Developer Guide

Welcome to the interactive Yetibot Developer Guide

  1. Running Yetibot locally
  2. Building your first command
  3. Command pipeline
    1. Command arguments
      1. Data
    2. Push down operations
    3. Timeouts
    4. Testing commands
  4. Working with the database
  5. Parser
  6. GraphQL API
  7. Web dashboard

These docs are are work in progress. If something is missing and you have questions just ask in #dev channl on Yetibot Slack. Feel free to open an issue to ask questions or suggest changes.

Running Yetibot locally

Building your first command

Command pipeline

Command arguments

Use raw all to view arguments.


Push down operations

Yetibot has the ability to look ahead into the command pipeline and optionally consume subsequent commands from a command handler function. In relational algebra this is often referred to as pushing down an operation.

Consider history | random as an example. The history command looks ahead at the next commands in the pipeline, sees that it's a command it knows how to consume (random), and tells the command execution pipeline to skip it for efficiency and instead handle itself.

Imagine the default naive method Yetibot would execute if this were not the case:

  1. Load all history into a data structure
  2. Pass the data structure to random which calls rand-nth on the collection. According to the public Yetibot dashboard at the time of writing we have 12,723 entries in the history table. This isn't that much but you can imagine how slow it'd be as the table continues to grow.

The full set of commands that history consumes are defined in yetibot.core.commands.history/consumeables:

(def consumables #{"head" "tail" "count" "grep" "random"})

See the history command for a full example.


Testing commands

  1. Using command-execution-info
  2. Mocking with Midje

Working with the database



Web dashboard