Augmenting HasJob With Docspad

About two months ago, Niketh came up with an idea to augment the job application process on Hasjob. Having made the quite awesome product, Docspad for handling documents in web applications, he certainly noticed the lack of the option to let a job applicant upload a CV while applying through Hasjob. While the user could add details about himself/herself in a text field, nothing tells as much about an applicant as a personalized CV. As the code for the website is open sourced and written in Python, I took this opportunity to check out the Docspad APIs and spend a weekend contributing to a Python based open source project!

But before I dive into the task and the experience, let’s first discuss more about Hasjob and Docspad.

Hasjob is HasGeek’s job board. Hasgeek, based out of Bangalore, is known for organizing technology events for geeks and developers in India. Some of these events include Hackathons and Hacknights, which are loved among the geeks. HasGeek lives up to its name by providing the source code to their website and its components under open source licenses on Github.

With the goal of making digital content more accessible, Niketh and Aditya founded Bookpad (@bookpadinc). Docspad is their first product, which provides an easy solution for developers to make applications with document handling capabilities. The set of APIs make it easy for a developer to provide users with the option to upload and view documents through the web browser on any device ranging from mobile to desktop.

Now, the primary task was to provide a job applicant with the option to upload a document as his CV and allow the employer to view the same. Moreover, we wanted the integration to make as little changes to the current workflow and code architecture. The reason behind this was two-fold :

  • allow seamless transition to the new workflow.

  • avoid making changes to the database layer or code architecture to make it easy to integrate and make it live at the earliest.

As a result, we settled on providing the user the option to upload a PDF, DOC or a DOCX file while filling up the rest of the application. To avoid making changes to the database layer, this document is only temporarily stored on the HasJob server until it is uploaded to Docspad. Thereafter, a URL based on the docid returned by Docspad Upload API is simply attached to the apply_message field, which is the one that the employer receives in an email allowing the employer to view the document with or without downloading it to his device. The Docspad HTML5 based viewer can be embeded right into the application, allowing us to let the user view the uploaded content from within Hasjob’s website.

This process is quite simple to code, made simpler by the quite flexible Flask microframework used for the Hasjob site along with the easy Docspad API.

Some highlights of the project were :-

  1. Wrapping around Docspad API

    Since Docspad didn’t have any publicly available SDKs at the time, I began by building a wrapper around the APIs. This code later became the boilerplate for a Python SDK for Docspad, which will soon be stable. SDKs for several other languages and platforms are in the pipeline too. SDKs not only make it easier to build applications on, but also provide sample code for using the APIs for those who like to build their own tools.

  2. Lastuser and HasGeek Auth

    HasGeek uses HasGeek Auth for authorization and user management for its sites. It is powered by Lastuser, which has been developed by HasGeek themselves! At that time, I was still getting to know more about HasGeek’s open sourced components and was quite impressed with their dedication to sharing all these tools that they had developed.

In the end, it turned out to be a great weekend project and let me get acquainted with Flask and Docspad API. Now, every time I encounter a site that doesn’t provide good support for uploading/viewing documents, I wonder all the things they could do with the API! In fact, there are several other job portals that could make use of the API in quite the same way.

While I didn’t follow up much after sending them the pull request, Niketh has been in touch with them and informs me that they are quite pleased with the implementation. Eagerly awaiting for the integration to go live!

Do leave your suggestions, about the blog or the project, in the comments below.

Comments