Anki/pylib/anki/models.py
Damien Elmes 61e86cc29d new change notetype implementation for the frontend
- changes can now be undone
- the same field can now be mapped to multiple target fields, allowing
fields to be cloned
- the old Qt dialog has been removed
- the old col.models.change() API calls the new code, to avoid
breaking existing consumers. It requires the field map to always
be passed in, but that appears to have been the common case.
- closes #1175
2021-06-10 22:19:24 +10:00

530 lines
17 KiB
Python

# Copyright: Ankitects Pty Ltd and contributors
# License: GNU AGPL, version 3 or later; http://www.gnu.org/licenses/agpl.html
from __future__ import annotations
import copy
import pprint
import sys
import time
import traceback
from typing import Any, Dict, List, NewType, Optional, Sequence, Tuple, Union
import anki # pylint: disable=unused-import
import anki._backend.backend_pb2 as _pb
from anki.collection import OpChanges, OpChangesWithId
from anki.consts import *
from anki.errors import NotFoundError
from anki.lang import without_unicode_isolation
from anki.stdmodels import StockNotetypeKind
from anki.utils import checksum, from_json_bytes, to_json_bytes
# public exports
NotetypeNameId = _pb.NotetypeNameId
NotetypeNameIdUseCount = _pb.NotetypeNameIdUseCount
NotetypeNames = _pb.NotetypeNames
ChangeNotetypeInfo = _pb.ChangeNotetypeInfo
ChangeNotetypeIn = _pb.ChangeNotetypeIn
# legacy types
NotetypeDict = Dict[str, Any]
NoteType = NotetypeDict
FieldDict = Dict[str, Any]
TemplateDict = Dict[str, Union[str, int, None]]
NotetypeId = NewType("NotetypeId", int)
sys.modules["anki.models"].NoteType = NotetypeDict # type: ignore
class ModelsDictProxy:
def __init__(self, col: anki.collection.Collection):
self._col = col.weakref()
def _warn(self) -> None:
traceback.print_stack(file=sys.stdout)
print("add-on should use methods on col.models, not col.models.models dict")
def __getitem__(self, item: Any) -> Any:
self._warn()
return self._col.models.get(NotetypeId(int(item)))
def __setitem__(self, key: str, val: Any) -> None:
self._warn()
self._col.models.save(val)
def __len__(self) -> int:
self._warn()
return len(self._col.models.all_names_and_ids())
def keys(self) -> Any:
self._warn()
return [str(nt.id) for nt in self._col.models.all_names_and_ids()]
def values(self) -> Any:
self._warn()
return self._col.models.all()
def items(self) -> Any:
self._warn()
return [(str(nt["id"]), nt) for nt in self._col.models.all()]
def __contains__(self, item: Any) -> bool:
self._warn()
return self._col.models.have(item)
class ModelManager:
# Saving/loading registry
#############################################################
def __init__(self, col: anki.collection.Collection) -> None:
self.col = col.weakref()
self.models = ModelsDictProxy(col)
# do not access this directly!
self._cache = {}
def __repr__(self) -> str:
d = dict(self.__dict__)
del d["col"]
return f"{super().__repr__()} {pprint.pformat(d, width=300)}"
def save(
self,
m: NotetypeDict = None,
# no longer used
templates: bool = False,
updateReqs: bool = True,
) -> None:
"Save changes made to provided note type."
if not m:
print("col.models.save() should be passed the changed notetype")
return
self.update(m, preserve_usn=False)
# legacy
def flush(self) -> None:
