MAINT: Factors-out asset db schema

This commit is contained in:
jfkirk
2015-12-29 14:16:02 -05:00
parent 0b588219af
commit 004c8dc141
4 changed files with 169 additions and 163 deletions
+2
View File
@@ -50,6 +50,8 @@ from zipline.assets.futures import (
from zipline.assets.asset_writer import (
check_version_info,
write_version_info,
)
from zipline.assets.asset_db_schema import (
ASSET_DB_VERSION,
_version_table_schema,
)
+154
View File
@@ -0,0 +1,154 @@
import sqlalchemy as sa
def generate_asset_db_metadata(bind=None):
metadata = sa.MetaData(bind=bind)
_version_table_schema(metadata)
_equities_table_schema(metadata)
_futures_exchanges_schema(metadata)
_futures_root_symbols_schema(metadata)
_futures_contracts_schema(metadata)
_asset_router_schema(metadata)
return metadata
# Define a version number for the database generated by these writers
# Increment this version number any time a breaking change is made to the
# schema and readers of the database
ASSET_DB_VERSION = 0
# A list of the names of all tables in the assets db
asset_db_table_names = ['version_info', 'equities', 'futures_exchanges',
'futures_root_symbols', 'futures_contracts',
'asset_router']
def _equities_table_schema(metadata):
return sa.Table(
'equities',
metadata,
sa.Column(
'sid',
sa.Integer,
unique=True,
nullable=False,
primary_key=True,
),
sa.Column('symbol', sa.Text),
sa.Column('company_symbol', sa.Text, index=True),
sa.Column('share_class_symbol', sa.Text),
sa.Column('fuzzy_symbol', sa.Text, index=True),
sa.Column('asset_name', sa.Text),
sa.Column('start_date', sa.Integer, default=0, nullable=False),
sa.Column('end_date', sa.Integer, nullable=False),
sa.Column('first_traded', sa.Integer, nullable=False),
sa.Column('exchange', sa.Text),
)
def _futures_exchanges_schema(metadata):
return sa.Table(
'futures_exchanges',
metadata,
sa.Column(
'exchange',
sa.Text,
unique=True,
nullable=False,
primary_key=True,
),
sa.Column('timezone', sa.Text),
)
def _futures_root_symbols_schema(metadata):
return sa.Table(
'futures_root_symbols',
metadata,
sa.Column(
'root_symbol',
sa.Text,
unique=True,
nullable=False,
primary_key=True,
),
sa.Column('root_symbol_id', sa.Integer),
sa.Column('sector', sa.Text),
sa.Column('description', sa.Text),
sa.Column(
'exchange',
sa.Text,
sa.ForeignKey('futures_exchanges.exchange'),
),
)
def _futures_contracts_schema(metadata):
return sa.Table(
'futures_contracts',
metadata,
sa.Column(
'sid',
sa.Integer,
unique=True,
nullable=False,
primary_key=True,
),
sa.Column('symbol', sa.Text, unique=True, index=True),
sa.Column(
'root_symbol',
sa.Text,
sa.ForeignKey('futures_root_symbols.root_symbol'),
index=True
),
sa.Column('asset_name', sa.Text),
sa.Column('start_date', sa.Integer, default=0, nullable=False),
sa.Column('end_date', sa.Integer, nullable=False),
sa.Column('first_traded', sa.Integer, nullable=False),
sa.Column(
'exchange',
sa.Text,
sa.ForeignKey('futures_exchanges.exchange'),
),
sa.Column('notice_date', sa.Integer, nullable=False),
sa.Column('expiration_date', sa.Integer, nullable=False),
sa.Column('auto_close_date', sa.Integer, nullable=False),
sa.Column('contract_multiplier', sa.Float),
)
def _asset_router_schema(metadata):
return sa.Table(
'asset_router',
metadata,
sa.Column(
'sid',
sa.Integer,
unique=True,
nullable=False,
primary_key=True),
sa.Column('asset_type', sa.Text),
)
def _version_table_schema(metadata):
return sa.Table(
'version_info',
metadata,
sa.Column(
'id',
sa.Integer,
unique=True,
nullable=False,
primary_key=True,
),
sa.Column(
'version',
sa.Integer,
unique=True,
nullable=False,
),
# This constraint ensures a single entry in this table
sa.CheckConstraint('id <= 1'),
)
+13 -158
View File
@@ -27,6 +27,11 @@ import sqlalchemy as sa
from zipline.errors import SidAssignmentError, AssetDBVersionError
from zipline.assets._assets import Asset
from zipline.assets.asset_db_schema import (
generate_asset_db_metadata,
asset_db_table_names,
ASSET_DB_VERSION,
)
SQLITE_MAX_VARIABLE_NUMBER = 999
@@ -164,12 +169,6 @@ def _generate_output_dataframe(data_subset, defaults):
return output
# Define a version number for the database generated by these writers
# Increment this version number any time a breaking change is made to the
# schema and readers of the database
ASSET_DB_VERSION = 0
def check_version_info(version_table, expected_version):
"""
Checks for a version value in the version table.
