deardurham / dear-petition Goto Github PK

I saw a comment in our Django entrypoint script that states "If only .env files supported variable expansion..." I think this might be a use case for Dockerize? https://github.com/jwilder/dockerize

@copelco Do you think there is value in adding Dockerize to our stack at this point, or have you more or less already emulated this functionality?

Currently the file number from the most recent record is displayed on the petition. If there are multiple records represented on a petition, the petition file number field should say:

Multiple, see below

This should be done in the mapper module.

Add name and date to appropriate fields. Check the "Petitioner's Attorney" checkbox.

Check out petition #105 on production and figure out why it is producing a 500 error.

Currently we use Mailgun but it seems to be hard and/or impossible to configure the project to send mail using Mailgun for free. Let's try SendGrid. This will allow us to send users emails, for example for creating their accounts.

The current Pipfile breaks the Heroku deployment. It has been renamed to Pipfile.bak until it's resolved.

I'm open to the idea of moving to using Pipfile if we can:

• Remove existing requirements files so we don't have two sets of requirements
• Update documentation on how to use Pipfiles
• Update Dockerfiles to use it
• Ensure Heroku deployments still work

sudo apt-get install -qq libtiff4-dev libjpeg8-dev libfreetype6-dev liblcms1-dev libwebp-dev
E: Package 'libtiff4-dev' has no installation candidate
E: Package 'liblcms1-dev' has no installation candidate
travis_time:end:135462cd:start=1565636043033281299,finish=1565636043537010181,duration=503728882
The command "sudo apt-get install -qq libtiff4-dev libjpeg8-dev libfreetype6-dev liblcms1-dev libwebp-dev" failed and exited with 100 during .�[0m

Your build has been stopped.

https://travis-ci.org/deardurham/dear-petition/builds/570985937

Per #139, the API needs to accept these new fields and add them to the petition.

1. The GeneratePetitionSerializer needs to be updated for this.
2. The PDF data fields are below and should be added to the map_petitioner method:
   • NamePetitioner
   • StreetAddr
   • MailAddr
   • City
   • State
   • ZipCode

Add additional comments field to petition generation form.

Only CHARGED offenses should appear on the petition

Right now, technically anyone can see records used to generate petitions in the tool. This was a concern that was raised during the meeting with Gina. Either we need to partition user files/cases so each user can only see their own or only admins can see the records. We will need to discuss the approach we want to take next meeting.

Ideally, we could test a PR before it's even merged. Currently, too many environment variables are required in app.json. I'm not sure how to get around this yet.

CI only runs tests for the users app, since that's all that exists currently. Adding a few tests, even to just verify logins are required, would be ideal.

Will need to discuss what this looks like next meeting.

Add option to allow PDFs to be opened in the browser, rather than as an attachment.

We have been testing in Chrome, but Joseph uses Firefox. Some text appears too large in Firefox (dates in offense section do not fit in column). Also, saving as PDF in Firefox blanks the petition in Adobe Acrobat. Investigate these issues and determine if there is a fix.

When developing locally on the frontend and proxying to https://dear-petition-staging.herokuapp.com/, POST requests fail:

2020-06-11T16:41:58.400638+00:00 heroku[router]: at=info method=POST path="/petition/api/token/" host=dear-petition-staging.herokuapp.com request_id=2e5d7840-7c25-4894-a00d-d7012d9c9520 fwd="127.0.0.1,71.70.229.1" dyno=web.1 connect=0ms service=67ms status=200 bytes=1031 protocol=https
2020-06-11T16:42:03.967550+00:00 heroku[router]: at=info method=POST path="/petition/api/batch/" host=dear-petition-staging.herokuapp.com request_id=afe4aa2b-5789-4d64-b828-d3233a34754a fwd="127.0.0.1,71.70.229.1" dyno=web.1 connect=1ms service=10ms status=403 bytes=475 protocol=https

Performing the same actions directly on staging work successfully.

Check out record #103 on production and figure out why it is missing the disposition method.

This field is not present in traffic accidents

When parsed properly, record.data["Offense Record"]["Records"] will contain JSON like:

[
    {
        "Action": "CHARGED",
        "Description": "SPEEDING(80 mph in a 65 mph zone)",
        "Severity": "INFRACTION",
        "Law": "G.S. 20-141(B)",
        "Code": "4450"
    },
    {
        "Action": "ARRAIGNED",
        "Description": "SPEEDING(80 mph in a 65 mph zone)",
        "Severity": "INFRACTION",
        "Law": "G.S. 20-141(B)",
        "Code": "4450"
    },
    {
        "Action": "CONVICTED",
        "Description": "IMPROPER EQUIP - SPEEDOMETER",
        "Severity": "INFRACTION",
        "Law": "G.S. 20-123.2",
        "Code": "4418"
    }
]

Ideally, these would be displayed on the CIPRS tab in a table, likely somewhere near here in the template.

The password reset views were not migrated to the React SPA. We can either:

1. Restore the existing Django views
2. Create API endpoint(s) to handle the reset and implement FE in React

This was mentioned in the meeting with Gina, but we were not provided an example. Let's discuss next meeting.

Petitions must be specific to district or superior courts. If a batch contains both CR and CRS records, then they must be entered into separate petition forms with the appropriate box checked.

Change county name here to caps.

In export app, offense dispositions need to be mapped to the following codes within the petition PDF:

• Dismissal without Leave by DA (VD)
• Dismissed by Court (DC)
• Deferred Prosecution Dismissal (DPD)
• Discharge and Dismissal (DD)
• Conditional Discharge (DD)
• No Probable Cause (NPC)
• Never To Be Served (NTBS)
• Guilty to Lesser (Glty to Lesser)

On the petition generation page, the agency autocomplete allows selection of the same agency numerous times. Ideally the choices list would exclude the currently selected agencies.

