DbMod
Database enabled modules for ruby.
Description
The db_mod
gem is a simple framework that helps you organise your
database access functions into modular libraries that can be included
in your projects to give them selective access to facets of your data.
For the moment db_mod
only supports PostgreSQL databases via the
pg
gem. This gem is still in the early stages of development and no
guarantees will be made about backwards compatibility until v0.1.0.
Issues, feature or pull requests, comments and feedback all welcomed.
Installation
From the command line:
gem install db_mod
Or in your Gemfile
:
gem 'db_mod'
And then in your script:
require 'db_mod'
Usage
The database connection
At its most basic, db_mod
provides db_connect
, query
, and
transaction
:
require 'db_mod'
module MyFunctions
include DbMod
def get_stuff
query 'SELECT * FROM stuff'
end
def do_complicated_thing(input)
transaction do # calls BEGIN
query 'INSERT ...'
query 'UPDATE ...'
query 'DELETE ...'
output = query 'SELECT ...'
fail if output.empty? # calls ROLLBACK
end # calls COMMIT
end
end
include MyFunctions
db_connect(
db: 'mydb',
host: 'localhost', # defaults to local socket
port: 5432, # this is the default
user: 'myuser', # default is ENV['USER']
pass: 'password' # attempts trusted connection by default
)
get_stuff.each do |thing|
thing['id'] # => '1'
# ...
end
Module instances: DbMod.create
Each module also comes with its own create
function,
which instantiates an object exposing all of the module's functions.
# Standard connection options can be used.
# db_connect will be called.
db = MyFunctions.create db: 'mydb'
# Or an existing connection object can be passed
db = MyFunctions.create PGconn.connect # ...
db.get_stuff
The connection object: @conn
The connection created by db_connect
or create
will be stored
in the instance variable @conn
. This instance variable may be
set explicitly instead of calling db_connect
, allowing arbitrary
sharing of database connections between modules and objects.
See notes below on using this technique with def_prepared
.
Module heirarchies
By including multiple modules into the same class or object, they
will all use the same connection object supplied by db_connect
.
This connection object is stored in the instance variable @conn
and can be supplied manually:
module DbAccess
# includes DbMod and defines do_things
include Db::Things
# includes DbMod and defines do_stuff
include Db::Stuff
def things_n_stuff
transaction do
do_things
do_stuff
end
end
end
db = DbAccess.create db: 'mydb'
db.things_n_stuff
Declaring SQL statements
Prepared statement methods: DbMod.def_prepared
Modules which include DbMod
can declare prepared statements
using the module function def_prepared
. These statements will
be prepared on the connection when db_connect
is called.
A method will be defined in the module with the same name as
the prepared statement, provided as a convenience for executing
the statement:
module Db
module Things
# Statements can use named parameters:
def_prepared :foo, <<-SQL
SELECT *
FROM foo
WHERE id = $id
AND b > $minimum_value
AND c > $minimum_value
SQL
end
module Stuff
# Indexed parameters also work:
def_prepared :bar, <<-SQL
INSERT INTO bar
(a, b, c)
VALUES
($1, $2, $1)
SQL
end
module ComplicatedStuff
# Statements on included modules will also
# be executed when db_connect is called.
include Things
include Stuff
def complicated_thing!(id, min)
transaction do
foo(id: id, minimum_value: min).each do |thing|
bar(thing['a'], thing['b'])
end
end
end
end
end
include Db::ComplicatedStuff
db_connect db: 'mydb'
complicated_thing!(1, 2)
Note that if an existing connection is supplied to create
or @conn
then declared statements will not be automatically prepared. In this
case the module function prepare_all_statements(conn)
can be used
to prepare all statements declared in the module or any included
modules on the given connection.
db = Db::ComplicatedStuff.create my_conn
Db::ComplicatedStuff.prepare_all_statements my_conn
Saved statement methods: DbMod.def_statement
def_statement
works in just the same way as def_prepared
, except that
the SQL queries are saved in memory rather than being sent to the database
at connection time. This is useful for queries that will only be run once
or twice during a program's execution.
Configuring defined statements
db_mod
contains a simple framework for extending these statement methods
and prepared methods with additional result processing. A block can be
passed to +def_prepared+ and +def_statement+ definitions, where a basic
DSL is made available for additional method configuration.
JSON and CSV formatting
Statement and prepared methods can be configured on declaration by using
as(:csv)
and as(:json)
, which will convert the result set to a string
formatted as either a CSV document or an array of JSON objects, respectively.
# db_mod makes no attempt to load these
require 'csv'
require 'json'
module Reports
include DbMod
def_prepared(:foo, 'SELECT a, b FROM foo WHERE bar_id = $id') { as(:csv) }
def_statement(:bar, 'SElECT c, d FROM bar WHERE foo_id = $1') { as(:json) }
end
include Reports
db_connect db: 'testdb'
foo(id: 1) # => "a,b\n1,2\n3,4\n..."
bar(2) # => '[{"c":"5","d":"6"},...]'