pass
# Caching
#############################################################
# A lot of existing code expects to be able to quickly and
# frequently obtain access to an entire notetype, so we currently
# need to cache responses from the backend. Please do not
# access the cache directly!
_cache: Dict[NotetypeId, NotetypeDict] = {}
def _update_cache(self, nt: NotetypeDict) -> None:
self._cache[nt["id"]] = nt
def _remove_from_cache(self, ntid: NotetypeId) -> None:
if ntid in self._cache:
del self._cache[ntid]
def _get_cached(self, ntid: NotetypeId) -> Optional[NotetypeDict]:
return self._cache.get(ntid)
def _clear_cache(self) -> None:
self._cache = {}
# Listing note types
#############################################################
def all_names_and_ids(self) -> Sequence[NotetypeNameId]:
return self.col._backend.get_notetype_names()
def all_use_counts(self) -> Sequence[NotetypeNameIdUseCount]:
return self.col._backend.get_notetype_names_and_counts()
# legacy
def allNames(self) -> List[str]:
return [n.name for n in self.all_names_and_ids()]
def ids(self) -> List[NotetypeId]:
return [NotetypeId(n.id) for n in self.all_names_and_ids()]
# only used by importing code
def have(self, id: NotetypeId) -> bool:
if isinstance(id, str):
id = int(id)
return any(True for e in self.all_names_and_ids() if e.id == id)
# Current note type
#############################################################
def current(self, forDeck: bool = True) -> NotetypeDict:
"Get current model."
m = self.get(self.col.decks.current().get("mid"))
if not forDeck or not m:
m = self.get(self.col.conf["curModel"])
if m:
return m
return self.get(NotetypeId(self.all_names_and_ids()[0].id))
def setCurrent(self, m: NotetypeDict) -> None:
"""Legacy. The current notetype is now updated on note add."""
self.col.set_config("curModel", m["id"])
# Retrieving and creating models
#############################################################
def id_for_name(self, name: str) -> Optional[NotetypeId]:
try:
return NotetypeId(self.col._backend.get_notetype_id_by_name(name))
except NotFoundError:
return None
def get(self, id: NotetypeId) -> Optional[NotetypeDict]:
"Get model with ID, or None."
# deal with various legacy input types
if id is None:
return None
elif isinstance(id, str):
id = int(id)
nt = self._get_cached(id)
if not nt:
try:
nt = from_json_bytes(self.col._backend.get_notetype_legacy(id))
self._update_cache(nt)
except NotFoundError:
return None
return nt
def all(self) -> List[NotetypeDict]:
"Get all models."
return [self.get(NotetypeId(nt.id)) for nt in self.all_names_and_ids()]
def byName(self, name: str) -> Optional[NotetypeDict]:
"Get model with NAME."
id = self.id_for_name(name)
if id:
return self.get(id)
else:
return None
def new(self, name: str) -> NotetypeDict:
"Create a new model, and return it."
# caller should call save() after modifying
nt = from_json_bytes(
self.col._backend.get_stock_notetype_legacy(StockNotetypeKind.BASIC)
)
nt["flds"] = []
nt["tmpls"] = []
nt["name"] = name
return nt
def rem(self, m: NotetypeDict) -> None:
"Delete model, and all its cards/notes."
self.remove(m["id"])
def remove_all_notetypes(self) -> None:
for nt in self.all_names_and_ids():
self._remove_from_cache(NotetypeId(nt.id))
self.col._backend.remove_notetype(nt.id)
def remove(self, id: NotetypeId) -> OpChanges:
"Modifies schema."
self._remove_from_cache(id)
return self.col._backend.remove_notetype(id)
def add(self, m: NotetypeDict) -> OpChangesWithId:
"Replaced with add_dict()"
self.ensureNameUnique(m)
out = self.col._backend.add_notetype_legacy(to_json_bytes(m))
m["id"] = out.id
self._mutate_after_write(m)
return out
def add_dict(self, m: NotetypeDict) -> OpChangesWithId:
"Notetype needs to be fetched from DB after adding."
self.ensureNameUnique(m)
return self.col._backend.add_notetype_legacy(to_json_bytes(m))
def ensureNameUnique(self, m: NotetypeDict) -> None:
existing_id = self.id_for_name(m["name"])
if existing_id is not None and existing_id != m["id"]:
m["name"] += "-" + checksum(str(time.time()))[:5]
def update(self, m: NotetypeDict, preserve_usn: bool = True) -> None:
"Add or update an existing model. Use .update_dict() instead."
self._remove_from_cache(m["id"])
self.ensureNameUnique(m)
m["id"] = self.col._backend.add_or_update_notetype(
json=to_json_bytes(m), preserve_usn_and_mtime=preserve_usn
)
self.setCurrent(m)
self._mutate_after_write(m)
def update_dict(self, m: NotetypeDict) -> OpChanges:
"Update a NotetypeDict. Caller will need to re-load notetype if new fields/cards added."
