Skip to content

open-austin/eviction-hearing-parser

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

eviction-hearing-parser

Parse registers of actions for Travis County court hearings and calendar data.

Coverage Status

The data is scraped from Travis County's official judicial records, and is being used to create this dashboard, and to support the Eviction Solidarity Network's court tracking efforts where volunteers track court hearings to ensure that tenant protections are being enforced and tenants are connected to resources. Additionally, the data will be used to help analyze trends in evictions to target campaigns towards repeat, bad actor landlords and craft policy solutions to combat the eviction crisis.

For instructions on using the scraper, just keep reading. For instructions on contrubuting to this project, see the instructions for developers. If you have any questions or experience any problems using this, contact Matt (matt@open-austin.org) and/or Alex (apiazza@trla.org).


Command Line Tools Instructions

First, some setup:

  1. Clone this GitHub repository to your local computer.

  2. Download Chromedriver and put it in the root directory of this project.

  3. Set up a local PostgreSQL database.

  4. Install Python3 (3.8 suggested) if you don't already have it.

  5. Navigate to the eviction-hearing-parser directory in the command line using the command:
    cd path_to_project/eviction-hearing-parser.

  6. Create a virtual environment using the command:
    python3 -m venv venv
    If you'd rather not use a virtual environment, that should work too, and you can skip steps 6 and 7.

  7. Activate the virtual environment with the command:
    source venv/bin/activate

  8. Install the required libraries with:
    pip install -r requirements.txt

One more note - web scraping can be finnicky, and we've tried to anticipate and handle any errors that may occur, but you may experience errors sometimes. When you do, just re-run the same command a few times, or try re-running it without the --showbrowser option, and it should work eventually. If not, or if the errors are frequent, let us know!

Now for the fun part! We have 5 command line tools:

1) Parse Hearings

To use this command line utility, feed in a "CSV" file containing a series of case ID numbers on separate lines. The scraper will fetch the register of actions for each case from the database published by Travis County. Then it extracts information about the last scheduled hearing in the register (regardless of whether the hearing is in the past or future). It will output a JSON file collecting this information for each of the case IDs. Step-by-step instructions:

  1. Create a CSV file with a list of case IDs that you want to query (check out test_input.csv in this repo for an example of how this file should look).

  2. In the project directory and with your virtual environment activated, execute the command line utility with a command in this format:

python parse_hearings.py --infile [your input CSV file] --outfile [name of new output file] --county [name of county]

For instance, if you use the following command to scrape the three case IDs included in test_input.csv:

python parse_hearings.py --infile test_input.csv--outfile result.json --travis

A new file called result.json will appear in your project directory with scraped data from those three cases and the data for these cases will be added to your database tables (specifically the case_detail, disposition, and event tables).

If you want to see your Chrome browser in action, add the --showbrowser command. For example:

python parse_hearings.py --infile test_input.csv --outfile result.json --county travis --showbrowser

2) Parse Settings

This command line utility scrapes court calendar data from a specified date range using the Court Calendar link on Travis County's website. Only settings with category "Civil" are scraped.

For example, the command

python parse_settings.py afterdate beforedate --outfile result.json --county travis

will scrape all settings on or after afterdate and on or before beforedate (dates should be formatted like: mm-dd-yyy), output results to result.json, and add the appropriate rows to the setting table in your database. For example:

python parse_settings.py 9-1-2020 9-7-2020 --outfile result.json --county travis

Add --showbrowser to the end of the command to see the browser as it is scraping:

python parse_settings.py 9-1-2020 9-7-2020 --outfile result.json --county travis --showbrowser

3) Parse Filings

This tool does the same thing as Parse Hearings, except its input is a date range rather than a CSV with case numbers (it finds all the case numbers in the given date range and scrapes their data rather than being told the case numbers). It will also look in your database for still active cases and rescrape those. This way if a case has new data it will be updated in your database.

So, the command

python parse_filings.py 9-1-2020 9-7-2020 result.json

will scrape data for all cases that occurred on or aftr September 1, 2020 and on or before September 7, 2020. To do the same thing while showing the browser, use:

python parse_filings.py 9-1-2020 9-7-2020 result.json --showbrowser

4) Schedule

This script allows you to automatically run the scraper repeatedly on a schedule. The command

python schedule.py

will start the schedule. When the schedule is running, it will perform a "scraper run" every day at 1:00AM EST. Each scraper run performs the same exact tasks as Parse Filings and Parse Settings, except that no .json files are created and there is no --showbrowser option.

