anki/pylib/tests/test_schedv2.py
Damien Elmes 5e0a761b87
Move away from Bazel (#2202)
(for upgrading users, please see the notes at the bottom)

Bazel brought a lot of nice things to the table, such as rebuilds based on
content changes instead of modification times, caching of build products,
detection of incorrect build rules via a sandbox, and so on. Rewriting the build
in Bazel was also an opportunity to improve on the Makefile-based build we had
prior, which was pretty poor: most dependencies were external or not pinned, and
the build graph was poorly defined and mostly serialized. It was not uncommon
for fresh checkouts to fail due to floating dependencies, or for things to break
when trying to switch to an older commit.

For day-to-day development, I think Bazel served us reasonably well - we could
generally switch between branches while being confident that builds would be
correct and reasonably fast, and not require full rebuilds (except on Windows,
where the lack of a sandbox and the TS rules would cause build breakages when TS
files were renamed/removed).

Bazel achieves that reliability by defining rules for each programming language
that define how source files should be turned into outputs. For the rules to
work with Bazel's sandboxing approach, they often have to reimplement or
partially bypass the standard tools that each programming language provides. The
Rust rules call Rust's compiler directly for example, instead of using Cargo,
and the Python rules extract each PyPi package into a separate folder that gets
added to sys.path.

These separate language rules allow proper declaration of inputs and outputs,
and offer some advantages such as caching of build products and fine-grained
dependency installation. But they also bring some downsides:

- The rules don't always support use-cases/platforms that the standard language
tools do, meaning they need to be patched to be used. I've had to contribute a
number of patches to the Rust, Python and JS rules to unblock various issues.
- The dependencies we use with each language sometimes make assumptions that do
not hold in Bazel, meaning they either need to be pinned or patched, or the
language rules need to be adjusted to accommodate them.

I was hopeful that after the initial setup work, things would be relatively
smooth-sailing. Unfortunately, that has not proved to be the case. Things
frequently broke when dependencies or the language rules were updated, and I
began to get frustrated at the amount of Anki development time I was instead
spending on build system upkeep. It's now about 2 years since switching to
Bazel, and I think it's time to cut losses, and switch to something else that's
a better fit.

The new build system is based on a small build tool called Ninja, and some
custom Rust code in build/. This means that to build Anki, Bazel is no longer
required, but Ninja and Rust need to be installed on your system. Python and
Node toolchains are automatically downloaded like in Bazel.

This new build system should result in faster builds in some cases:

- Because we're using cargo to build now, Rust builds are able to take advantage
of pipelining and incremental debug builds, which we didn't have with Bazel.
It's also easier to override the default linker on Linux/macOS, which can
further improve speeds.
- External Rust crates are now built with opt=1, which improves performance
of debug builds.
- Esbuild is now used to transpile TypeScript, instead of invoking the TypeScript
compiler. This results in faster builds, by deferring typechecking to test/check
time, and by allowing more work to happen in parallel.

As an example of the differences, when testing with the mold linker on Linux,
adding a new message to tags.proto (which triggers a recompile of the bulk of
the Rust and TypeScript code) results in a compile that goes from about 22s on
Bazel to about 7s in the new system. With the standard linker, it's about 9s.

Some other changes of note:

- Our Rust workspace now uses cargo-hakari to ensure all packages agree on
available features, preventing unnecessary rebuilds.
- pylib/anki is now a PEP420 implicit namespace, avoiding the need to merge
source files and generated files into a single folder for running. By telling
VSCode about the extra search path, code completion now works with generated
files without needing to symlink them into the source folder.
- qt/aqt can't use PEP420 as it's difficult to get rid of aqt/__init__.py.
Instead, the generated files are now placed in a separate _aqt package that's
added to the path.
- ts/lib is now exposed as @tslib, so the source code and generated code can be
provided under the same namespace without a merging step.
- MyPy and PyLint are now invoked once for the entire codebase.
- dprint will be used to format TypeScript/json files in the future instead of
the slower prettier (currently turned off to avoid causing conflicts). It can
automatically defer to prettier when formatting Svelte files.
- svelte-check is now used for typechecking our Svelte code, which revealed a
few typing issues that went undetected with the old system.
- The Jest unit tests now work on Windows as well.

If you're upgrading from Bazel, updated usage instructions are in docs/development.md and docs/build.md. A summary of the changes:

- please remove node_modules and .bazel
- install rustup (https://rustup.rs/)
- install rsync if not already installed  (on windows, use pacman - see docs/windows.md)
- install Ninja (unzip from https://github.com/ninja-build/ninja/releases/tag/v1.11.1 and
  place on your path, or from your distro/homebrew if it's 1.10+)
- update .vscode/settings.json from .vscode.dist
2022-11-27 15:24:20 +10:00

