py-tiny-orm - A Minimal SQLite ORM
py-tiny-orm
is a minimal Object-relational mapping system which stores
python classes in SQLite.
Classes in your code extend the base classes provided by this module, and
a backing model is automatically generated. This model can then be applied
to an SQLite cursor, and data read, updated, or searched.
See the quick start exmaple.
(Simple) Tables
"Table" Data Class
A "Table" is a data structure that contains program defined fields.
They MUST inherited from orm.Table
and contain typed properties, so that the
model can be generated.
The following data types can be used:
Python Type | SQLite Type |
---|
int | INTEGER |
str | TEXT |
bytes | BLOB |
float | REAL |
bool | SMALLINT |
Optional[ x ] | x will not have NOT NULL |
Table | INT + FOREIGN KEY |
"TableModel" Model Class
The Model generated for these tables is of the type TableModel
.
You retrieved this from the Table class method model
passing a sqlite cursor.
Table.create_tables
method generates the table in the database.
with sqlite3.connect(":memory:") as conn:
cursor = conn.cursor()
MyJoinTable.create_table(cursor)
model = MyJoinTalbe.model(cursor)
TableModel.all
all(self) -> 'List[ModelledTable]'
Returns all records on the current table.
Note: records will be loaded into memory before being returned,
in order to optimise the number of queries to realted tables.
TableModel.get
get(self, unique_id: int) -> Optional[T]
Gets a record by ID, or None if no record with that ID exists.
TableModel.get_many
get_many(self, *ids: int) -> Dict[int, T]
Gets all records that exist with ID in the supplied list.
Entries in the dict are not generated for records which do not exist.
TableModel.search
search(self, **kwargs: Any) -> List[T]
Gets records for this model which match the given filters.
You can filter using any field in the table, or by a foreign object.
class Foo(Table["Foo"]):
foo_id: int
class Bar(Table["Bar"]
bar_id: int
foo: Foo
name: str
Bar.model(cursor).search(name="Hello")
Bar.model(cursor).search(foo_id=1)
Bar.model(cursor).search(foo=Foo(1))
Bar.model(cursor).search(bar_id=123)
TableModel.store
store(self, record: T) -> bool
Writes a record to the database.
In all cases, this is done as an INSERT OR REPLACE statement.
If the ID field is not set, this may cause the ID of a record to change,
where it is matched via a unique key.
The ID field will be updated with the inserted row's ID.
Join Tables
A Join table represents a many-to-many mapping between two simple Tables.
These tables are referred to as the Left
and Right
tables; operations
are however bi-directional.
"JoinTable" Data Class
The data entity for a 'Join' Table.
A JoinTable has precisely two fields, which are foreign keys to two other
regular Tables, called Left and Right.
class RoleMapping(JoinTable[User, Role]):
user: User
role: Role
Note that the JoinModel does not currently return instances of the
JoinTable, but future features might; instead functions will return
sequences of the "Left" or "Right" Tables, based on the calls made.
"JoinModel" Model Class
As with Table
, the model is retrieved from the class method
model
using an sqlite3 cursor, and a create_tables
method
is available to generate the table in the database.
with sqlite3.connect(":memory:") as conn:
cursor = conn.cursor()
MyJoinTable.create_table(cursor)
model = MyJoinTalbe.model(cursor)
JoinModel.of_left
/ of_right
of_left(self, left: Left) -> List[Right]
of_right(self, right: Right) -> List[Left]
Returns all records which map to a given record.
JoinModel.ids_for_left
/ ids_for_right
ids_for_left(self, left: Left) -> List[int]
ids_for_right(self, left: Right) -> List[int]
Returns all id
s present which map to a given record.
This is the same extracting the IDs from the records returned by
the of_left
/of_right
function, but saves the overhead of doing
any extra foreign key lookups.
JoinModel.from_left
/ from_right
from_left(self, **kwargs: Any) -> List[Right]
from_right(self, **kwargs: Any) -> List[Left]
Returns all unique Right records which map to Left records that match
the given search criteria. No information about which Left they
matched is maintained.
This will have the same result as:
rights = Left.model().search(**kwargs)
lefts = {}
for right in right:
lefts = join_model.of_right(right)
lefts.update({l.left_id: l for l in lefts})
return lefts.values()
but this function will be considerably more efficient.
JoinModel.clear_left
/ clear_right
clear_left(self, left: Left) -> None
clear_right(self, right: Right) -> None
Deletes all records in the join table that feature the given record
JoinModel.store
store(self, left: Left, right: Right) -> bool
Adds a mapping between the supplied Left and Right
No action is taken if this mapping already exists
JoinModel.remove
remove(self, left: Left, right: Right) -> bool
Removes a mapping between the supplied Left and Right.
No action is taken if this mapping does not exist.