from SF in Nov 2019 is different from the format going forward in SF and the Dem primaries. The version for Dem primaries is 5.10.11.24.
  2. Dem primaries only have 1 contest; SF has multiple contests (and ballot types). We should obviously program for the SF case of multiple contests.
  3. I like the idea of having the Universal Tabulator read in a CvrExport.json file and spit out the rankings in .csv format. Because it takes a lot of time to read and export 3 million records, I think it'd be better to use option don't do batch elimination if the size of the batch is one #2 above: specify contest to export and output path (which for me will probably be the same input folder but nice to have the flexibility to specify output file).
  4. Is it too much to ask for the option to convert and export a single contest, specified by [contestId] or all contests, specified by [*]?

I'm around this weekend, but I don't check GitHub posts from my phone, so feel free to call, text or email if you want a quicker response.
4. Since Dominion tends to provide all of the json files (cvrs, candidates, precincts, etc) in a single directory, I like the idea of spitting out the contest files

@HEdingfield HEdingfield self-assigned this Mar 1, 2020
@HEdingfield
Copy link
Contributor

Relevant code that was used to temporarily test the new functionality:

    // for testing	
    String wyoming = "/Flair/Downloads/Dominion/Wyoming";	
    String sf = "/Flair/Downloads/Dominion/SF";	
    List<CastVoteRecord> castVoteRecords = new ArrayList<>();	
    DominionCvrReader test = new DominionCvrReader(wyoming);	
    try {	
      test.readCastVoteRecords(castVoteRecords);	
    } catch (CvrParseException e) {	
      e.printStackTrace();	
    }

@moldover
Copy link
Contributor Author

moldover commented Mar 4, 2020

Lol! So I was looking for this sample code yesterday and the squash commit didn't have it. Sadness. Going forward I'm going to use my judgement to squash or not. Good to have in the toolbox.

Check out TabulatorSession.convertDominionCvrJsonToGenericCsv for the higher level functionality the UI / CLI code will tie into. It just takes an input file to convert. Per @CalebKleppner because Dominion data is always packaged with canonical names we can just export to the input folder. If it's easy enough to collect an export file name from the user, I can incorporate that.

@moldover moldover changed the title Add command-line options to support Dominion conversion process Add user interface to support Dominion conversion process Mar 4, 2020
@HEdingfield
Copy link
Contributor

It's right thurr, breh! Even if you squash, the PR still has a record of every commit that went into it.

@HEdingfield
Copy link
Contributor

HEdingfield commented Mar 8, 2020

Per Slack conversation, decided to create a new panel in the UI that has the CLI-only functionality (convert-to-cdf and convert-dominion-to-generic-csv). Added #421 for the former.

HEdingfield added a commit that referenced this issue Mar 8, 2020
…408) (#419)

* Adds "convert-dominion-to-generic-csv" CLI option and updates README.
@HEdingfield
Copy link
Contributor

Also confirmed with @moldover that it's fine to change the conversion function to output in an "output" directory inside the folder with the JSONs instead of in the level above them (in the GUI the user will be able to specify where to save it), and to also include the timestamp as part of the filename.

@HEdingfield
Copy link
Contributor

Relevant discussion from Slack:

@HEdingfield:
Isn't our end goal to make the Dominion conversion process kind of behind-the-scenes? So instead of creating a separate button / area in the GUI for it, wouldn't it be better to just make the normal load function handle Dominion files? Or is this too advanced / too much work?

@moldover
so you’re right that, we do want to support Dominion as a first-class data source (just as we do for ES&S)
however, we expose the conversion process as a way to verify the conversion process itself.
we’ll need to do the same for any data source which has non-trivial input logic.
so to your question, ‘wouldn’t it be better to just support it as a data source’ …
yes, that would be better.
we could read from data sources AND export the converted csv files for verification
best of both worlds

@HEdingfield
Couldn't that just include providing the converted file as another output along with the results?

