Welcome to flash cards for software engineers.

Running the App

Invoke the training session with:

docker run -it keytrain -p <list of pack names to train on>

# Examples:
docker run -it keytrain -p git

docker run -it keytrain -p git kubectl

A series of flashcards will start printing out, for example:

➜  keytrain git:(master) ✗ docker run -it keytrain -p git kubectl
stop minikube and remove all associated resources

Type your answer and press return/enter. If it is a multi-step answer, enter the answer for each step and press return/enter between each.

Single-Step Answer:

docker run -it keytrain -p git kubectl
stop minikube and remove all associated resources
minikube delete

Multi-step Answer:

docker run -it keytrain -p git kubectl
add all locally modified files to a remote branch?
git add .
git commit -m "message"
git push

You will see either ✅ or ❌ in response to your answer.

Development

Run the image without re-building every time with:

docker run  -v $(pwd):/usr/src/app keytrain -p git kubectl

Design

Data Model

The primary abstractions are:

• Deck - the group of Cards a user is training with currently. A Deck can have Cards from one or more Packs.
• Pack - a group of Cards stored together in a logical group
• Card - the atomic unit of a concept to be trained on, containing one or more Turns
• Turn - a Prompt with an expected series of Responses (answers)
• Prompt - the prompt given to the user to solicit a response
• Response - the expected input from the user to answer the Prompt correctly

Flashcard Data Storage Format

All Cards are represented in yaml files taking inspiration from Kubernetes objects. There are currently two definitions representing a Pack and a Card. This format makes it easy to define simple cards and is also extensible to add in metadata as the system grows to use features like labels, selectors, annotations, etc.

apiVersion: beta1
type: Pack
metadata:
    name: Github CLI Commands
spec:
  template:
    metadata:
      labels:
        pack: github
    spec:
      cards:
      - name: Github Empty Commit
        turns:
        - prompt: push an empty commit
          response:
          - git commit --allow-empty -m "message"
      - name: Commit Local Changes
        turns:
        - prompt: add all locally modified files to a remote branch?
          response:
          - git add .
          - git commit -m "message"
          - git push


---
apiVersion: beta1
type: Card
metadata:
  name: Example Single Card
spec:
  name: A Sample prompt
  turns:
  - prompt: first prompt
    response:
    - first response
  - prompt: second prompt
    response:
    - second response

Brainstorm

• Filter cards to train on (ex: only cards I missed last time, cards with expectation of getting right < 80%, cards I haven't practiced with before, etc.). Implement this both for the full deck (global CLI arg) and with per Pack option (CLI option to each Pack)
• Create IntelliJ plugin to easily create cards/packs
• Add labels and selectors to gather certain cards
• Add annotations on cards for things like lastTimeCorrect, lastTimeIncorrect, probabilityCorrect
• Auto-import content from "cheat-sheet" style data banks, i.e. https://devhints.io/bash
• Define markup language for responses (regex-ish?) to allow for varying, correct answers
  • Define escape characters
• Add ability to say assumptions for a Pack/Card
  • i.e. "(kubectl pack) Assume namespace is default unless otherwise specified."