1333 lines
36 KiB
Python

# Copyright: Ankitects Pty Ltd and contributors
# License: GNU AGPL, version 3 or later; http://www.gnu.org/licenses/agpl.html
import copy
import os
import time
from typing import Callable, Dict
import pytest
from anki import hooks
from anki.consts import *
from anki.lang import without_unicode_isolation
from anki.scheduler import UnburyDeck
from anki.utils import int_time
from tests.shared import getEmptyCol as getEmptyColOrig
# This file is used to exercise both the legacy Python 2.1 scheduler,
# and the experimental new one in Rust. Most tests run on both, but a few
# tests have been implemented separately where the behaviour differs.
def is_2021() -> bool:
return "2021" in os.getenv("PYTEST_CURRENT_TEST")
def getEmptyCol():
col = getEmptyColOrig()
col.upgrade_to_v2_scheduler()
if is_2021():
col.set_v3_scheduler(True)
return col
def test_clock():
col = getEmptyCol()
if (col.sched.day_cutoff - int_time()) < 10 * 60:
raise Exception("Unit tests will fail around the day rollover.")
def checkRevIvl(col, c, targetIvl):
if is_2021():
return
min, max = col.sched._fuzzIvlRange(targetIvl)
assert min <= c.ivl <= max
def test_basics():
col = getEmptyCol()
col.reset()
assert not col.sched.getCard()
def test_new():
col = getEmptyCol()
col.reset()
assert col.sched.newCount == 0
# add a note
note = col.newNote()
note["Front"] = "one"
note["Back"] = "two"
col.addNote(note)
col.reset()
assert col.sched.newCount == 1
# fetch it
c = col.sched.getCard()
assert c
assert c.queue == QUEUE_TYPE_NEW
assert c.type == CARD_TYPE_NEW
# if we answer it, it should become a learn card
t = int_time()
col.sched.answerCard(c, 1)
assert c.queue == QUEUE_TYPE_LRN
assert c.type == CARD_TYPE_LRN
assert c.due >= t
# disabled for now, as the learn fudging makes this randomly fail
# # the default order should ensure siblings are not seen together, and
# # should show all cards
# m = col.models.current(); mm = col.models
# t = mm.new_template("Reverse")
# t['qfmt'] = "{{Back}}"
# t['afmt'] = "{{Front}}"
# mm.add_template(m, t)
# mm.save(m)
# note = col.newNote()
# note['Front'] = u"2"; note['Back'] = u"2"
# col.addNote(note)
# note = col.newNote()
# note['Front'] = u"3"; note['Back'] = u"3"
# col.addNote(note)
# col.reset()
# qs = ("2", "3", "2", "3")
# for n in range(4):
# c = col.sched.getCard()
# assert qs[n] in c.question()
# col.sched.answerCard(c, 2)
def test_newLimits():
col = getEmptyCol()
# add some notes
deck2 = col.decks.id("Default::foo")
for i in range(30):
note = col.newNote()
note["Front"] = str(i)
if i > 4:
note.note_type()["did"] = deck2
col.addNote(note)
# give the child deck a different configuration
c2 = col.decks.add_config_returning_id("new conf")
col.decks.set_config_id_for_deck_dict(col.decks.get(deck2), c2)
col.reset()
# both confs have defaulted to a limit of 20
assert col.sched.newCount == 20
# first card we get comes from parent
c = col.sched.getCard()
assert c.did == 1
# limit the parent to 10 cards, meaning we get 10 in total
conf1 = col.decks.config_dict_for_deck_id(1)
conf1["new"]["perDay"] = 10
col.decks.save(conf1)
col.reset()
assert col.sched.newCount == 10
# if we limit child to 4, we should get 9
conf2 = col.decks.config_dict_for_deck_id(deck2)
conf2["new"]["perDay"] = 4
col.decks.save(conf2)
col.reset()
assert col.sched.newCount == 9
def test_newBoxes():
col = getEmptyCol()
note = col.newNote()
note["Front"] = "one"
col.addNote(note)
col.reset()
c = col.sched.getCard()
conf = col.sched._cardConf(c)
conf["new"]["delays"] = [1, 2, 3, 4, 5]
col.decks.save(conf)
col.sched.answerCard(c, 2)
# should handle gracefully
conf["new"]["delays"] = [1]
col.decks.save(conf)
col.sched.answerCard(c, 2)
def test_learn():
col = getEmptyCol()
# add a note
note = col.newNote()
note["Front"] = "one"
note["Back"] = "two"
col.addNote(note)
# set as a new card and rebuild queues
col.db.execute(f"update cards set queue={QUEUE_TYPE_NEW}, type={CARD_TYPE_NEW}")
col.reset()
# sched.getCard should return it, since it's due in the past
c = col.sched.getCard()
assert c
conf = col.sched._cardConf(c)
conf["new"]["delays"] = [0.5, 3, 10]
col.decks.save(conf)
# fail it
col.sched.answerCard(c, 1)
# it should have three reps left to graduation
assert c.left % 1000 == 3
# it should be due in 30 seconds
t = round(c.due - time.time())
assert t >= 25 and t <= 40
# pass it once
col.sched.answerCard(c, 3)
# it should be due in 3 minutes
dueIn = c.due - time.time()
assert 178 <= dueIn <= 180 * 1.25
assert c.left % 1000 == 2
# check log is accurate
log = col.db.first("select * from revlog order by id desc")
assert log[3] == 3
assert log[4] == -180
assert log[5] == -30