@moldover
right
but there’s a few more wrinkles to consider -
Dominion data sources specify multiple contest Ids
Dominion data sources specify multiple contests (by Id)
and tabulation process is done against for a single contest at a time
(yes, one could tabulate all the contests for an election “at once”, but in actual practice, that is not what election admins do. at least according to Caleb.)
so actually getting all that figured out in a reasonable way is I think, a much bigger chunk of work, than just… plugging in a new fileReader to a new data source
but that said, I do think we should do it
so to recap: the “convert cvr” functionality is needed on it’s own right.

@tarheel
so if we wanted to support Dominion as a regular input, we would need the user to specify a contest ID param in the config also

@moldover
something like that, yes
the dem primaires only have one contest, but take SF - it has like 20 or 30 contests
Maybe it’s not that much work?

HEdingfield added a commit that referenced this issue Mar 25, 2020
* Changes main button bar into a menu bar.
* Adds `convert-to-cdf` functionality to GUI menu (fixes #421).
* Adds `convert-dominion-to-generic-csv` functionality to GUI menu (fixes #408).
* Updates README.md to reflect recent changes.
* Adds setUpAndStartService helper method.
HEdingfield added a commit that referenced this issue Mar 25, 2020
HEdingfield added a commit that referenced this issue Mar 25, 2020
* Dominion conversion output files now output to same directory as input files and include timestamp in filename (#408).

* Updates Dominion tests to account for new filename schema; Dominion test files are now cleaned up automatically; fixes bug with fileCompare not closing readers.

* All delete operations in `TabulatorTests` now handle exceptions with useful error messages if they fail.
@HEdingfield
Copy link
Contributor

Went with outputting to the input directory for now. Filed #437 for choosing output folder (and maybe also which contests are converted), and #438 for direct Dominion tabulation.

HEdingfield added a commit that referenced this issue May 24, 2020
### New features:
* Added support for converting Dominion JSON CVRs to generic .csv format (including precinct portions) (#404, #406, #407, #408, #415, #439)
* Added multiSeatBottomsUpPercentageThreshold option (#403)
* Added CLI option to convert Dominion CVR to generic .csv (#408)
* New GUI menu and conversion options (can now convert to CDF and convert Dominion to generic via the GUI) (#408, #421)
* Added Dominion Alaska CVR to sample_input folder

### Bug fixes:
* Batch elimination now works properly with singleSeatContinueUntilTwoCandidatesRemain (#396)
* In a multi-seat contest, if someone wins in the first round, we now automatically eliminate undeclared write-ins before we eliminate any other candidates; previously, we treated UWIs like a normal candidate, which meant we potentially eliminated other candidates with lower tallies first (#397)
* If UWI exceeds the winning threshold in the initial count, we no longer mistakenly elect this candidate (#398)

### Backend updates:
* Updated dependencies to latest version: JDK, JavaFX, Checkstyle google_checks.xml, Checkstyle plugin, org.openjfx.javafxplugin, org.beryx.jlink, org.apache.commons:commons-csv, org.apache.poi:poi-ooxml, com.fasterxml.jackson.core:jackson-*
* Added special code to test configs to obviate the need to update the version with each increment (#426)
* Updated tests and improved test coverage
* Copyright update (#414)
* Code cleanup
moldover pushed a commit that referenced this issue Aug 12, 2022
* Dominion conversion output files now output to same directory as input files and include timestamp in filename (#408).

* Updates Dominion tests to account for new filename schema; Dominion test files are now cleaned up automatically; fixes bug with fileCompare not closing readers.

* All delete operations in `TabulatorTests` now handle exceptions with useful error messages if they fail.
moldover pushed a commit that referenced this issue Aug 12, 2022
### New features:
* Added support for converting Dominion JSON CVRs to generic .csv format (including precinct portions) (#404, #406, #407, #408, #415, #439)
* Added multiSeatBottomsUpPercentageThreshold option (#403)
* Added CLI option to convert Dominion CVR to generic .csv (#408)
* New GUI menu and conversion options (can now convert to CDF and convert Dominion to generic via the GUI) (#408, #421)
* Added Dominion Alaska CVR to sample_input folder

### Bug fixes:
* Batch elimination now works properly with singleSeatContinueUntilTwoCandidatesRemain (#396)
* In a multi-seat contest, if someone wins in the first round, we now automatically eliminate undeclared write-ins before we eliminate any other candidates; previously, we treated UWIs like a normal candidate, which meant we potentially eliminated other candidates with lower tallies first (#397)
* If UWI exceeds the winning threshold in the initial count, we no longer mistakenly elect this candidate (#398)

