summaryrefslogtreecommitdiffstats
path: root/flashcards-project
diff options
context:
space:
mode:
authorEddy Pedroni <epedroni@pm.me>2024-09-26 10:02:15 +0200
committerEddy Pedroni <epedroni@pm.me>2024-09-26 10:02:15 +0200
commite65bef9c22244fc9bcd22a37d335f5f76ba16ff5 (patch)
tree9af6fa41bfee6fc03c3ab30cf1b23a82bdf8f2e7 /flashcards-project
parentce76b00d7b2ccac6843732f92becfabb753864a0 (diff)
Create separate packages for library and CLI
Diffstat (limited to 'flashcards-project')
-rw-r--r--flashcards-project/pyproject.toml14
-rw-r--r--flashcards-project/src/flashcards/__init__.py2
-rw-r--r--flashcards-project/src/flashcards/card.py12
-rw-r--r--flashcards-project/src/flashcards/parser.py108
-rw-r--r--flashcards-project/src/flashcards/scheduler.py62
-rw-r--r--flashcards-project/src/flashcards/scheduler_brutal.py79
-rw-r--r--flashcards-project/src/flashcards/session.py51
-rw-r--r--flashcards-project/src/flashcards/state_json.py23
8 files changed, 351 insertions, 0 deletions
diff --git a/flashcards-project/pyproject.toml b/flashcards-project/pyproject.toml
new file mode 100644
index 0000000..06a6931
--- /dev/null
+++ b/flashcards-project/pyproject.toml
@@ -0,0 +1,14 @@
+[build-system]
+requires = ["setuptools"]
+build-backend = "setuptools.build_meta"
+
+[project]
+name = "flashcards"
+authors = [
+ {name = "Eddy Pedroni", email = "epedroni@pm.me"},
+]
+description = "A library for memorising information with flashcards"
+requires-python = ">=3.12"
+dependencies = [
+]
+dynamic = ["version"]
diff --git a/flashcards-project/src/flashcards/__init__.py b/flashcards-project/src/flashcards/__init__.py
new file mode 100644
index 0000000..38c9936
--- /dev/null
+++ b/flashcards-project/src/flashcards/__init__.py
@@ -0,0 +1,2 @@
+from .session import Session
+from .scheduler import SCHEDULERS
diff --git a/flashcards-project/src/flashcards/card.py b/flashcards-project/src/flashcards/card.py
new file mode 100644
index 0000000..3278343
--- /dev/null
+++ b/flashcards-project/src/flashcards/card.py
@@ -0,0 +1,12 @@
+"""
+Defines a struct representing a single card. The struct takes the form:
+
+(front, back)
+"""
+from collections import namedtuple
+from hashlib import md5
+
+Card = namedtuple('Card', ['front', 'back'])
+
+def getId(card: Card) -> str:
+ return md5((card.front + card.back).encode("utf-8")).hexdigest()
diff --git a/flashcards-project/src/flashcards/parser.py b/flashcards-project/src/flashcards/parser.py
new file mode 100644
index 0000000..38abdcc
--- /dev/null
+++ b/flashcards-project/src/flashcards/parser.py
@@ -0,0 +1,108 @@
+"""
+Load .fcard files into dictionaries.
+
+The parser expects .fcard files in the following format:
+
+FRONT
+This is the front of the first card.
+
+BACK
+This is the back of the first card.
+
+FRONT
+This is another card.
+
+Multiple lines on the front are allowed.
+
+BACK
+Multiple lines on the back?
+
+Also allowed.
+
+FRONT
+...
+
+The cards are represented in dictionary entries of the form:
+
+id: card.Card
+"""
+from pathlib import Path
+from enum import Enum
+from typing import TextIO, Iterator
+
+from .card import Card, getId
+
+def _getCard(front_lines: list[str], back_lines: list[str]) -> tuple[str, Card]:
+ front_text = "".join(front_lines).strip()
+ back_text = "".join(back_lines).strip()
+ card = Card(front_text, back_text)
+ id = getId(card)
+ return id, card
+
+def _getCards(f: TextIO) -> Iterator[tuple[str, Card]]:
+ class State(Enum):
+ PARSE_FRONT = 1,
+ PARSE_BACK = 2
+
+ state = None
+ front_lines = []
+ back_lines = []
+
+ for i, line in enumerate(f):
+ match line.strip():
+ case "FRONT":
+ # Edge case: FRONT twice in a row
+ if state == State.PARSE_FRONT:
+ raise Exception(f"Unexpected 'FRONT': {f}:{i}")
+
+ # Next card is starting, wrap up current one
+ if state == State.PARSE_BACK:
+ yield _getCard(front_lines, back_lines)
+ front_lines.clear()
+ back_lines.clear()
+
+ state = State.PARSE_FRONT
+
+ case "BACK":
+ # Edge case: BACK without FRONT before it
+ if state != State.PARSE_FRONT:
+ raise Exception(f"Unexpected 'BACK': {f}:{i}")
+
+ state = State.PARSE_BACK
+
+ case _:
+ match state:
+ case State.PARSE_FRONT:
+ front_lines += line
+ case State.PARSE_BACK:
+ back_lines += line
+ # Edge case: file does not start with FRONT, flush preamble
+ case _:
+ continue
+
+ # Edge case: file did not end with contents of BACK
+ if state == State.PARSE_FRONT:
+ raise Exception(f"Unexpected end of file")
+
+ # Edge case: file was empty
+ if state is None:
+ return
+
+ yield _getCard(front_lines, back_lines)
+
+def parseFile(path: str) -> dict[str, Card]:
+ """
+ Parse a .fcard file and return a dictionary of Card instances indexed by ID.