# pass again
col.sched.answerCard(c, 3)
# it should be due in 10 minutes
dueIn = c.due - time.time()
assert 599 <= dueIn <= 600 * 1.25
assert c.left % 1000 == 1
# the next pass should graduate the card
assert c.queue == QUEUE_TYPE_LRN
assert c.type == CARD_TYPE_LRN
col.sched.answerCard(c, 3)
assert c.queue == QUEUE_TYPE_REV
assert c.type == CARD_TYPE_REV
# should be due tomorrow, with an interval of 1
assert c.due == col.sched.today + 1
assert c.ivl == 1
# or normal removal
c.type = CARD_TYPE_NEW
c.queue = QUEUE_TYPE_LRN
c.flush()
col.sched.reset()
col.sched.answerCard(c, 4)
assert c.type == CARD_TYPE_REV
assert c.queue == QUEUE_TYPE_REV
checkRevIvl(col, c, 4)
# revlog should have been updated each time
assert col.db.scalar("select count() from revlog where type = 0") == 5
def test_relearn():
col = getEmptyCol()
note = col.newNote()
note["Front"] = "one"
col.addNote(note)
c = note.cards()[0]
c.ivl = 100
c.due = col.sched.today
c.queue = CARD_TYPE_REV
c.type = QUEUE_TYPE_REV
c.flush()
# fail the card
col.reset()
c = col.sched.getCard()
col.sched.answerCard(c, 1)
assert c.queue == QUEUE_TYPE_LRN
assert c.type == CARD_TYPE_RELEARNING
assert c.ivl == 1
# immediately graduate it
col.sched.answerCard(c, 4)
assert c.queue == CARD_TYPE_REV and c.type == QUEUE_TYPE_REV
assert c.ivl == 2
assert c.due == col.sched.today + c.ivl
def test_relearn_no_steps():
col = getEmptyCol()
note = col.newNote()
note["Front"] = "one"
col.addNote(note)
c = note.cards()[0]
c.ivl = 100
c.due = col.sched.today
c.queue = CARD_TYPE_REV
c.type = QUEUE_TYPE_REV
c.flush()
conf = col.decks.config_dict_for_deck_id(1)
conf["lapse"]["delays"] = []
col.decks.save(conf)
# fail the card
col.reset()
c = col.sched.getCard()
col.sched.answerCard(c, 1)
assert c.queue == CARD_TYPE_REV and c.type == QUEUE_TYPE_REV
def test_learn_collapsed():
col = getEmptyCol()
# add 2 notes
note = col.newNote()
note["Front"] = "1"
col.addNote(note)
note = col.newNote()
note["Front"] = "2"
col.addNote(note)
# set as a new card and rebuild queues
col.db.execute(f"update cards set queue={QUEUE_TYPE_NEW}, type={CARD_TYPE_NEW}")
col.reset()
# should get '1' first
c = col.sched.getCard()
assert c.question().endswith("1")
# pass it so it's due in 10 minutes
col.sched.answerCard(c, 3)
# get the other card
c = col.sched.getCard()
assert c.question().endswith("2")
# fail it so it's due in 1 minute
col.sched.answerCard(c, 1)
# we shouldn't get the same card again
c = col.sched.getCard()
assert not c.question().endswith("2")
def test_learn_day():
col = getEmptyCol()
# add a note
note = col.newNote()
note["Front"] = "one"
col.addNote(note)
note = col.newNote()
note["Front"] = "two"
col.addNote(note)
col.sched.reset()
c = col.sched.getCard()
conf = col.sched._cardConf(c)
conf["new"]["delays"] = [1, 10, 1440, 2880]
col.decks.save(conf)
# pass it
col.sched.answerCard(c, 3)
# two reps to graduate, 1 more today
assert c.left % 1000 == 3
assert col.sched.counts() == (1, 1, 0)
c.load()
ni = col.sched.nextIvl
assert ni(c, 3) == 86400
# answer the other dummy card
col.sched.answerCard(col.sched.getCard(), 4)
# answering the first one will place it in queue 3
c = col.sched.getCard()
col.sched.answerCard(c, 3)
assert c.due == col.sched.today + 1
assert c.queue == QUEUE_TYPE_DAY_LEARN_RELEARN
assert not col.sched.getCard()
# for testing, move it back a day
c.due -= 1
c.flush()
col.reset()
assert col.sched.counts() == (0, 1, 0)
c = col.sched.getCard()
# nextIvl should work
assert ni(c, 3) == 86400 * 2
# if we fail it, it should be back in the correct queue
col.sched.answerCard(c, 1)
assert c.queue == QUEUE_TYPE_LRN
if is_2021():
col.undo()
else:
col.undo_legacy()
col.reset()
c = col.sched.getCard()
col.sched.answerCard(c, 3)
# simulate the passing of another two days
c.due -= 2
c.flush()
col.reset()
# the last pass should graduate it into a review card
assert ni(c, 3) == 86400
col.sched.answerCard(c, 3)
assert c.queue == CARD_TYPE_REV and c.type == QUEUE_TYPE_REV
# if the lapse step is tomorrow, failing it should handle the counts
# correctly
c.due = 0
c.flush()
col.reset()
assert col.sched.counts() == (0, 0, 1)
conf = col.sched._cardConf(c)
conf["lapse"]["delays"] = [1440]
col.decks.save(conf)
c = col.sched.getCard()
col.sched.answerCard(c, 1)
assert c.queue == QUEUE_TYPE_DAY_LEARN_RELEARN
assert col.sched.counts() == (0, 0, 0)
def test_reviews():
col = getEmptyCol()
# add a note
note = col.newNote()
note["Front"] = "one"
note["Back"] = "two"
col.addNote(note)
# set the card up as a review card, due 8 days ago
c = note.cards()[0]
c.type = CARD_TYPE_REV
c.queue = QUEUE_TYPE_REV
c.due = col.sched.today - 8
c.factor = STARTING_FACTOR
c.reps = 3
c.lapses = 1