### Backend updates:
* Updated dependencies to latest version: JDK, JavaFX, Checkstyle google_checks.xml, Checkstyle plugin, org.openjfx.javafxplugin, org.beryx.jlink, org.apache.commons:commons-csv, org.apache.poi:poi-ooxml, com.fasterxml.jackson.core:jackson-*
* Added special code to test configs to obviate the need to update the version with each increment (#426)
* Updated tests and improved test coverage
* Copyright update (#414)
* Code cleanup
moldover added a commit that referenced this issue Aug 15, 2022
* Changed package from com.rcv to network.brightspots.rcv.

* remove unused tests: fix for #197

* add tiebreak test #197

* Allows user to cancel during interactive tiebreaker (fixes #132); adds example config file for testing interactive tiebreaker with sequential multi-seat enabled; adds commented command in build.gradle for executing CLI arguments.

* add missing precinct test

* Reconciles branch with `develop`.

* Re-adds interactive tiebreaker sample input.

* Continues fixing recently broken tests.

* Reverts name back to "Universal RCV Tabulator".

* add new random seed config field

* updated tests

* incorporate new option into tabulation logic and add test case

* onlyOneWinnerPerRound -> allowOnlyOneWinnerPerRound

* update test case to include a tie-break

* update tests

* address comments

* tests

* add test case

* update one test for each winnerElectionMode option

* updating tests

* fix tests

* fix minor output bug and add test case for previousRoundCountsThenRandom tie-break option

* fix minor issue with transfer tallies

* update test configs

* update a ton of JSON files!

* fix bad source path

* Increment version from 0.1.0 to 0.1.1. (#373)

* Increment version from 0.1.1 to 0.1.2

* Increment version from 0.1.2 to 1.0.0 (#384)

* v1.0.0

* Increment version from 1.0.0 to 1.0.1 (#389)

* Fix broken tests (#393)

* Increments version from "1.0.1" to "1.0.1_20200320_internal" (#423)

* Adds multiSeatBottomsUpPercentageThreshold (#428)

* Tests for Dominion CSV conversion routine (#429)

* test for Dominion csv conversion routine #415

* 2 more tests for cvr conversion

* updates for Dominion conversion tests: PR #425

* fix for PR #425

Co-authored-by: jonmoldover <jonmoldover@gmail.com>

* Adds AUTOMATED_TEST_VERSION code so tests can remain stable as app version updates (fixes #426). (#435)

* Fixes Dominion conversion output filenames and path (#436)

* Dominion conversion output files now output to same directory as input files and include timestamp in filename (#408).

* Updates Dominion tests to account for new filename schema; Dominion test files are now cleaned up automatically; fixes bug with fileCompare not closing readers.

* All delete operations in `TabulatorTests` now handle exceptions with useful error messages if they fail.

* read and write precinct portion and precinct (#433)

* read and write precinct portion + precinct

* edits for precinct portion PR#433

* Update src/main/java/network/brightspots/rcv/DominionCvrReader.java

Co-Authored-By: HEdingfield <HEdingfield@users.noreply.github.com>

* Update src/main/java/network/brightspots/rcv/DominionCvrReader.java

Co-Authored-By: HEdingfield <HEdingfield@users.noreply.github.com>

Co-authored-by: HEdingfield <HEdingfield@users.noreply.github.com>

* UWI test (#440)

* add UWI test files

* add an additional test to validate my fix of #397

* use a better name for the first test

* make batch elimination work properly with singleSeatContinueUntilTwoCandidatesRemain (#441)

* make batch elimination work properly with singleSeatContinueUntilTwoCandidatesRemain

* cleaned up wording

* v1.1.0 (#454)

### New features:
* Added support for converting Dominion JSON CVRs to generic .csv format (including precinct portions) (#404, #406, #407, #408, #415, #439)
* Added multiSeatBottomsUpPercentageThreshold option (#403)
* Added CLI option to convert Dominion CVR to generic .csv (#408)
* New GUI menu and conversion options (can now convert to CDF and convert Dominion to generic via the GUI) (#408, #421)
* Added Dominion Alaska CVR to sample_input folder

