Deployed Version: https://serene-dusk-25063.herokuapp.com/
Given a list of Google Doc articles, this application will do two things:
- Extract hyperlinks and text.
- Count the number of potential spelling/grammar errors.
It will then display both metrics on a results page.
The intended use case for this application is streamlining the guest post review process. Rather than having to open each document and manually check for the aforementioned information, you can simply place their Google Doc URLs in this tool and it will extract it for you in a matter of seconds.
This is useful since guest posts are quite commoditized. Most SEO's don't spend a ton of time reviewing them.
I'll also be adding more features (i.e. a plagiarism check) to streamline the guest post review process further.
I built this application using the following technologies.
Back-End:
- Rails 6.1.3.2
- Ruby 3.0.1
- PostgreSQL
- RSpec (for testing)
- Google Drive API
- Grammarbot
Front-End:
- Bulma
Here's a guide to using this application (including some of the errors you may run into if you input things incorrectly).
Visit this URL to access the application:
https://serene-dusk-25063.herokuapp.com/
Note: This app is hosted on Heroku using a free plan. If it hasn't been used in a while, the server will take some time to load. Just let it sit for a few seconds.
Next, you'll arrive at a screen that looks like this. Enter your Google Doc share URLs in the text area. Make sure to:
- Place one URL per line.
- Link only to files that are native Google Docs. If you feed the application .docx files, it will skip those files and show you an error where the results otherwise would've been (I'll show you what I mean in a second).
- Link only to files that have permissions "Anyone with this link can edit." The tool accesses Google Doc files through a service account. If your permissions are set to something more private, those files will also be skipped.
Once you've entered your Google Doc URLs, hit "Create Batch." The app will pause for a few seconds. During this time, it's working with the Google Drive API to scan your documents and pull the necessary information.
As of writing, the application extracts all the anchor texts in each article along with their hyperlinks and lists them for you (grouped by article). You can click on the anchor text and it will open in a new tab to show you what link the writer used.
The app also runs a spellcheck on the article using Grammarbot. It will tally the number of potential errors and display it in the box as well (the yellow sections in the screenshot above).
If you entered a .docx URL, that article's associated box will simply contain an error message.
Click the "View Article on Google Drive" button to see which one it is. Then, return to the Homepage and create a new batch, replacing the faulty URL you initially input with a proper native Google Doc URL.
This app uses a PostgreSQL databse containing two tables:
- Batches
- GuestPosts
When you enter your Google Doc URLs on the first screen, you're creating a Batch. The :urls column of that record will contain a single string with all of the URLs you've entered. A callback in the Batch model then calls the create_guest_posts method, which relies on a helper method to split that long string into several URLs and generate GuestPosts from them.
Each GuestPost model has a column containing the file's Google URL. If you entered an invalid URL (including one with incorrect permissions), the GuestPost record will not be created at all. This is to prevent improper files from breaking the application in subsequent steps.
The mechanisms that scan each GuestPost and extract its hyperlinks and URLs are model methods. Every single time you reload the results page, it will crawl through the document and grab the data again. This is so you can do things like edit the article and reload the results screen to see updated data rather than having to go back to the first step.