c.ivl = 100
c.start_timer()
c.flush()
# save it for later use as well
cardcopy = copy.copy(c)
# try with an ease of 2
##################################################
c = copy.copy(cardcopy)
c.flush()
col.reset()
col.sched.answerCard(c, 2)
assert c.queue == QUEUE_TYPE_REV
# the new interval should be (100) * 1.2 = 120
checkRevIvl(col, c, 120)
assert c.due == col.sched.today + c.ivl
# factor should have been decremented
assert c.factor == 2350
# check counters
assert c.lapses == 1
assert c.reps == 4
# ease 3
##################################################
c = copy.copy(cardcopy)
c.flush()
col.sched.answerCard(c, 3)
# the new interval should be (100 + 8/2) * 2.5 = 260
checkRevIvl(col, c, 260)
assert c.due == col.sched.today + c.ivl
# factor should have been left alone
assert c.factor == STARTING_FACTOR
# ease 4
##################################################
c = copy.copy(cardcopy)
c.flush()
col.sched.answerCard(c, 4)
# the new interval should be (100 + 8) * 2.5 * 1.3 = 351
checkRevIvl(col, c, 351)
assert c.due == col.sched.today + c.ivl
# factor should have been increased
assert c.factor == 2650
# leech handling
##################################################
conf = col.decks.get_config(1)
conf["lapse"]["leechAction"] = LEECH_SUSPEND
col.decks.save(conf)
c = copy.copy(cardcopy)
c.lapses = 7
c.flush()
# setup hook
hooked = []
def onLeech(card):
hooked.append(1)
hooks.card_did_leech.append(onLeech)
col.sched.answerCard(c, 1)
if not is_2021():
assert hooked
assert c.queue == QUEUE_TYPE_SUSPENDED
c.load()
assert c.queue == QUEUE_TYPE_SUSPENDED
assert "leech" in c.note().tags
def review_limits_setup() -> tuple[anki.collection.Collection, Dict]:
col = getEmptyCol()
parent = col.decks.get(col.decks.id("parent"))
child = col.decks.get(col.decks.id("parent::child"))
pconf = col.decks.get_config(col.decks.add_config_returning_id("parentConf"))
cconf = col.decks.get_config(col.decks.add_config_returning_id("childConf"))
pconf["rev"]["perDay"] = 5
col.decks.update_config(pconf)
col.decks.set_config_id_for_deck_dict(parent, pconf["id"])
cconf["rev"]["perDay"] = 10
col.decks.update_config(cconf)
col.decks.set_config_id_for_deck_dict(child, cconf["id"])
m = col.models.current()
m["did"] = child["id"]
col.models.save(m, updateReqs=False)
# add some cards
for i in range(20):
note = col.newNote()
note["Front"] = "one"
note["Back"] = "two"
col.addNote(note)
# make them reviews
c = note.cards()[0]
c.queue = CARD_TYPE_REV
c.type = QUEUE_TYPE_REV
c.due = 0
c.flush()
return col, child
def test_review_limits():
col, child = review_limits_setup()
tree = col.sched.deck_due_tree().children
# (('parent', 1514457677462, 5, 0, 0, (('child', 1514457677463, 5, 0, 0, ()),)))
assert tree[0].review_count == 5 # parent
assert tree[0].children[0].review_count == 10 # child
# .counts() should match
col.decks.select(child["id"])
col.sched.reset()
assert col.sched.counts() == (0, 0, 10)
# answering a card in the child should decrement parent count
c = col.sched.getCard()
col.sched.answerCard(c, 3)
assert col.sched.counts() == (0, 0, 9)
tree = col.sched.deck_due_tree().children
assert tree[0].review_count == 4 # parent
assert tree[0].children[0].review_count == 9 # child
def test_button_spacing():
col = getEmptyCol()
note = col.newNote()
note["Front"] = "one"
col.addNote(note)
# 1 day ivl review card due now
c = note.cards()[0]
c.type = CARD_TYPE_REV
c.queue = QUEUE_TYPE_REV
c.due = col.sched.today
c.reps = 1
c.ivl = 1
c.start_timer()
c.flush()
col.reset()
ni = col.sched.nextIvlStr
wo = without_unicode_isolation
assert wo(ni(c, 2)) == "2d"
assert wo(ni(c, 3)) == "3d"
assert wo(ni(c, 4)) == "4d"
# if hard factor is <= 1, then hard may not increase
conf = col.decks.config_dict_for_deck_id(1)
conf["rev"]["hardFactor"] = 1
col.decks.save(conf)
assert wo(ni(c, 2)) == "1d"
def test_overdue_lapse():
# disabled in commit 3069729776990980f34c25be66410e947e9d51a2
return
col = getEmptyCol() # pylint: disable=unreachable
# add a note
note = col.newNote()
note["Front"] = "one"
col.addNote(note)
# simulate a review that was lapsed and is now due for its normal review
c = note.cards()[0]
c.type = CARD_TYPE_REV
c.queue = QUEUE_TYPE_LRN
c.due = -1
c.odue = -1
c.factor = STARTING_FACTOR
c.left = 2002
c.ivl = 0
c.flush()
# checkpoint
col.save()
col.sched.reset()
assert col.sched.counts() == (0, 2, 0)
c = col.sched.getCard()
col.sched.answerCard(c, 3)
# it should be due tomorrow
assert c.due == col.sched.today + 1
# revert to before
col.rollback()
# with the default settings, the overdue card should be removed from the
# learning queue
col.sched.reset()
assert col.sched.counts() == (0, 0, 1)
def test_nextIvl():