### Bug fixes:
* Batch elimination now works properly with singleSeatContinueUntilTwoCandidatesRemain (#396)
* In a multi-seat contest, if someone wins in the first round, we now automatically eliminate undeclared write-ins before we eliminate any other candidates; previously, we treated UWIs like a normal candidate, which meant we potentially eliminated other candidates with lower tallies first (#397)
* If UWI exceeds the winning threshold in the initial count, we no longer mistakenly elect this candidate (#398)

### Backend updates:
* Updated dependencies to latest version: JDK, JavaFX, Checkstyle google_checks.xml, Checkstyle plugin, org.openjfx.javafxplugin, org.beryx.jlink, org.apache.commons:commons-csv, org.apache.poi:poi-ooxml, com.fasterxml.jackson.core:jackson-*
* Added special code to test configs to obviate the need to update the version with each increment (#426)
* Updated tests and improved test coverage
* Copyright update (#414)
* Code cleanup

* Hart (#458)

* add contest id to config

* add xml parsing module

* First cut at Hart CVR reader + test assets.

* cleanup paths and add expected results

* add full cvr data

* refactor contestId from Integer to String

* add school board test

* fix xml parsing

* logging

* re-add accidentally removed import

* cleanup for PR#548

* updates for PR #458

* updates for PR #458

* test other file reading options

* more cleanup for PR #458

* more cleanup for PR #458

* Fixes tests not working in Windows due to mangled UTF-8 BOM marker.

* Handles potential NPE if chosen directory contains no files.

Co-authored-by: HEdingfield <hylton@groupagree.com>

* Update dependencies (#459)

* Updates to latest version:
    * Gradle 6.5.1
    * JDK 14.0.1
    * JavaFX 14.0.1
    * Checkstyle google_checks.xml 8.34
    * Checkstyle plugin 8.34
    * org.openjfx.javafxplugin 0.0.9
    * org.beryx.jlink 2.20.0
    * com.fasterxml.jackson.core:jackson-* 2.11.1
    * org.junit.jupiter.junit-jupiter-* 5.6.2
* Gets tests working with Gradle 6.5.1 (fixes #283)
* Updates `APP_VERSION` to temporary value 1.2.0_alpha
* Addresses Checkstyle warnings
* Fixes versions in Hart tests

* make contestId a real config field (#463)

* make contestId a real config field

* update .txt

* Providers dropdown (#462)

* Changes "Provider" field for CVR source to be required, bases it on an enum of valid values, and changes it to a ChoiceBox at the start of the "CVR Files" tab in the GUI  (progress on #460 and #461).
* Adds log message to help identify which CVR source validation errors pertain to.
* Gets rid of `SimpleObjectProperty` warnings during compile (fixes #280).

* Hart GUI (#466)

* Provider ComboBox now dynamically enables and disables fields and buttons when different providers are selected. (Progress on #461... actually progress on GUI phase 2!)
* Enables `DirectoryChooser` for Hart so it now functions through the GUI. (Fixes #460)
* Validation improved to fail if `firstVoteColumnIndex`, `firstVoteRowIndex`, `idColumnIndex`, `precinctColumnIndex` are defined for CVRs from non-ES&S providers.
* Removes superfluous logging.

* direct Dominion tabulation (#470)

* direct Dominion tabulation

* enable specifying a Dominion CVR source path in the UI

* output generic CSV and add test coverage

* remove extra log line

* filter out CVRs from other contests during the loading

* Changes `contestId` so it's specified on a per-CVR basis (fixes #456) (#474)

* First cut at Clear Ballot rcv reader (#475)

* First cut at Clear Ballot rcv reader
Add Kansas Dem Primary regression test
Add Clear Ballot reference files
#400

* updates for PR #475

* updates for PR #475

* updates for PR #475

* Enable Clear Ballot tabulation in GUI (progress on #400) (#480)

* GUI redesign phase 1 (#492)