@@ -215,143 +214,6 @@ def write_version_info(version_table, version_value):
sa.insert(version_table, values={'version': version_value}).execute()
# A list of the names of all tables in the assets db
asset_db_table_names = ['version_info', 'equities', 'futures_exchanges',
'futures_root_symbols', 'futures_contracts',
'asset_router']
def _equities_table_schema(metadata):
return sa.Table(
'equities',
metadata,
sa.Column(
'sid',
sa.Integer,
unique=True,
nullable=False,
primary_key=True,
),
sa.Column('symbol', sa.Text),
sa.Column('company_symbol', sa.Text, index=True),
sa.Column('share_class_symbol', sa.Text),
sa.Column('fuzzy_symbol', sa.Text, index=True),
sa.Column('asset_name', sa.Text),
sa.Column('start_date', sa.Integer, default=0, nullable=False),
sa.Column('end_date', sa.Integer, nullable=False),
sa.Column('first_traded', sa.Integer, nullable=False),
sa.Column('exchange', sa.Text),
)
def _futures_exchanges_schema(metadata):
return sa.Table(
'futures_exchanges',
metadata,
sa.Column(
'exchange',
sa.Text,
unique=True,
nullable=False,
primary_key=True,
),
sa.Column('timezone', sa.Text),
)
def _futures_root_symbols_schema(metadata, futures_exchanges):
return sa.Table(
'futures_root_symbols',
metadata,
sa.Column(
'root_symbol',
sa.Text,
unique=True,
nullable=False,
primary_key=True,
),
sa.Column('root_symbol_id', sa.Integer),
sa.Column('sector', sa.Text),
sa.Column('description', sa.Text),
sa.Column(
'exchange',
sa.Text,
sa.ForeignKey(futures_exchanges.c.exchange),
),
)
def _futures_contracts_schema(metadata, futures_root_symbols,
futures_exchanges):
return sa.Table(
'futures_contracts',
metadata,
sa.Column(
'sid',
sa.Integer,
unique=True,
nullable=False,
primary_key=True,
),
sa.Column('symbol', sa.Text, unique=True, index=True),
sa.Column(
'root_symbol',
sa.Text,
sa.ForeignKey(futures_root_symbols.c.root_symbol),
index=True
),
sa.Column('asset_name', sa.Text),
sa.Column('start_date', sa.Integer, default=0, nullable=False),
sa.Column('end_date', sa.Integer, nullable=False),
sa.Column('first_traded', sa.Integer, nullable=False),
sa.Column(
'exchange',
sa.Text,
sa.ForeignKey(futures_exchanges.c.exchange),
),
sa.Column('notice_date', sa.Integer, nullable=False),
sa.Column('expiration_date', sa.Integer, nullable=False),
sa.Column('auto_close_date', sa.Integer, nullable=False),
sa.Column('contract_multiplier', sa.Float),
)
def _asset_router_schema(metadata):
return sa.Table(
'asset_router',
metadata,
sa.Column(
'sid',
sa.Integer,
unique=True,
nullable=False,
primary_key=True),
sa.Column('asset_type', sa.Text),
)
def _version_table_schema(metadata):
return sa.Table(
'version_info',
metadata,
sa.Column(
'id',
sa.Integer,
unique=True,
nullable=False,
primary_key=True,
),
sa.Column(
'version',
sa.Integer,
unique=True,
nullable=False,
),
# This constraint ensures a single entry in this table
sa.CheckConstraint('id <= 1'),
)
class AssetDBWriter(with_metaclass(ABCMeta)):
"""
Class used to write arbitrary data to SQLite database.
@@ -480,23 +342,16 @@ class AssetDBWriter(with_metaclass(ABCMeta)):
constraints : bool, optional
If True, create SQL ForeignKey and PrimaryKey constraints.
"""
metadata = sa.MetaData(bind=engine)
tables_already_exist = self.check_for_tables(engine)
metadata = generate_asset_db_metadata(bind=engine)
tables_already_exist = self.check_for_tables(engine=engine)
self.version_info = metadata.tables['version_info']
self.equities = metadata.tables['equities']
self.futures_exchanges = metadata.tables['futures_exchanges']
self.futures_root_symbols = metadata.tables['futures_root_symbols']
self.futures_contracts = metadata.tables['futures_contracts']
self.asset_router = metadata.tables['asset_router']
self.equities = _equities_table_schema(metadata)
self.futures_exchanges = _futures_exchanges_schema(metadata)
self.futures_root_symbols = _futures_root_symbols_schema(
metadata=metadata,
futures_exchanges=self.futures_exchanges,
)
self.futures_contracts = _futures_contracts_schema(
metadata=metadata,
futures_root_symbols=self.futures_root_symbols,
futures_exchanges=self.futures_exchanges,
)
self.asset_router = _asset_router_schema(metadata)
self.version_info = _version_table_schema(metadata)
# Create the SQL tables if they do not already exist.
metadata.create_all(checkfirst=True)
-5
View File
@@ -38,8 +38,6 @@ from zipline.assets import (
)
from zipline.assets.asset_writer import (
split_delimited_symbol,
check_version_info,
ASSET_DB_VERSION,
asset_db_table_names,
SQLITE_MAX_VARIABLE_NUMBER
)
@@ -106,9 +104,6 @@ class AssetFinder(object):
for table_name in asset_db_table_names:
setattr(self, table_name, metadata.tables[table_name])
# Check the version info of the db for compatibility
check_version_info(self.version_info, ASSET_DB_VERSION)
# Cache for lookup of assets by sid, the objects in the asset lookup
# may be shared with the results from equity and future lookup caches.
#