col = getEmptyCol()
note = col.newNote()
note["Front"] = "one"
note["Back"] = "two"
col.addNote(note)
col.reset()
conf = col.decks.config_dict_for_deck_id(1)
conf["new"]["delays"] = [0.5, 3, 10]
conf["lapse"]["delays"] = [1, 5, 9]
col.decks.save(conf)
c = col.sched.getCard()
# new cards
##################################################
ni = col.sched.nextIvl
assert ni(c, 1) == 30
assert ni(c, 2) == (30 + 180) // 2
assert ni(c, 3) == 180
assert ni(c, 4) == 4 * 86400
col.sched.answerCard(c, 1)
# cards in learning
##################################################
assert ni(c, 1) == 30
assert ni(c, 2) == (30 + 180) // 2
assert ni(c, 3) == 180
assert ni(c, 4) == 4 * 86400
col.sched.answerCard(c, 3)
assert ni(c, 1) == 30
assert ni(c, 2) == 180
assert ni(c, 3) == 600
assert ni(c, 4) == 4 * 86400
col.sched.answerCard(c, 3)
# normal graduation is tomorrow
assert ni(c, 3) == 1 * 86400
assert ni(c, 4) == 4 * 86400
# lapsed cards
##################################################
c.type = CARD_TYPE_RELEARNING
c.ivl = 100
c.factor = STARTING_FACTOR
c.flush()
assert ni(c, 1) == 60
assert ni(c, 3) == 100 * 86400
assert ni(c, 4) == 101 * 86400
# review cards
##################################################
c.type = CARD_TYPE_REV
c.queue = QUEUE_TYPE_REV
c.ivl = 100
c.factor = STARTING_FACTOR
c.flush()
# failing it should put it at 60s
assert ni(c, 1) == 60
# or 1 day if relearn is false
conf["lapse"]["delays"] = []
col.decks.save(conf)
assert ni(c, 1) == 1 * 86400
# (* 100 1.2 86400)10368000.0
assert ni(c, 2) == 10368000
# (* 100 2.5 86400)21600000.0
assert ni(c, 3) == 21600000
# (* 100 2.5 1.3 86400)28080000.0
assert ni(c, 4) == 28080000
assert without_unicode_isolation(col.sched.nextIvlStr(c, 4)) == "10.8mo"
def test_bury():
col = getEmptyCol()
note = col.newNote()
note["Front"] = "one"
col.addNote(note)
c = note.cards()[0]
note = col.newNote()
note["Front"] = "two"
col.addNote(note)
c2 = note.cards()[0]
# burying
col.sched.bury_cards([c.id], manual=True) # pylint: disable=unexpected-keyword-arg
c.load()
assert c.queue == QUEUE_TYPE_MANUALLY_BURIED
col.sched.bury_cards(
[c2.id], manual=False
) # pylint: disable=unexpected-keyword-arg
c2.load()
assert c2.queue == QUEUE_TYPE_SIBLING_BURIED
col.reset()
assert not col.sched.getCard()
col.sched.unbury_deck(deck_id=col.decks.get_current_id(), mode=UnburyDeck.USER_ONLY)
c.load()
assert c.queue == QUEUE_TYPE_NEW
c2.load()
assert c2.queue == QUEUE_TYPE_SIBLING_BURIED
col.sched.unbury_deck(
deck_id=col.decks.get_current_id(), mode=UnburyDeck.SCHED_ONLY
)
c2.load()
assert c2.queue == QUEUE_TYPE_NEW
col.sched.bury_cards([c.id, c2.id])
col.sched.unbury_deck(deck_id=col.decks.get_current_id())
col.reset()
assert col.sched.counts() == (2, 0, 0)
def test_suspend():
col = getEmptyCol()
note = col.newNote()
note["Front"] = "one"
col.addNote(note)
c = note.cards()[0]
# suspending
col.reset()
assert col.sched.getCard()
col.sched.suspend_cards([c.id])
col.reset()
assert not col.sched.getCard()
# unsuspending
col.sched.unsuspend_cards([c.id])
col.reset()
assert col.sched.getCard()
# should cope with rev cards being relearnt
c.due = 0
c.ivl = 100
c.type = CARD_TYPE_REV
c.queue = QUEUE_TYPE_REV
c.flush()
col.reset()
c = col.sched.getCard()
col.sched.answerCard(c, 1)
assert c.due >= time.time()
due = c.due
assert c.queue == QUEUE_TYPE_LRN
assert c.type == CARD_TYPE_RELEARNING
col.sched.suspend_cards([c.id])
col.sched.unsuspend_cards([c.id])
c.load()
assert c.queue == QUEUE_TYPE_LRN
assert c.type == CARD_TYPE_RELEARNING
assert c.due == due
# should cope with cards in cram decks
c.due = 1
c.flush()
did = col.decks.new_filtered("tmp")
col.sched.rebuild_filtered_deck(did)
c.load()
assert c.due != 1
assert c.did != 1
col.sched.suspend_cards([c.id])
c.load()
assert c.due != 1
assert c.did != 1
assert c.odue == 1
def test_filt_reviewing_early_normal():
def to_int(val: float) -> int:
if is_2021():
return round(val)
else:
return int(val)
col = getEmptyCol()
note = col.newNote()
note["Front"] = "one"
col.addNote(note)
c = note.cards()[0]
c.ivl = 100
c.queue = CARD_TYPE_REV
c.type = QUEUE_TYPE_REV
# due in 25 days, so it's been waiting 75 days
c.due = col.sched.today + 25
c.mod = 1
c.factor = STARTING_FACTOR
c.start_timer()
c.flush()
col.reset()
assert col.sched.counts() == (0, 0, 0)
# create a dynamic deck and refresh it
did = col.decks.new_filtered("Cram")
col.sched.rebuild_filtered_deck(did)
col.reset()
# should appear as normal in the deck list
assert sorted(col.sched.deck_due_tree().children)[0].review_count == 1
# and should appear in the counts
assert col.sched.counts() == (0, 0, 1)
# grab it and check estimates
c = col.sched.getCard()