+ """
+ with open(path, "r") as f:
+ return {id : card for id, card in _getCards(f)}
+
+def parseFiles(paths: list[str]) -> dict[str, Card]:
+ """
+ Parse a list of .fcard files and return a dictionary of Card instances indexed by ID.
+ """
+ cards = {}
+ for p in paths:
+ cards |= parseFile(p)
+ return cards
+
diff --git a/flashcards-project/src/flashcards/scheduler.py b/flashcards-project/src/flashcards/scheduler.py
new file mode 100644
index 0000000..a9d9470
--- /dev/null
+++ b/flashcards-project/src/flashcards/scheduler.py
@@ -0,0 +1,62 @@
+from typing import Protocol
+from abc import abstractmethod
+
+from .card import Card
+
+class Scheduler(Protocol):
+ """
+ Schedulers must implement this interface to be usable in a session.
+ """
+ @abstractmethod
+ def __init__(self, cards: dict[str, Card], state: dict):
+ """
+ Create a new instance of the scheduler from a dictionary of
+ Cards indexed by ID and a scheduler-specific state as a dict.
+ """
+ raise NotImplementedError
+
+ @abstractmethod
+ def practice(self, size: int) -> list[str]:
+ """
+ Return a list of card IDs of the requested size, if possible.
+ This list is intended for practice.
+ """
+ raise NotImplementedError
+
+ @abstractmethod
+ def test(self, size: int) -> list[str]:
+ """
+ Return a list of card IDs of the requested size, if possible.
+ This list is intended to test the player's knowledge.
+ """
+ raise NotImplementedError
+
+ @abstractmethod
+ def update(self, results: dict[str, int]) -> None:
+ """
+ Takes a dictionary of card IDs and integers, where the integer
+ is 0 if the player failed to guess the other side of the card,
+ of 1 if the player succeeded.
+ """
+ raise NotImplementedError
+
+ @abstractmethod
+ def getState(self) -> dict:
+ """
+ Return the scheduler's state for storage.
+ """
+ raise NotImplementedError
+
+SCHEDULERS = ["brutal"]
+
+def getSchedulerClass(name: str) -> Scheduler:
+ """
+ Returns the class object for the requested scheduler, if one exists.
+ """
+ match name:
+ case "brutal":
+ from .scheduler_brutal import SchedulerBrutal
+ return SchedulerBrutal
+ case _:
+ raise Exception(f"Unknown scheduler: {name}")
+
diff --git a/flashcards-project/src/flashcards/scheduler_brutal.py b/flashcards-project/src/flashcards/scheduler_brutal.py
new file mode 100644
index 0000000..ebbc0ff
--- /dev/null
+++ b/flashcards-project/src/flashcards/scheduler_brutal.py
@@ -0,0 +1,79 @@
+from random import shuffle
+
+from .scheduler import Scheduler
+from .card import Card
+
+HISTORY_DEPTH = 8
+
+class SchedulerBrutal(Scheduler):
+ """
+ The brutal scheduler tracks how well the player has consolidated each card
+ and also how often the card has been shown.
+
+ Using this information, it prioritizes cards that have been shown less
+ frequently and recently, which means the player will often see totally new
+ cards in test sessions.