* Splits Output tab into new Contest Info and Output tabs.
* Redesigns GUI CVR Files tab, adds Clear button, and changes Add button so it only clears the file path to make it easier to manually enter multiple sources that share fields.
* Improves visual presentation of Candidate tab; adds Clear button and adds checkBoxCandidateExcluded when adding a candidate.
* Reorganizes presentation of rules in "Winning Rules" and "Voter Error Rules" tabs.
* Winner Election Mode and Tiebreak Mode now start undefined with all relevant fields disabled; choosing specific modes enables applicable fields.
* Changes Winner Election Modes and Tiebreak Modes to be more user-friendly, including necessary migration logic to update older config files.
* Adds `continueUntilTwoCandidatesRemain` boolean (fixes #481).
* Expands footprint of GUI window to 1200x1000.
* Implements bordered boxes.

* remove Dominion conversion feature (#487)

* update test files

* remove conversion feature from code

* update test configs

* remove unnecessary check

* allow overvote delimiter for ES&S source files (#503)

* test

* update winner election mode logic

* comments

* final comments

* Read and Tabulate Unisyn Xml Cdf (#504)

* plumbing for xml cdf reading

* add contestId to json and xml CDF parse logic
update unisyn cvr test to use the full (but still incorrect) cvr data.  This test is expected to fail.
add candidate parsing code to xml CDF parser

* dont parse Candidate data from CDF at runtime - this enforces the same constraints on validation as all other file types.  The un-used candidate parsing code is left in for implementing future helper functions which will automatically populate config files.

* Handle ContestSelections in XML correctly

* add NIST example 2 (seems to be correctly formed)

* fill out CDF XML class definitions needed for tabulation

* add some helpers

* cleanup pre-processing code

* add more class fields around contest selections
fix ranking parsing
add example Unisyn contests

* handle write-ins

* get first 2 Unisyn regression tests working :)

* fix first NIST CDF XML test

* cleanup test names

* parse GpUnit from CVRs

* remove un-used assets

* updates for PR #504

* updated data from Unisyn with GpUnit (precinct) parsing validated

* add more Unisyn regression tests
tests share a single cvr input file

* tabulate all elections - in practice we should not see more than one

* updates for PR #504

* updates for PR #504

* Sundry UX fixes (#509)

* Converts `overvoteRule` from a `ChoiceBox` to an array of `RadioButton`s; changes `overvoteRule` string display in config files and adds migration logic.
* Disables `decimalPlacesForVoteArithmetic` and `nonIntegerWinningThreshold` except when `winnerElectionMode` is "Multi-winner allow only one winner per round" or "Multi-winner allow multiple winners per round" (fixes #500).
* Adds suggested values for `overvoteLabel`, `undervoteLabel`, and ES&S column and row indices as requested by @chughes297 (overriding what was in the original GUI mock-ups).
* Clarifies language to address confusing UX in `passesBasicCvrSourceValidation()` for CVR-specific errors.

* Cdf Json Fixes (#506)

* plumbing for xml cdf reading

* add contestId to json and xml CDF parse logic
update unisyn cvr test to use the full (but still incorrect) cvr data.  This test is expected to fail.
add candidate parsing code to xml CDF parser

* dont parse Candidate data from CDF at runtime - this enforces the same constraints on validation as all other file types.  The un-used candidate parsing code is left in for implementing future helper functions which will automatically populate config files.

* Handle ContestSelections in XML correctly

* add NIST example 2 (seems to be correctly formed)

* fill out CDF XML class definitions needed for tabulation

* add some helpers

* cleanup pre-processing code

* add more class fields around contest selections
fix ranking parsing
add example Unisyn contests

* handle write-ins

* get first 2 Unisyn regression tests working :)

* fix first NIST CDF XML test

* cleanup test names

* parse GpUnit from CVRs

* remove un-used assets

* updates for PR #504

* updated data from Unisyn with GpUnit (precinct) parsing validated

* add more Unisyn regression tests
tests share a single cvr input file

* tabulate all elections - in practice we should not see more than one

* updates for PR #504

* better logging

* first cut at fixing JSON CDF reader:
- all static election objects are parsed at beginning of cvr processing
- during processing contest selections are linked to candidate objects
- candidates are stored in config - not pulled from CDF file
- fixed NIST test set 0 config and CDF json to match the NIST 1500-103 UML
- verified tabulation results are correct for Nist test set 0 test