Right now the only attorneys we have as options in the tool are Arnetta and Henry, who are no longer with DEAR. Joseph is to send us a list of attorneys that we should have as options in the tool.

Each row in the Petition PDF's offense section is keyed off of an incrementing index with these field names:

• File No.(s): Fileno:1
• Date of Arrest: ArrestDate:1
• Offense Description: Description:1
• Date of Offense: DOOF:1
• Disposition: Disposition:1
• Date of Disposition: DispositionDate:1

For each record.data["Offense Record"]["Records"], map:

• File No: record.file_no -> Fileno:N
• Date of Arrest: record.offense_date -> ArrestDate:N
• Offense Description: record.data["Offense Record"]["Records"][N]["Description"] -> Description:N
• Date of Offense: record.offense_date -> DOOF:N
• Disposition: record.disposition_method -> Disposition:N
• Date of Disposition: record.disposed_on -> DispositionDate:N

DEAR team wants to be able to fill in client information for petition (Address, SSN, driver's license #, etc.). This information does not come from CIPRS, so needs to be filled in manually.

Not Guilty fields differ from Dismissed Charges:

• The application will need to reference the “Verdict” field for the phrase “Not Guilty”.
• The entry in the petition’s “Disposition” column should read “Not Guilty” spelled out.

AC:

• dear_petition.petition.types.not_guilty module exists, modeled after dear_petition.petition.types.dismissed
• not_guilty.get_offense_records() returns an OffenseRecord queryset where:
  • Q(action="CHARGED")
  • and Q(offense__verdict__iexact="Not Guilty")
• dear_petition.petition.types.main.TYPE_MAP includes mapping from constants. NOT_GUILTY to get_offense_records() above
• dear_petition.petition.etl.load.create_batch_petitions() includes a section to create distinct Petition objects

It is possible for a CIPRS record to have offense split into multiple tables on the record, and the parsers currently do not handle that well.

Adobe Acrobat can't see auto-filled text in the petition PDFs. This post implies it may be harder than expected. This issue was opened in pdfrw, but it's unresolved.

This may not be needed if #12 is a good enough workaround.

The React SPA is very simple to get running after cloning the repo:

npm i
npm run start

Ideally a frontend developer could point the backend API based on what was being developed:

1. To staging (maybe the default?)
2. To the Django docker container

This can be customized in package.json under the proxy setting:

{
  "proxy": "http://localhost:8000/"
}

AC:

• Determine best way to override proxy
• README is updated with how to configure

I tried to go through the non-Docker local setup docs today, and was surprised that they stop after installing requirments, without instructions about setting up a database, running the server, etc. locally. If the project supports non-Docker setups, then we should either

• a) give instructions on how to set up the rest of the project locally without Docker, or
• b) remove the non-Docker section

Also set up and document pre-commit hook for black

Up to date petitioner addresses are not available on CIPRS records. Fillable address fields should be added to the generate page so that they can be included on all petitions:

• Name (pre-filled with batch["label"])
• Address line 1
• Address line 2
• City
• State
• Zipcode

They should be submitted alongside the other JSON to the generate API endpoint, e.g:

{
    "petition": 11,
    "ssn": "111-111-1111",
    "drivers_license": "12345678",
    "name": "",
    "address_1": "",
    "address_2": "",
    "city": "",
    "state": "",
    "zipcode": "",
}

/petition/api/users/ will always return a single item list based on the logged in user. If the user is is_staff, it will contain a URL path to the admin. If not, it will be blank.

AC:

• Add admin URL to navbar if user api returns it

These are findings of Guilty. There are two big differences from the Dismissed Charge fields.

• First, the application will pull the entry for the petition’s “Offense Description” column from the “Convicted” row of the CIPRS record rather than from the “Charged” row.
• Second, the application will need to reference the “Verdict” field for the phrase “Guilty”.

AC:

• dear_petition.petition.types.misdemeanor module exists, modeled after dear_petition.petition.types.dismissed
• not_guilty.get_offense_records() returns an OffenseRecord queryset where:
  • Q(action="CONVICTED")
  • and Q(offense__verdict__iexact="Guilty")
• dear_petition.petition.types.main.TYPE_MAP includes mapping from constants. MISDEMEANOR to get_offense_records() above
• dear_petition.petition.etl.load.create_batch_petitions() includes a section to create distinct AOC-CR-281 Petition objects

If there's only a single offense date, the tool should calculate the age and fill in the Age PDF field. However, if there are multiple offense dates represented on a petition, just fill the field with Varies.

Users should be able to manually override the court record parsed fields.

510 South Dillard Street, Durham, NC 27701

This is a trickier but common situation where the defendant is charged with a higher-classification offense (typically a felony) but pleads guilty to a lesser-included offense (such as a misdemeanor).

• The “Charged” offense should be listed on the petition while the “Convicted” offense should not.
• The “Guilty to Lesser” phrase is found in the “Plea” field.
• The “Disposition Method” field lists “Disposed by Judge”.
  • Disposition Method = Disposed by Judge
  • Plea = Guilty to Lesser

Notes: DJ listed the phrase that is found on the CIPRS record. These are generally located in the “Disposition Method” field of the CIPRS record. In parenthesis, I listed the corresponding code that should be displayed in the petition’s “Disposition” column.

This form is used as an addendum to any expunction petition when there are additional LEO agencies and/or offenses that need to be listed and don’t fit on the petition.