243 views
--- tags: glamhack title: Sir Dridbot Glamhacker --- ## Challenge For the first time at a #GLAMhack, this time a virtual edition, we're making heavy use of real time collaboration platforms like Slack. Chatbots are a kind of app that interacts with users through chat. **Sir Dridbot Glamhack** is a deployment of the [Dridbot](https://github.com/hackathons-ftw/dridbot) open source chatbot based on the [Hubot](https://github.com/slackapi/hubot-slack) framework, developed by us originally for use in the [Climathon](https://blog.datalets.ch/023/#hellosodabot). For example, you can ask it to search or retrieve open data sets. See the [introductory video](https://vimeo.com/414580564), use the chatbot, suggest ways to improve it, contribute to the source code - or just fork it and create your own GLAM bot! ### Demo To try out our bot without spamming your Slack channel, start a Direct Message (Ctrl-Shift-K) with **Sir Dridbot Glamhacker**. Type `help` to see a list of commands you can use. If you're feeling brave, invite Dridbot to your channel and type something like `glambot fresh data` into your chat to get a response. Join the [#project_chatbots](https://app.slack.com/client/T01389WTSSX/C014FPW1DJT) channel on Slack if you have any questions/ feedback/ complaints. The source code is in the [dridbot/glamhack2020](https://github.com/hackathons-ftw/dridbot/tree/glamhack2020) branch. Find our VR-avatars in [Lichtspiel Glamhack](https://hubs.mozilla.com/3UZuSXs/lichtspiel-glamhack) ([Spoke](https://hubs.mozilla.com/spoke/projects/RmnCFj7), [Source](http://all.utou.ch/is/stumm/lichtspiel-vr.zip)) and [Silky Evil Huddle](https://hubs.mozilla.com/xnZpsxG/silky-evil-huddle/) on Mozilla Hubs. ### Team - Jeremy ([@jrmysvr](https://github.com/jrmysvr)) - Oleg ([@loleg](https://github.com/loleg)) [![](https://avatars.slack-edge.com/2020-06-05/1165158430805_b0137d248088d8460f71_96.png)](https://app.slack.com/client/T01389WTSSX/C014FPW1DJT) # Results (1) **Film recommendations** A new chatbot has been written to query films from the [Lichtspiel archives](https://lichtspiel.ch/en/database/) catalog, using an [open data package](https://github.com/schoolofdata-ch/lichtspiel-films), posting a result to the channel on a chosen category. The development is happening in the `glamhack2020` branch and you can leave comments on the [Pull Request](https://github.com/hackathons-ftw/dridbot/pull/5). ![](https://us-east-1.linodeobjects.com/dribdat/uploads/upload_74abc7c42370b4ad931b62bea3b3f40a.jpg) (2) **New open data sources** In addition to [opendata.swiss](https://opendata.swiss), where the Swiss [OpenGLAM datasets](https://opendata.swiss/en/organization/openglam) are found, we added the more _GLAM-orous_ data portal of the UK's [Natural History Museum](https://data.nhm.ac.uk/) to Dridbot's search results. ![](https://us-east-1.linodeobjects.com/dribdat/uploads/upload_6747c13a7b03c8faab19eaa7cb88fa41.png) ![](https://us-east-1.linodeobjects.com/dribdat/uploads/upload_15b9798d8a177b3f752c4e7b2573cba1.jpg) (3) **New frontiers for chatbots** [Hubot](https://hubot.github.com/) supports a variety of chat platforms besides Slack, and we thought it would be neat to adapt it to the next generation of chat: _immersive virtual reality_. In a third hack that we're confident the [Lord of Hackathons](https://www.blick.ch/digital/mark-zuckerberg-als-avatar-facebook-mit-virtuellem-chat-und-neuer-vr-brille-id5587727.html) will approve, we took a stab at adding film recommendations