The Metabob team hosted its first Hackathon, BobHacks, September 25 & 26th, bringing 350 onboard assigned with the age-old task plaguing Developers: make a cool thing! I mean, it’s not quite so simplistic, but our participants did use a public application (given through Metabob’s own API), collecting information about their own/public repositories as well as results for any analyses/information about the codebase.

This allowed participants to then use that API dataset in order to build a brand new dev tool; the real goal (though cool things are always appreciated) was dedicated to helping developers mine useful insights as to what is happening within their repositories! 

With just 24 hours available, participants worked day and night discovering, collaborating, and building the right tool for the task. To aid in the ideation process, Metabob’s own Director of AI Ben Reaves and Backend Developer James Luo hosted workshops, enabling attendees to use these insights for the betterment of their tools, while project execution waged on into the wee small hours of the morning. James discussed the Metabob API, and how to access and parse it using JavaScript. Ben’s workshop was dedicated to topic modeling for python code review, and a method used by the Metabob team called tomotopy

After 24 hours of building, learning, publishing and anticipating, it was time for judging. Judges Howie Xu, Yuan Shen, and Metabob’s own Avi Gopal narrowed down the submissions to 6 finalists.

Our Judging Criteria was as follows: 

  • Originality: Does it do something entirely novel, or at least take a fresh approach to an old problem?
  • Execution: Is the hack usable in its current state? Is the user experience smooth? Does everything appear to work? Is it well designed?
  • Usefulness: Is the hack practical? Is it something people would actually use? Does it fulfill a real need people have?
  • Presentation: How well was the project presented? Did it make the hack more compelling? Did it give a good idea of its purpose?
  • Learning: Did the team stretch themselves? Did they try to learn something new? What kind of projects have they worked on before?

Amongst a product demo, question and discussion period, and insightful deliberation, 3 stellar projects were given top billing! 

  • 3rd Place was given to Metalyzer, creating a search feature that gives users all the details/statistics of a given problem, based on the repo’s name or reference ID. (You can see Metalyzer's project and team here).
  • 2nd Place was given to Kyusei, a personalization as well as pattern recognition tool for user repositories. (You can see Kyusei’s project and team here).
  • Our 1st Place winner was Automatic Issue Creation! Kushal Gopal made a command line utility tool that takes all the analysed problems on a specific repository and automatically creates issues on the project tracking software Jira! (You can see Kushal’s full project here).

It’s a tall ask to have a development tool made and delivered in just a day’s time, let alone have it be intuitive and useful to developers. Our final contestants did the unthinkable with ease, and landed the first Metabob Hackathon back down to Earth with time to spare. We want to thank all of our attendees for making the first ever BobHacks marvelous, as well as the incredible team here who worked back at mission control for making this event possible.

  • You can see the full event at Metabob’s YouTube page, as well as the project finalists here
About Metabob: 
  • Metabob is a Python static code analysis tool used to debug and visualize Python code. Using a combination of conventional status analysis, attention-based models, and a 360-degree Visualizer, Metabob is able to detect where problems are, how they interact with other aspects of your codebase, as well as offer plain-text recommendations on how to fix them.
  • Metabob’s model is written on open source code from python developers across the globe, creating a dynamic, uniquely intuitive base to help identify where problems will likely occur and debug python code.