Testing Helpers

This page describes docculus.testing.fixtures, which provides pytest markers to skip tests based on whether an optional dependency is installed.

Prerequisites: You'll need to know a bit of Python and pytest. pytest must be installed to use these fixtures; docculus.testing is not a runtime dependency of docculus itself.

Overview

Several docculus features depend on optional packages: docculus.store's ready-to-use document stores require persista, docculus.utils.fake requires faker, and docculus.analysis's report-printing functions require rich. docculus.testing.fixtures exposes, for each of these, a pair of pytest markers:

• <dep>_available: skip the test unless <dep> is installed
• <dep>_not_available: skip the test if <dep> is installed

Markers are currently provided for faker and persista: faker_available/faker_not_available and persista_available/persista_not_available.

Skipping Tests Based on Optional Dependencies

Use <dep>_available to only run a test when the corresponding package is installed, for example a test that exercises InMemoryDocumentStore (which requires persista):

from docculus.store import InMemoryDocumentStore
from docculus.testing.fixtures import persista_available


@persista_available
def test_in_memory_document_store_set_get():
    with InMemoryDocumentStore() as store:
        store.set_many([...])

Use <dep>_not_available for the opposite case, e.g. verifying that a helpful error is raised when a required optional dependency is missing:

import pytest

from docculus.testing.fixtures import persista_not_available


@persista_not_available
def test_in_memory_document_store_requires_persista():
    with pytest.raises(RuntimeError, match="'persista' package is required"):
        from docculus.store import InMemoryDocumentStore

        InMemoryDocumentStore()

Available Markers

Import markers directly from docculus.testing.fixtures, for example:

from docculus.testing.fixtures import faker_available, persista_available

Generating Fake Documents

docculus.utils.fake.generate_fake_documents (requires the faker extra, docculus[faker]) generates synthetic Document objects for tests and examples, each with a unique id ("doc-{i}"), a Faker-generated paragraph as content, and metadata containing a fake author and topic:

from docculus.utils.fake import generate_fake_documents

docs = generate_fake_documents(n=10)

Content is not guaranteed to be unique across documents -- Faker.paragraph() has no built-in uniqueness constraint, so deduplicate the result yourself (see deduplicate_documents) if strict uniqueness matters.

API Reference

See the reference documentation and reference documentation for the full API.