to [Mozilla Hubs](https://hubs.mozilla.com/#/) using an adaptation of [hubs-discord-bot](https://github.com/MozillaReality/hubs-discord-bot). Once we write a custom [hubot adapter](https://hubot.github.com/docs/adapters/), you'll be able to chat with the Glamhacker bot inside of [VR spaces](https://hubs.mozilla.com/xnZpsxG/silky-evil-huddle/), such as this one quickly thrown together for the demo using [Spoke](https://hubs.mozilla.com/spoke/). ![](https://us-east-1.linodeobjects.com/dribdat/uploads/upload_f5d4b1727d563897dd4debd6a182f103.jpg) Special thanks to **Fabien Benetou** ([@utopiah](https://twitter.com/utopiah)) for his helpful script and guidance, and sharing an incredible look into the future of XR experience and VR for digital culture, [as he explains here](https://www.youtube.com/watch?v=Evhh69VBG2s) with a Wikipedia in VR. [Watch this](https://twitter.com/utopiah/status/1268485058245152772) and be blown away. ![](https://us-east-1.linodeobjects.com/dribdat/uploads/upload_9847e3674383b1ed1fd0208c2ce980dd.jpg) Inspired by the Lichtspiel, we created a 3D rendition of the iconic film theatre in Bern: [![](https://us-east-1.linodeobjects.com/dribdat/uploads/upload_ebcba8d25048c935a9f29a17159ddbde.jpg)](https://hubs.mozilla.com/3UZuSXs/lichtspiel-glamhack) Our avatars are in the [Lichtspiel Glamhack](https://hubs.mozilla.com/3UZuSXs/lichtspiel-glamhack) VR chatroom on Mozilla Hubs, and you can visit [Spoke](https://hubs.mozilla.com/spoke/projects/RmnCFj7) or download the [Sources](http://all.utou.ch/is/stumm/lichtspiel-vr.zip). During the event we had a very simple room called [Silky Evil Huddle](https://hubs.mozilla.com/xnZpsxG/silky-evil-huddle/). ### Screenshots A couple of impressions of the work in progress ... ![](https://us-east-1.linodeobjects.com/dribdat/uploads/upload_c59e936a1e85e1080423fad46f5bb6f9.png) ![](https://us-east-1.linodeobjects.com/dribdat/uploads/upload_d1f59c919964721abde7e4defc62b67f.png) ![](https://us-east-1.linodeobjects.com/dribdat/uploads/upload_11948c4d87472ac05e25f786110f2c24.png) ![](https://us-east-1.linodeobjects.com/dribdat/uploads/upload_c3fc8630358b447748a4fd6b01945542.jpg) Thanks to everyone at [#GLAMhack](https://hack.glam.opendata.ch/) for another brilliant event, and booyah Coronavirus! 🦠 --- ### Brainstorming 6.6.2020 - [x] ask the user a set of questions to look up in the data - [x] prefetch the whole dataset & store it in memory - [x] query the data more intelligently, less [randomly](https://twitter.com/loleg/status/1268974606725140489/photo/1) - [ ] see if there's any way to get images from their service - [ ] ask Lichtspiel to update the dataset - [ ] ...??? - [ ] [avoid A.I. wrath](https://us-east-1.linodeobjects.com/dribdat/uploads/upload_872854965294fed5e1ebf032e9a511a2.jpg) - [ ] [be kind to bots](https://twitter.com/loleg/status/1269255240903262208) - [ ] profit!!! ### Brainstorming 5.6.2020 How about asking a user 3 or 4 questions, and making a film recommendation? I'm itching to use the [Lichtspiel](https://github.com/schoolofdata-ch/lichtspiel-films) open dataset from a couple of years ago (we might get it updated while we're at it). In my [livestream](https://youtu.be/RQu-ojs751Y?t=1327) (22:30) I did a quick run through this data. [This image](https://catholic-link.org/wp-content/uploads/2013/10/movie.png) and this [online quiz](https://twitter.com/sodacpr/status/1268982829372801025) might inspire thinking about what questions one might ask a chatbot to get a film recommendation.