assert col.sched.answerButtons(c) == 4
assert col.sched.nextIvl(c, 1) == 600
assert col.sched.nextIvl(c, 2) == to_int(75 * 1.2) * 86400
assert col.sched.nextIvl(c, 3) == to_int(75 * 2.5) * 86400
assert col.sched.nextIvl(c, 4) == to_int(75 * 2.5 * 1.15) * 86400
# answer 'good'
col.sched.answerCard(c, 3)
checkRevIvl(col, c, int(75 * 2.5))
assert c.due == col.sched.today + c.ivl
assert not c.odue
# should not be in learning
assert c.queue == QUEUE_TYPE_REV
# should be logged as a cram rep
assert col.db.scalar("select type from revlog order by id desc limit 1") == 3
# due in 75 days, so it's been waiting 25 days
c.ivl = 100
c.due = col.sched.today + 75
c.flush()
col.sched.rebuild_filtered_deck(did)
col.reset()
c = col.sched.getCard()
assert col.sched.nextIvl(c, 2) == 100 * 1.2 / 2 * 86400
assert col.sched.nextIvl(c, 3) == 100 * 86400
assert col.sched.nextIvl(c, 4) == to_int(100 * (1.3 - (1.3 - 1) / 2)) * 86400
def test_filt_keep_lrn_state():
col = getEmptyCol()
note = col.newNote()
note["Front"] = "one"
col.addNote(note)
# fail the card outside filtered deck
c = col.sched.getCard()
conf = col.sched._cardConf(c)
conf["new"]["delays"] = [1, 10, 61]
col.decks.save(conf)
col.sched.answerCard(c, 1)
assert c.type == CARD_TYPE_LRN and c.queue == QUEUE_TYPE_LRN
assert c.left % 1000 == 3
col.sched.answerCard(c, 3)
assert c.type == CARD_TYPE_LRN and c.queue == QUEUE_TYPE_LRN
# create a dynamic deck and refresh it
did = col.decks.new_filtered("Cram")
col.sched.rebuild_filtered_deck(did)
col.reset()
# card should still be in learning state
c.load()
assert c.type == CARD_TYPE_LRN and c.queue == QUEUE_TYPE_LRN
assert c.left % 1000 == 2
# should be able to advance learning steps
col.sched.answerCard(c, 3)
# should be due at least an hour in the future
assert c.due - int_time() > 60 * 60
# emptying the deck preserves learning state
col.sched.empty_filtered_deck(did)
c.load()
assert c.type == CARD_TYPE_LRN and c.queue == QUEUE_TYPE_LRN
assert c.left % 1000 == 1
assert c.due - int_time() > 60 * 60
def test_preview():
# add cards
col = getEmptyCol()
note = col.newNote()
note["Front"] = "one"
col.addNote(note)
c = note.cards()[0]
orig = copy.copy(c)
note2 = col.newNote()
note2["Front"] = "two"
col.addNote(note2)
# cram deck
did = col.decks.new_filtered("Cram")
cram = col.decks.get(did)
cram["resched"] = False
col.decks.save(cram)
col.sched.rebuild_filtered_deck(did)
col.reset()
# grab the first card
c = col.sched.getCard()
if is_2021():
passing_grade = 4
else:
passing_grade = 2
assert col.sched.answerButtons(c) == passing_grade
assert col.sched.nextIvl(c, 1) == 600
assert col.sched.nextIvl(c, passing_grade) == 0
# failing it will push its due time back
due = c.due
col.sched.answerCard(c, 1)
assert c.due != due
# the other card should come next
c2 = col.sched.getCard()
assert c2.id != c.id
# passing it will remove it
col.sched.answerCard(c2, passing_grade)
assert c2.queue == QUEUE_TYPE_NEW
assert c2.reps == 0
assert c2.type == CARD_TYPE_NEW
# the other card should appear again
c = col.sched.getCard()
assert c.id == orig.id
# emptying the filtered deck should restore card
col.sched.empty_filtered_deck(did)
c.load()
assert c.queue == QUEUE_TYPE_NEW
assert c.reps == 0
assert c.type == CARD_TYPE_NEW
def test_ordcycle():
col = getEmptyCol()
# add two more templates and set second active
m = col.models.current()
mm = col.models
t = mm.new_template("Reverse")
t["qfmt"] = "{{Back}}"
t["afmt"] = "{{Front}}"
mm.add_template(m, t)
t = mm.new_template("f2")
t["qfmt"] = "{{Front}}2"
t["afmt"] = "{{Back}}"
mm.add_template(m, t)
mm.save(m)
# create a new note; it should have 3 cards
note = col.newNote()
note["Front"] = "1"
note["Back"] = "1"
col.addNote(note)
assert col.card_count() == 3
conf = col.decks.get_config(1)
conf["new"]["bury"] = False
col.decks.save(conf)
col.reset()
# ordinals should arrive in order
for i in range(3):
c = col.sched.getCard()
assert c.ord == i
col.sched.answerCard(c, 4)
def test_counts_idx():
if is_2021():
pytest.skip("old sched only")
col = getEmptyCol()
note = col.newNote()
note["Front"] = "one"
note["Back"] = "two"
col.addNote(note)
col.reset()
assert col.sched.counts() == (1, 0, 0)
c = col.sched.getCard()
# counter's been decremented but idx indicates 1
assert col.sched.counts() == (0, 0, 0)
assert col.sched.countIdx(c) == 0
# answer to move to learn queue
col.sched.answerCard(c, 1)
assert col.sched.counts() == (0, 1, 0)
# fetching again will decrement the count
c = col.sched.getCard()
assert col.sched.counts() == (0, 0, 0)
assert col.sched.countIdx(c) == 1
# answering should add it back again
col.sched.answerCard(c, 1)
assert col.sched.counts() == (0, 1, 0)