+ """
+ def __init__(self, cards: dict[str, Card], state: dict):
+ self._cards = cards
+ self._state = {}
+
+ # Synchronise state with current card collection
+ for id, card in self._cards.items():
+ history = state.get(id, [None] * HISTORY_DEPTH)
+
+ # Adjust history if depth has changed
+ if len(history) > HISTORY_DEPTH:
+ history = history[-HISTORY_DEPTH:]
+ elif len(history) < HISTORY_DEPTH:
+ history = ([None] * (HISTORY_DEPTH - len(history))) + history
+
+ self._state[id] = history
+
+ def practice(self, size: int) -> list[str]:
+ return self._schedule(size)
+
+ def test(self, size: int) -> list[str]:
+ return self._schedule(size)
+
+ def update(self, results: dict[str, int]) -> None:
+ # Add card result to sliding window, or None if card was not shown
+ self._state = {id: history[1:] + [results.get(id, None)]
+ for id, history in self._state.items()}
+
+ def getState(self) -> dict:
+ return self._state
+
+ @staticmethod
+ def _consolidationIndex(history: list, weights: range) -> float:
+ """
+ Consolidation index is a measure of how well the player has guessed the card recently
+ """
+ relevant_history = [(h, w) for h, w in zip(history, weights) if h is not None]
+ weighted_history = sum([h * w for h, w in relevant_history])
+ total_weights = sum([w for h, w in relevant_history])
+ return weighted_history / total_weights if total_weights > 0 else 0.0
+
+ @staticmethod
+ def _exposureIndex(history: list) -> float:
+ """
+ Exposure index is a measure of how much and how recently a card has been shown
+ """
+ return sum([i + 1 for i, h in enumerate(history) if h is not None])
+
+ def _schedule(self, size: int) -> list[str]:
+ weights = range(10, 10 + HISTORY_DEPTH)
+ cards = [id for id, card in self._cards.items()]
+
+ # First sort by consolidation index
+ cards.sort(key=lambda id: SchedulerBrutal._consolidationIndex(self._state[id], weights))
+
+ # Next sort by exposure index
+ cards.sort(key=lambda id: SchedulerBrutal._exposureIndex(self._state[id]))
+
+ # Return least exposed and least consolidated cards, shuffled
+ cards = cards[0:size]
+
+ shuffle(cards)
+
+ return cards
diff --git a/flashcards-project/src/flashcards/session.py b/flashcards-project/src/flashcards/session.py
new file mode 100644
index 0000000..da444dd
--- /dev/null
+++ b/flashcards-project/src/flashcards/session.py
@@ -0,0 +1,51 @@
+from typing import Iterator, Callable
+
+from .card import Card
+from .scheduler import getSchedulerClass
+from .parser import parseFiles
+from .state_json import load, save
+
+class Session:
+ """
+ Represents a play session. During a session, multiple practice and test runs
+ can be made with the same scheduler.
+ """
+ def __init__(self, scheduler_name: str, card_files: list[str], state_file: str):
+ self._cards = parseFiles(card_files)
+ self._state_file = state_file
+ self._scheduler = getSchedulerClass(scheduler_name)(self._cards, load(state_file))
+
+ def practice(self, size: int) -> Iterator[Card]:
+ """
+ Yields cards for a practice run of the requested size.
+
+ Practice runs do not affect the scheduler state.
+ """
+ ids = self._scheduler.practice(size)
+ for id in ids:
+ yield self._cards[id]
+
+ def test(self, size: int) -> Iterator[tuple[Card, Callable]]:
+ """
+ Yields cards for a test run of the requested size.
+
+ A function is yielded with each card that takes single boolean argument.
+ The UI is expected to call the function for each card to indicate whether
+ the user correctly guessed the card (True) or not (False).
+
+ Multiple subsequent calls to the same function overwrite past results.
+
+ When the test run is done, the scheduler state is updated with the
+ collected results
+ """
+ ids = self._scheduler.practice(size)
+ results = {}
+
+ for id in ids:
+ def result(correct: bool) -> None:
+ results[id] = int(correct)
+ yield self._cards[id], result
+
+ self._scheduler.update(results)
+ save(self._state_file, self._scheduler.getState())
+
diff --git a/flashcards-project/src/flashcards/state_json.py b/flashcards-project/src/flashcards/state_json.py
new file mode 100644
index 0000000..673d904
--- /dev/null
+++ b/flashcards-project/src/flashcards/state_json.py
@@ -0,0 +1,23 @@
+"""
+Helper functions to store scheduler state as json
+"""
+import json
+from pathlib import Path
+
+def save(file: str, state: dict) -> None:
+ """
+ Dump the specified state dictionary in JSON format
+ """
+ with open(file, "w") as f:
+ json.dump(state, f)
+
+def load(file: str) -> dict:
+ """
+ Load the state from the specified file and return
+ an empty dictionary silently if the file doesn't exist.
+ """
+ try:
+ with open(file, "r") as f:
+ return json.load(f)
+ except:
+ return {}