bitsmiths-audit 2.2.2

The Bitsmiths Audit Product

BITSMITHS AUDIT

This is the second tier base/core Bitsmiths python library. The purpose of this product is provide field level row auditing for any tables or objects within a product.

This auditing product is used by our own open-source products as well as commercial products which require detailed auditing. The nice thing is, that if you are using any Bitsmiths products down the line, then you already have a working auditing solution within your project that is ready to go.

Note that we package the generated Mettle code in this package, which means that the Mettle version required in this module is important.

Tables & Setup

Audit is designed to use a relational database with a schema/namespace of audit. You will need to create this database schema yourself. It also requires four relational database tables.

The package provides the SQL code to create these tables. There are two ways to access the table creation SQL.

1. You can run bs-audit -g postgresql at the command line, this will print all the SQL to stdout.
2. You can import the bs_audit module and get the SQL as shown below:

import bs_audit

print(bs_audit.get_table_sql('postgresql'))

Note! Currently only postgresql SQL is added to the package. If you want us to support another database let us know and we will put it into the development pipeline.

Library Objects

These are the objects in the audit product with a short description of their intended purpose.

Audit

The Audit class (as well as it's AuditAsync counter part for async code) is the method in which one uses the audit product. Each time you make an insert, update, or delete to a table, you can choose to audit that table using this class.

There are a couple of presumptions that the Audit object makes:

• That you are auditing an object that inherits from mettle.io.ISerializable. This is because audit will use the _deserialize() method to read the objects fields.
• That the name of the auditing object has been entered into the audit.cfg table.
• That you are also using the auditing object from within some kind of AutoTransaction - see bitsmiths-lib.AutoTransaction.

Below is an example of intended use:

• We assume the Pod object contains an instance of the Audit object.
• We assume the Pod object contains an instance of a database connection.
• We assume the brec is a mettle table record of some kind.
• We assume the user identifier has been configured in the pod, and the Audit.cfg table has been configured to use the modified_by column.

import copy
from mettle.lib import AutoTransaction

def save_bank_record(self, brec: tBank):
	with AutoTransaction(self.pod.audit, self.pod) as at:
		bank_tbl = self.pod.dao.dBank(self.pod.dbcon)

		brec.modified_by = self.pod.usr_id

		if bank_tbl.try_select_one_deft(brec.id):
			orig = copy.copy(bank_tbl.rec)
			bank_tbl.update(brec)
			self.pod.audit.diff(orig, bank_tbl.rec)  # audits the difference between the original and new record
		else:
			bank_tbl.insert(brec)
			self.pod.audit.diff(None, bank_tbl.rec)  # audits a brand new record.

		at.commit()

TableWatch

This is a simply object that is designed to check on an interval for table changes. You may want to implement to update internal caches or something if a table has row changes. Assumes that you are using the Audit objects whenever you do insert, update, or deletes to table rows else it won't know.

Change History

2.2.2

Type	Description
Fix	Fixed a sql error on the postgres trigger script.

2.2.1

Type	Description
New	Upgraded to use bitsmiths-mettle version 2.2.3

2.2.0

Type	Description
Breaking	Upgraded to use bitsmiths-mettle version 2.2.2

2.1.6

Type	Description
New	Implemented auditing auto triggers
New	Using the new python fetch method from bitsmiths-mettle (2.1.14)
Breaking	Table changes and new triggers required on the database (see version_2_1_6_upgrade.postgresql.sql in the repo)
New	Implemented a new function to read audit trails easier.

Example of audit trail query/function:

select * from audit.show_trail('{"config"}', null, '{"value"}', '2022-05-01', null, '{"U"}');

2.1.5

Type	Description
New	Implemented the dataclass feature from the latest mettle (2.1.13) version.

2.1.4

Type	Description
Bug	Fixed a regression bug in base_audit.search method that was raising SQL exceptions.

2.1.3

Type	Description
Bug	Forced the truncation of long column values when doing auditing.
Bug	Auditing of column values now works with derived dictionary and list types.

2.1.0

Type	Description
Breaking	Rename from triceraclops to audit
Breaking	Moved the dao import method from util.py to the init.py.
New	Pulled in the mettle via PYPI, updated project to cater for new changes in mettle.
New	You can now read the database creation from the package, also added an entry point for this.