jstyle21/okcscrape3

OKCupid webscraper in Python 3

okcscrape3

(okcupid scraper in Python 3)

Notice:

I've stopped working on this since OKC changes their html structure so much and I got a job so fuggit.

What it does

Scrape and log usernames and profile text content from the popular dating website OKCupid.

Here's what it captures:

How to use it

okcscrape3 is a command line app, so install it, run it on the command line, and check out the help text like so:

To actually view and scrape OKC profiles, you will need to provide a cookies.json file, which is basically OKC user dredentials needed to look at profiles. If you have logged into OKC on Chrome before, you can find all the cookie values in the advanced settings. Here's what the file should contain:

[
  {
    "domain": ".okcupid.com",
    "name": "authlink",
    "value": "<value>"
  },
  {
    "domain": ".okcupid.com",
    "name": "nano",
    "value": "<value>"
  },
  {
    "domain": ".okcupid.com",
    "name": "override_session",
    "value": "<value>"
  },
  {
    "domain": ".okcupid.com",
    "name": "secure_check",
    "value": "<value>"
  },
  {
    "domain": ".okcupid.com",
    "name": "secure_login",
    "value": "<value>"
  },
  {
    "domain": ".okcupid.com",
    "name": "session",
    "value": "<value>"
  }
]

Detailed command overview

okcscrape3 <args>

Calling okcscrape3 using only arguments and no subroutine will allow you to set internal configuration variables such as

--base-url

okcscrape3 <args> findusers <args>

The findusers subroutine will launch a Selenium Chromedriver instance and navigate to the OKC "browse profiles" page to scrape usernames. The usernames, along with the date they were gathered and a boolean flag indicating whether the profile associated with that username has been fetched yet, will be stored in a .csv in the data folder of the package installation directory. A cookies file is not required to use this subroutine.

okcscrape3 <args> fetchusers <args>

The fetchusers subroutine will launch a Selenium Chromedriver instance, navigate to profiles using the usernames gathered by findusers, and grab the profile contents. You must provide the package with a cookies.json, because in order to access other user's profiles, the Chromedriver instance must be "logged in". The profile data is stored in a JSON as a list of dicts.

okcscrape3 <args> print-config

Print the contents of the config.ini file.

okcscrape3 <args> download-webdriver

Download the chromedriver.exe from https://chromedriver.storage.googleapis.com/2.41/chromedriver_win32.zip for Selenium to use.

How to install it

Clone the repo or download + extract it, spin up a terminal/shell (I've only tested this on Windows), navigate to the top level 'okcscrape3' folder and use pip to install with the following command:

py -m pip install .

Or just however you access pip, which would essentially be:

<prefix to get to pip> pip install .

Here's an example of what that looks like:

If the 'Scripts' folder of the Python 3 installation on which you just installed okcscrape3 is on your %PATH%, you can simply install the package and call the app by typing 'okcscrape3' in the command prompt.

I haven't comprehensively tested the installation process, so let me know if you run into issues.