Settings

For clients that wish to temporarily alter any of the previously described behaviors, a handful of settings can be controlled at the module level. All values default to True unless otherwise noted.

HIDE_TRACEBACK_IN_HOOKS

When an exception occurs in patched method, this traceback is hidden by default for pytest. If this information is required to debug a complex issue enable it as follows:

import datafiles

datafiles.settings.HIDE_TRACEBACK_IN_HOOKS = False

HOOKS_ENABLED

When running unit tests for a client using datafiles, it can be helpful to disable automatic loading/saving of models for performance and to avoid writing files to disk:

import datafiles

def pytest_runtest_setup(item):
    """Disable file storage during unit tests."""
    datafiles.settings.HOOKS_ENABLED = False

INDENT_YAML_BLOCKS

In datafiles >= 0.4, YAML blocks are now indented by default, like so:

items:
  - 1
  - 2
  - 3

To make it easier to upgrade to this version, a client can disable this functionality:

import datafiles

datafiles.settings.INDENT_YAML_BLOCKS = False

to produce YAML like:

items:
- 4
- 5
- 6

MINIMIZE_LIST_DIFFS

When serializing lists, datafiles intentionally deviates from the semantic representation of an empty list to optimize for the use case of storing YAML files in version control.

By storing any empty list of items as:

items:
  -

adding or removing an item always results in a one-line change. Where as adding items to items: [] produces a noisier diff and requires knowledge of the YAML specification to edit files by hand.

To disable this behavior:

import datafiles

datafiles.settings.MINIMIZE_LIST_DIFFS = False

YAML_LIBRARY

This setting controls the underlying YAML library used to read and write files. The following options are available:

• 'ruamel.yaml' (default)
• 'PyYAML'

WRITE_DELAY

One some file systems, the modification time of a file (st_mtime) is unchanged if a file is read immediately after writing. This may cause intermittent issues if your use case involves rapidly changing files.

To compensate for this, a short delay can be inserted after datafiles writes to the file system:

import datafiles

datafiles.settings.WRITE_DELAY = 0.01  # seconds