self._remove_from_cache(m["id"])
self.ensureNameUnique(m)
return self.col._backend.update_notetype_legacy(to_json_bytes(m))
def _mutate_after_write(self, nt: NotetypeDict) -> None:
# existing code expects the note type to be mutated to reflect
# the changes made when adding, such as ordinal assignment :-(
updated = self.get(nt["id"])
nt.update(updated)
# Tools
##################################################
def nids(self, ntid: NotetypeId) -> List[anki.notes.NoteId]:
"Note ids for M."
if isinstance(ntid, dict):
# legacy callers passed in note type
ntid = ntid["id"]
return self.col.db.list("select id from notes where mid = ?", ntid)
def useCount(self, m: NotetypeDict) -> int:
"Number of note using M."
return self.col.db.scalar("select count() from notes where mid = ?", m["id"])
# Copying
##################################################
def copy(self, m: NotetypeDict, add: bool = True) -> NotetypeDict:
"Copy, save and return."
m2 = copy.deepcopy(m)
m2["name"] = without_unicode_isolation(
self.col.tr.notetypes_copy(val=m2["name"])
)
m2["id"] = 0
if add:
self.add(m2)
return m2
# Fields
##################################################
def fieldMap(self, m: NotetypeDict) -> Dict[str, Tuple[int, FieldDict]]:
"Mapping of field name -> (ord, field)."
return {f["name"]: (f["ord"], f) for f in m["flds"]}
def fieldNames(self, m: NotetypeDict) -> List[str]:
return [f["name"] for f in m["flds"]]
def sortIdx(self, m: NotetypeDict) -> int:
return m["sortf"]
# Adding & changing fields
##################################################
def new_field(self, name: str) -> FieldDict:
assert isinstance(name, str)
nt = from_json_bytes(
self.col._backend.get_stock_notetype_legacy(StockNotetypeKind.BASIC)
)
field = nt["flds"][0]
field["name"] = name
field["ord"] = None
return field
def add_field(self, m: NotetypeDict, field: FieldDict) -> None:
"Modifies schema."
m["flds"].append(field)
def remove_field(self, m: NotetypeDict, field: FieldDict) -> None:
"Modifies schema."
m["flds"].remove(field)
def reposition_field(self, m: NotetypeDict, field: FieldDict, idx: int) -> None:
"Modifies schema."
oldidx = m["flds"].index(field)
if oldidx == idx:
return
m["flds"].remove(field)
m["flds"].insert(idx, field)
def rename_field(self, m: NotetypeDict, field: FieldDict, new_name: str) -> None:
assert field in m["flds"]
field["name"] = new_name
def set_sort_index(self, nt: NotetypeDict, idx: int) -> None:
"Modifies schema."
assert 0 <= idx < len(nt["flds"])
nt["sortf"] = idx
# legacy
newField = new_field
def addField(self, m: NotetypeDict, field: FieldDict) -> None:
self.add_field(m, field)
if m["id"]:
self.save(m)
def remField(self, m: NotetypeDict, field: FieldDict) -> None:
self.remove_field(m, field)
self.save(m)
def moveField(self, m: NotetypeDict, field: FieldDict, idx: int) -> None:
self.reposition_field(m, field, idx)
self.save(m)
def renameField(self, m: NotetypeDict, field: FieldDict, newName: str) -> None:
self.rename_field(m, field, newName)
self.save(m)
# Adding & changing templates
##################################################
def new_template(self, name: str) -> TemplateDict:
nt = from_json_bytes(
self.col._backend.get_stock_notetype_legacy(StockNotetypeKind.BASIC)
)
template = nt["tmpls"][0]
template["name"] = name
template["qfmt"] = ""
template["afmt"] = ""
template["ord"] = None
return template
def add_template(self, m: NotetypeDict, template: TemplateDict) -> None:
"Modifies schema."
m["tmpls"].append(template)
def remove_template(self, m: NotetypeDict, template: TemplateDict) -> None:
"Modifies schema."