def test_counts_idx_new():
if not is_2021():
pytest.skip("new sched only")
col = getEmptyCol()
note = col.newNote()
note["Front"] = "one"
note["Back"] = "two"
col.addNote(note)
note = col.newNote()
note["Front"] = "two"
note["Back"] = "two"
col.addNote(note)
col.reset()
assert col.sched.counts() == (2, 0, 0)
c = col.sched.getCard()
# getCard does not decrement counts
assert col.sched.counts() == (2, 0, 0)
assert col.sched.countIdx(c) == 0
# answer to move to learn queue
col.sched.answerCard(c, 1)
assert col.sched.counts() == (1, 1, 0)
assert col.sched.countIdx(c) == 1
# fetching next will not decrement the count
c = col.sched.getCard()
assert col.sched.counts() == (1, 1, 0)
assert col.sched.countIdx(c) == 0
def test_repCounts():
col = getEmptyCol()
note = col.newNote()
note["Front"] = "one"
col.addNote(note)
note = col.newNote()
note["Front"] = "two"
col.addNote(note)
col.reset()
# lrnReps should be accurate on pass/fail
assert col.sched.counts() == (2, 0, 0)
col.sched.answerCard(col.sched.getCard(), 1)
assert col.sched.counts() == (1, 1, 0)
col.sched.answerCard(col.sched.getCard(), 1)
assert col.sched.counts() == (0, 2, 0)
col.sched.answerCard(col.sched.getCard(), 3)
assert col.sched.counts() == (0, 2, 0)
col.sched.answerCard(col.sched.getCard(), 1)
assert col.sched.counts() == (0, 2, 0)
col.sched.answerCard(col.sched.getCard(), 3)
assert col.sched.counts() == (0, 1, 0)
col.sched.answerCard(col.sched.getCard(), 4)
assert col.sched.counts() == (0, 0, 0)
note = col.newNote()
note["Front"] = "three"
col.addNote(note)
note = col.newNote()
note["Front"] = "four"
col.addNote(note)
col.reset()
# initial pass and immediate graduate should be correct too
assert col.sched.counts() == (2, 0, 0)
col.sched.answerCard(col.sched.getCard(), 3)
assert col.sched.counts() == (1, 1, 0)
col.sched.answerCard(col.sched.getCard(), 4)
assert col.sched.counts() == (0, 1, 0)
col.sched.answerCard(col.sched.getCard(), 4)
assert col.sched.counts() == (0, 0, 0)
# and failing a review should too
note = col.newNote()
note["Front"] = "five"
col.addNote(note)
c = note.cards()[0]
c.type = CARD_TYPE_REV
c.queue = QUEUE_TYPE_REV
c.due = col.sched.today
c.flush()
note = col.newNote()
note["Front"] = "six"
col.addNote(note)
col.reset()
assert col.sched.counts() == (1, 0, 1)
col.sched.answerCard(col.sched.getCard(), 1)
assert col.sched.counts() == (1, 1, 0)
def test_timing():
col = getEmptyCol()
# add a few review cards, due today
for i in range(5):
note = col.newNote()
note["Front"] = f"num{str(i)}"
col.addNote(note)
c = note.cards()[0]
c.type = CARD_TYPE_REV
c.queue = QUEUE_TYPE_REV
c.due = 0
c.flush()
# fail the first one
col.reset()
c = col.sched.getCard()
col.sched.answerCard(c, 1)
# the next card should be another review
c2 = col.sched.getCard()
assert c2.queue == QUEUE_TYPE_REV
# if the failed card becomes due, it should show first
c.due = int_time() - 1
c.flush()
col.reset()
c = col.sched.getCard()
assert c.queue == QUEUE_TYPE_LRN
def test_collapse():
col = getEmptyCol()
# add a note
note = col.newNote()
note["Front"] = "one"
col.addNote(note)
# and another, so we don't get the same twice in a row
note = col.newNote()
note["Front"] = "two"
col.addNote(note)
col.reset()
# first note
c = col.sched.getCard()
col.sched.answerCard(c, 1)
# second note
c2 = col.sched.getCard()
assert c2.nid != c.nid
col.sched.answerCard(c2, 1)
# first should become available again, despite it being due in the future
c3 = col.sched.getCard()
assert c3.due > int_time()
col.sched.answerCard(c3, 4)
# answer other
c4 = col.sched.getCard()
col.sched.answerCard(c4, 4)
assert not col.sched.getCard()
def test_deckDue():
col = getEmptyCol()
# add a note with default deck
note = col.newNote()
note["Front"] = "one"
col.addNote(note)
# and one that's a child
note = col.newNote()
note["Front"] = "two"
default1 = note.note_type()["did"] = col.decks.id("Default::1")
col.addNote(note)
# make it a review card
c = note.cards()[0]
c.queue = QUEUE_TYPE_REV
c.due = 0
c.flush()
# add one more with a new deck
note = col.newNote()
note["Front"] = "two"
note.note_type()["did"] = col.decks.id("foo::bar")
col.addNote(note)
# and one that's a sibling
note = col.newNote()
note["Front"] = "three"
note.note_type()["did"] = col.decks.id("foo::baz")
col.addNote(note)
col.reset()
assert len(col.decks.all_names_and_ids()) == 5
tree = col.sched.deck_due_tree().children
assert tree[0].name == "Default"
# sum of child and parent
assert tree[0].deck_id == 1
assert tree[0].review_count == 1
assert tree[0].new_count == 1
# child count is just review
child = tree[0].children[0]
assert child.name == "1"
assert child.deck_id == default1
assert child.review_count == 1