As the code is currently set up, the scraper calls Parse Filings with today's date as the beforedate and seven days ago as the afterdate, and it calls Parse Settings with seven days ago as afterdate and 90 days from now as beforedate. These parameters, as well as the time and frequency at which the scraper runs, can be adjusted with minor adjustments to the schedule.py script. For example, changing

sched.add_job(scrape_filings_and_settings_task, 'interval', days=1, start_date='2020-10-12 03:00:00', timezone='US/Eastern')

to

sched.add_job(scrape_filings_and_settings_task, 'interval', days=7, start_date='2020-10-12 03:00:00', timezone='US/Eastern')

will cause the scraper to run every 7 days rather than every day.

For as long as you don't exit the process, the schedule will continue to run locally. But if you want it to run even when your computer is off, you can do what we do - use Heroku.

5) Parse Filings and Settings Since Date

Given a date, this script will populate your database will all of the hearings and settings data on or after the given date and on or before the current date. It also divides the tasks up into weeks, so if one week fails it can just move onto the next week. Once it's done it will print to the console all the weeks for which it failed. If you've set up an email account for this project, as described here, it will also email you the weeks that failed.

For example, the command

python get_all_filings_settings_since_date.py 9-1-2020

gets all data from September 1, 2020 up until the current date.


Instructions for Contributing Developers

  1. Fork this project's repository and clone it to your local computer.

  2. Set up a local PostgreSQL database.

  3. Create your .env file.

  4. Set up a virtual environment, install requirements, and install chromedriver as described here.

  5. Write code.

  6. Make sure the scraper still works by running the tests in the tests folder and making sure the command line tools (parse_hearings.py, parse_settings.py, parse_filings.py) successfully populate the database and don't throw errors. To run the tests, just use the command pytest.

  7. Ideally, add your own tests (if it makes sense to do so).

  8. When you're done, make a pull request from your fork. If the PR completes a specific issue, include "closes #{issue_number}" in the description of your PR.


Database Set Up Instructions

If you're using this scraper to get data, follow the instructions below. We're working on making an option to use these tools without having a database, but haven't done that quite yet. If you're a developer contributing to the project, you can also follow these instructions, or you can skip them and follow the instructions here, which is quicker and easier but also slightly less ideal.

Setting up a Local PostgreSQL Instance

  1. Setup PostgresQL. Tutorials for MacOS users here and here. For Windows users here.

  2. Create a local database by entering createdb desired_database_name in the command line.

  3. Add the schema to your database using the command pg_restore -O -x -c -d database_name_from_previous_step evictions_dump.dump. The file evictions_dump.dump can be found in the "sql" folder on this GitHub. When running this command, make sure you're in the same directory as this file. You may see an error message or two, but that should be fine.

  4. If using pg_restore didn't work, you can run the commands in each of the .sql files in the "sql" folder to create all the necessary tables, indexes, views, and constraints.

pgAdmin Set Up (optional)

pgAdmin can be useful to access your database in a web browser. Here's how to get set up with that. For first time users, specifically ones adding a local server, these instructions may not be perfect, but hopefully they'll help.

  1. Install pgAmin4 on your computer.

  2. Once you have a pgAdmin window open, click on Add New Server.

  3. Give the server a name under General, then go to Connection and fill out the first 5 fields (host, port, database, user, password) according to your database credentials. If it's just a local server, which it probably is unless you're accessing the test database, you may only have to fill out some of these fields.

  4. Click Save and you should see the server on the left. Click on it, then click Databases, then scroll until you find the one that matches your database name.

  5. Click on your database name, then Schemas, then Tables, then any of the table names, and you can see all of its column names and some other info.

  6. To query the actual data, click on the Query Tool (should be on the top left, and looks like a lightning bolt). You can then query the data using SQL. For example, SELECT * FROM setting will show you all the data in the setting table.


Environment Variable Instructions

Create a file named ".env" in the root directory of this project, and add the following two lines:

LOCAL_DEV=true
LOCAL_DATABASE_URL=postgres://localhost/your_local_database_name

If you'd like to receive error emails, create a new Gmail account and configure it to allow less secure apps. Then add the following two lines to your .env file:

ERROR_EMAIL_ADDRESS=your_email_address
ERROR_EMAIL_ADDRESS_PASSWORD=your_password

Test Database Uri

If you're a developer choosing to use the test database rather than set up a local database, set LOCAL_DATABSE_URL to test_database_uri. The URI is kind of a secret and changes periodically, so email Alex at apiazza@trla.org to get it. The drawback of this method is that if multiple people are developing using the test database, any data you add for testing purposes may be removed / changed.


Using Heroku to Schedule Scraper Runs

No instructions yet... try Google for now :)

Releases

No releases published

Packages

No packages published