* update ResultsWriter CVR generation code to create Candidate objects and link to them from ContestSelection objects.
verified updated NIST test 0 cvr output

* updates for PR #504

* fix typo

* update test asset with contest name

* better handling for cdf reader parse errors

* update all CDF json assets

* Updates to CDF parsing logic:
Candidate validation and CandidateId validation are done while parsing CVRs (not during static parsing)
xml parser throws unrecognizedCandidatesException (instead of just CVRParserException)
Add json getCvrContest helper
json parser logic upgrades:
 handle writeIns in both places
 check for rank in both places
 check for HasIndication
 check for IsAllocable
 synthesize cvrId

* cleanup for PR #506

* updates for PR #505

* Throw if undeclared write-in is found but label has not been defined.

* make some fields per-source instead of per-config (#517)

* create new class

* helpful errors in CLI; don't allow newer version in config than in app

* make fields per-source

* comments

* Jon's changes

* fix almost all tests

* fix Dominion tests

* update validation and documentation for 4 fields that moved to CVR level (#518)

* update validation and documentation for 4 fields that moved to CVR level

* Fixes broken tests and a couple of typos.

Co-authored-by: HEdingfield <hylton@groupagree.com>

* camelCase strings for enums in config files (#519)

* use camelCase for enum values in configs and update migration logic accordingly
* disable editing of CVR source rows in GUI table

* Final 1.2.0 polish (#526)

* Updates checkstyle from 8.34 to 8.36.2, and google_checks.xml.
* Creates and implements separate `checkstyle-suppressions.xml` file to suppress Checkstyle warnings that don't make sense for our project (fixes #489).
* Addresses relevant Checkstyle warnings (fixes #490).
* Addresses IntelliJ warnings.
* Pixel-pushing for Linux (get thicc, boi)
* Changes all logging to use new `fine()`, `info()`, `warning()`, and `severe()` methods.
* Gets rid of now-unused `onEditCommit` functions for the CVR Files and Candidates tables.
* Makes util classes `final`, with private constructors to prevent instantiation.
* Renames `TieBreakMode`, `tieBreakMode`, `TieBreak`, and `tieBreak` to `TiebreakMode`, `tiebreakMode`, `Tiebreak`, and `tiebreak`, respectively. This word has now lost all meaning.
* Removes unnecessary usages of `.toString()`.
* Standardizes exceptions as `exception` instead of `e` to comply with VVSG requirement on variable names (5.2.5 paragraph c) that only index loops are allowed to have single-character variable names.
* Fixes broken test.
* Minor fixes to hints.
* Get rid of the Devilish Double Spaces after periods.

* Minor cleanups (#527)

* fix README

* CDF JSON: throw if a CandidateObject can't be found during parsing.

* Cleanup CDF class (type) names as output from ResultsWriter.

* fix label

* Handle missing precinct data for older Dominion data sets #533 (#534)

* Handle missing precinct data for older Dominion data sets #533
* Add test data with missing precinct data for older Dominion data sets #533

* Fix for #536 - allow multiple CDF files (#537)

* Fix for #536 - allow multiple CDF files
Add regression test with multiple CDF files

* fix overvote delimiter logic and update test (#539)

* Fixes for bad merge between `master` and `develop`.

* Add README and LICENSE for split repo

* latest updates from develop

* update cdf output names for #584

* add dominion multi file tests

* add licence file #551

* remove test_data as submodule #612

* remove test_data submodule #612

* remove submodule reference from docs #612

Co-authored-by: HEdingfield <hylton+git@groupagree.com>
Co-authored-by: hedingfield <hylton@groupagree.com>
Co-authored-by: Louis Eisenberg <tarheel@gmail.com>
Co-authored-by: HEdingfield <HEdingfield@users.noreply.github.com>
Co-authored-by: Armin Samii <artoonie@gmail.com>
Co-authored-by: Armin Samii <armin.samii@gmail.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

No branches or pull requests

4 participants