Microsoft Bot Framework bot for viewing traffic cameras.
- Selects a city/set of traffic cameras to view. Sets this as state in the conversation.
- Examples:
view bay areaView the traffic cameras for Seattle
- Lists the names of all of the cameras for the currently-selected city.
- Examples:
list
- This is how you view a camera image. The bot keeps an index of a number of alternative names (e.g. variations on "st", "street", etc.) where possible, and uses some heuristics (including a Lucene index) to try and figure out what the user meant.
- The most reliable way to get a result, obviously, is to enter an exact camera name (case-insensitive).
- But certain more-conversational styles will often work as well, e.g. "Show me the traffic camera at Sunset Blvd".
- These are based on the phrase-search query scoring in the Lucene index, so ymmv.
- If multiple cameras may match, the user is presented with a numbered menu...
- If the user was presented with a choice list by the bot, they can enter the appropriate integer to view a given camera.
- The choice list stays in conversational state until another one needs to be presented.
If you have a new set of traffic cameras you would like the bot to support, follow these steps:
- Look at the existing examples.
- Create a new class in the Data folder, TrafficCamBot.Data namespace.
- Extend CameraDataServiceBase. (You may also just implement ICameraDataService, but you'll lose lots of nice stuff like the Lucene index by default).
- Return a unique name from the Name property. Note that this name should be displayable to the user (i.e. user-friendly and meaningful).
- In the constructor, call SetCameraNames with a list of all of the camera names your service will understand/support.
- Implement GetImageUrlForCamera (basically, given the name of one of your cameras, return the appropriate CameraImage structure).
- Send a pull request.