assert len(m["tmpls"]) > 1
m["tmpls"].remove(template)
def reposition_template(
self, m: NotetypeDict, template: TemplateDict, idx: int
) -> None:
"Modifies schema."
oldidx = m["tmpls"].index(template)
if oldidx == idx:
return
m["tmpls"].remove(template)
m["tmpls"].insert(idx, template)
# legacy
newTemplate = new_template
def addTemplate(self, m: NotetypeDict, template: TemplateDict) -> None:
self.add_template(m, template)
if m["id"]:
self.save(m)
def remTemplate(self, m: NotetypeDict, template: TemplateDict) -> None:
self.remove_template(m, template)
self.save(m)
def moveTemplate(self, m: NotetypeDict, template: TemplateDict, idx: int) -> None:
self.reposition_template(m, template, idx)
self.save(m)
def template_use_count(self, ntid: NotetypeId, ord: int) -> int:
return self.col.db.scalar(
"""
select count() from cards, notes where cards.nid = notes.id
and notes.mid = ? and cards.ord = ?""",
ntid,
ord,
)
# Changing notetypes of notes
##########################################################################
def get_single_notetype_of_notes(
self, note_ids: Sequence[anki.notes.NoteId]
) -> NotetypeId:
return NotetypeId(
self.col._backend.get_single_notetype_of_notes(note_ids=note_ids)
)
def change_notetype_info(
self, *, old_notetype_id: NotetypeId, new_notetype_id: NotetypeId
) -> bytes:
return self.col._backend.get_change_notetype_info(
old_notetype_id=old_notetype_id, new_notetype_id=new_notetype_id
)
def change_notetype_of_notes(self, input: ChangeNotetypeIn) -> OpChanges:
"""Assign a new notetype, optionally altering field/template order.
To get defaults, use
input = ChangeNotetypeIn()
input.ParseFromString(col.models.change_notetype_info(...))
input.note_ids.extend([...])
The new_fields and new_templates lists are relative to the new notetype's
field/template count. Each value represents the index in the previous
notetype. -1 indicates the original value will be discarded.
"""
return self.col._backend.change_notetype(input)
# legacy API - used by unit tests and add-ons
def change(
self,
m: NotetypeDict,
nids: List[anki.notes.NoteId],
newModel: NotetypeDict,
fmap: Dict[int, Optional[int]],
cmap: Optional[Dict[int, Optional[int]]],
) -> None:
# - maps are ord->ord, and there should not be duplicate targets
self.col.modSchema(check=True)
assert fmap
field_map = self._convert_legacy_map(fmap, len(newModel["flds"]))
if not cmap or newModel["type"] == MODEL_CLOZE or m["type"] == MODEL_CLOZE:
template_map = []
else:
template_map = self._convert_legacy_map(cmap, len(newModel["tmpls"]))
self.col._backend.change_notetype(
ChangeNotetypeIn(
note_ids=nids,
new_fields=field_map,
new_templates=template_map,
old_notetype_id=m["id"],
new_notetype_id=newModel["id"],
current_schema=self.col.db.scalar("select scm from col"),
)
)
def _convert_legacy_map(
self, old_to_new: Dict[int, Optional[int]], new_count: int
) -> List[int]:
"Convert old->new map to list of old indexes"
new_to_old = {v: k for k, v in old_to_new.items() if v is not None}
out = []
for idx in range(new_count):
try:
val = new_to_old[idx]
except KeyError:
val = -1
out.append(val)
return out
# Schema hash
##########################################################################
def scmhash(self, m: NotetypeDict) -> str:
"Return a hash of the schema, to see if models are compatible."
s = ""
for f in m["flds"]:
s += f["name"]
for t in m["tmpls"]:
s += t["name"]
return checksum(s)
# Cloze
##########################################################################
def _availClozeOrds(
self, m: NotetypeDict, flds: str, allowEmpty: bool = True
) -> List[int]:
print("_availClozeOrds() is deprecated; use note.cloze_numbers_in_fields()")
note = _pb.Note(fields=[flds])
return list(self.col._backend.cloze_numbers_in_note(note))