Queries returning one row, column or value
To save a lot of repetetive unboxing of query results, methods that return
only one row, or rows with only one column, or only one row with a single
value, can be marked as such using the single
extension.
module Getters
include DbMod
def_prepared(:user, 'SELECT * FROM user WHERE id = $1') { single(:row) }
def_prepared(:name, 'SELECT name FROM user WHERE id = $1') { single(:value) }
def_statement(:ids, 'SELECT id FROM user') { single(:column) }
end
# ...
user(1) # => { "id" => "1", "name" => "username" }
name(1) # => "username"
ids # => ['1', '2', '3', ...]
When no results are returned, :column
returns []
while :row
and
:value
will return nil
. To raise an exception instead of returning
nil
, use :row!
and :value!
instead.
def_statement(:a, 'SELECT 1 WHERE true = false') { single(:value) }
def_statement(:b, 'SELECT 1 WHERE true = false') { single(:value!) }
a # => nil
b # => fail
Default parameter values
Arbitrary default parameter values can be supplied using defaults
.
For methods with named parameters:
def_statement(:a, 'SELECT * FROM foo WHERE id = $id AND y > $min') do
defaults min: 10
end
# ...
a(id: 1) # === a(id: 1, min: 10)
For methods with fixed parameters:
def_statement(:a, %(
SELECT *
FROM foo
WHERE x = $1
AND y < $2
AND z > $3
)) { defaults 4, 5, 6 }
# ...
a # === a(4, 5, 6)
a(1) # === a(1, 5, 6)
a(1, 2) # === a(1, 2, 6)
a(1, 2, 3) # === a(1, 2, 3)
A proc may be given as the default value for any parameter. The proc should accept one parameter, which will be the argument list/hash (depending on if the statement uses numbered or named arguments) and should return a single value to be used for this execution of the query.
def_prepared(:default_min, %(
SELECT default_min
FROM defaults
WHERE foo_id = $1
)) { single(:value) }
def_prepared(:report, 'SELECT * FROM a WHERE b = $c AND d > $e') do
defaults min: ->(args) { default_min(args[:c]) }
as :json
end
The above example shows how the proc will be executed using the instance
of the DbMod
module as scope, so statement, prepared, or other methods
may be accessed.
Note that the argument list will be partially constructed at the time it
is received by the proc; other default values may or may not yet have been
populated. Defaults will be populated in the order that they are declared
using defaults
.
Custom method return values
Besides the built-in result transformations provided by as
and single
,
db_mod
also allows arbitrary control over the return value of statement
and prepared methods using a block provided via returning
:
def_prepared(:a, 'SELECT name, sound FROM animals') do
# Block parameter is the SQL result set
returning do |animals|
animals.map do |animal|
"the #{animal['name']} goes #{animal['sound']}"
end.join ' and '
end
end
def_prepared(:important_report!, 'SELECT address FROM email WHERE id = $1') do
# 'single' and 'as' will transform the result set
# before it is passed to the block
single(:value)
returning { |email| send_email(email, a) }
end
# Block has instance-level scope
def send_email(address, body)
# ...
end
# ...
important_report!(123)
# === send_email('ex@mp.le', 'the sheep goes baa and the cow goes moo')
Default configuration settings
To save typing, modules may declare a default_method_settings
block that will
be applied to all following def_statement
and def_prepared
definitions. The
dsl used is the same as for individual method configuration blocks.
require 'csv'
module CsvReports
include DbMod
default_method_settings do
as(:csv).returning { |csv| send_report(csv) }
end
def send_report(csv)
# ...
end
def_prepared(:report_a, 'SELECT * FROM report_a WHERE user_id = $1')
def_prepared(:report_b, 'SELECT * FROM report_b WHERE user_id = $1')
def_prepared(:report_c, 'SELECT * FROM report_c WHERE user_id = $1')
end
module ReportEmailer
include CsvReports
default_method_settings do
single(:column)
returning do |ids|
ids.each { |id| send_all_reports_for(id) }
end
end
def send_all_reports_for(user)
report_a(user)
report_b(user)
report_c(user)
end
def_statement(:send_all_reports, 'SELECT id FROM user')
def_statement(:send_priority_reports, 'SELECT id FROM user WHERE priority')
# individual settings may be overridden
def_statement(:send_all_a_reports, 'SELECT id FROM user') do
returning { |ids| ids.each { |id| report_a(id) } }
end
def_statement(:send_reports, 'SELECT id FROM user WHERE name = $name') do
single(:value).returning { |id| send_all_reports_for(id) }
end
end
Defaults don't cascade automatically from one module to another module that has included it. However the following sorts of things work if you need more flexibility in reusing default settings:
BASE_SETTINGS = ->() { single(:row).as(:json) }
module A
include DbMod
# Use base settings with overrides in block
default_method_settings(BASE_SETTINGS) { as(:csv) }
end
A.default_method_settings # => { as: :csv, single: :row }
module B
include A
# Inherit settings from A, overrides as named args
default_method_settings(A.default_method_settings, as: :csv)
end