assert child.new_count == 0
# code should not fail if a card has an invalid deck
c.did = 12345
c.flush()
col.sched.deck_due_tree()
def test_deckTree():
col = getEmptyCol()
col.decks.id("new::b::c")
col.decks.id("new2")
# new should not appear twice in tree
names = [x.name for x in col.sched.deck_due_tree().children]
names.remove("new")
assert "new" not in names
def test_deckFlow():
col = getEmptyCol()
# add a note with default deck
note = col.newNote()
note["Front"] = "one"
col.addNote(note)
# and one that's a child
note = col.newNote()
note["Front"] = "two"
note.note_type()["did"] = col.decks.id("Default::2")
col.addNote(note)
# and another that's higher up
note = col.newNote()
note["Front"] = "three"
default1 = note.note_type()["did"] = col.decks.id("Default::1")
col.addNote(note)
col.reset()
assert col.sched.counts() == (3, 0, 0)
# should get top level one first, then ::1, then ::2
for i in "one", "three", "two":
c = col.sched.getCard()
assert c.note()["Front"] == i
col.sched.answerCard(c, 3)
def test_reorder():
col = getEmptyCol()
# add a note with default deck
note = col.newNote()
note["Front"] = "one"
col.addNote(note)
note2 = col.newNote()
note2["Front"] = "two"
col.addNote(note2)
assert note2.cards()[0].due == 2
found = False
# 50/50 chance of being reordered
for i in range(20):
col.sched.randomize_cards(1)
if note.cards()[0].due != note.id:
found = True
break
assert found
col.sched.order_cards(1)
assert note.cards()[0].due == 1
# shifting
note3 = col.newNote()
note3["Front"] = "three"
col.addNote(note3)
note4 = col.newNote()
note4["Front"] = "four"
col.addNote(note4)
assert note.cards()[0].due == 1
assert note2.cards()[0].due == 2
assert note3.cards()[0].due == 3
assert note4.cards()[0].due == 4
col.sched.reposition_new_cards(
[note3.cards()[0].id, note4.cards()[0].id],
starting_from=1,
shift_existing=True,
step_size=1,
randomize=False,
)
assert note.cards()[0].due == 3
assert note2.cards()[0].due == 4
assert note3.cards()[0].due == 1
assert note4.cards()[0].due == 2
def test_forget():
col = getEmptyCol()
note = col.newNote()
note["Front"] = "one"
col.addNote(note)
c = note.cards()[0]
c.queue = QUEUE_TYPE_REV
c.type = CARD_TYPE_REV
c.ivl = 100
c.due = 0
c.flush()
col.reset()
assert col.sched.counts() == (0, 0, 1)
col.sched.forgetCards([c.id])
col.reset()
assert col.sched.counts() == (1, 0, 0)
def test_resched():
col = getEmptyCol()
note = col.newNote()
note["Front"] = "one"
col.addNote(note)
c = note.cards()[0]
col.sched.set_due_date([c.id], "0")
c.load()
assert c.due == col.sched.today
assert c.ivl == 1
assert c.queue == QUEUE_TYPE_REV and c.type == CARD_TYPE_REV
# make it due tomorrow
col.sched.set_due_date([c.id], "1")
c.load()
assert c.due == col.sched.today + 1
assert c.ivl == 1
def test_norelearn():
col = getEmptyCol()
# add a note
note = col.newNote()
note["Front"] = "one"
col.addNote(note)
c = note.cards()[0]
c.type = CARD_TYPE_REV
c.queue = QUEUE_TYPE_REV
c.due = 0
c.factor = STARTING_FACTOR
c.reps = 3
c.lapses = 1
c.ivl = 100
c.start_timer()
c.flush()
col.reset()
col.sched.answerCard(c, 1)
col.sched._cardConf(c)["lapse"]["delays"] = []
col.sched.answerCard(c, 1)
def test_failmult():
col = getEmptyCol()
note = col.newNote()
note["Front"] = "one"
note["Back"] = "two"
col.addNote(note)
c = note.cards()[0]
c.type = CARD_TYPE_REV
c.queue = QUEUE_TYPE_REV
c.ivl = 100
c.due = col.sched.today - c.ivl
c.factor = STARTING_FACTOR
c.reps = 3
c.lapses = 1
c.start_timer()
c.flush()
conf = col.sched._cardConf(c)
conf["lapse"]["mult"] = 0.5
col.decks.save(conf)
c = col.sched.getCard()
col.sched.answerCard(c, 1)
assert c.ivl == 50
col.sched.answerCard(c, 1)
assert c.ivl == 25
# cards with a due date earlier than the collection should retain
# their due date when removed
def test_negativeDueFilter():
col = getEmptyCol()
# card due prior to collection date
note = col.newNote()
note["Front"] = "one"
note["Back"] = "two"
col.addNote(note)
c = note.cards()[0]
c.due = -5
c.queue = QUEUE_TYPE_REV
c.ivl = 5
c.flush()
# into and out of filtered deck
did = col.decks.new_filtered("Cram")
col.sched.rebuild_filtered_deck(did)
col.sched.empty_filtered_deck(did)
col.reset()
c.load()
assert c.due == -5
# hard on the first step should be the average of again and good,
# and it should be logged properly
def test_initial_repeat():
col = getEmptyCol()
note = col.newNote()
note["Front"] = "one"
note["Back"] = "two"
col.addNote(note)
col.reset()
c = col.sched.getCard()
col.sched.answerCard(c, 2)
# should be due in ~ 5.5 mins
expected = time.time() + 5.5 * 60
assert expected - 10 < c.due < expected * 1.25
ivl = col.db.scalar("select ivl from revlog")
assert ivl == -5.5 * 60