119 lines
4 KiB
Python
119 lines
4 KiB
Python
"""
|
|
Tests for document processing and ingestion.
|
|
"""
|
|
|
|
import pytest
|
|
from pathlib import Path
|
|
from unittest.mock import patch, MagicMock
|
|
|
|
|
|
@pytest.mark.integration
|
|
class TestDocumentIngestion:
|
|
"""Test document ingestion pipeline."""
|
|
|
|
def test_text_file_ingestion(self, sample_kb, clean_db, sample_text_file):
|
|
"""Test ingesting a text file (requires llama_index)."""
|
|
pytest.importorskip("llama_index", reason="llama_index required for ingestion")
|
|
|
|
from app.core.ingestion import ingest_file
|
|
|
|
# Mock embedding generation
|
|
with patch('app.core.ingestion.OllamaEmbedding') as mock_embed:
|
|
mock_instance = MagicMock()
|
|
mock_instance.get_text_embedding.return_value = [0.1] * 768
|
|
mock_embed.return_value = mock_instance
|
|
|
|
result = ingest_file(
|
|
file_path=str(sample_text_file),
|
|
collection_name=sample_kb["name"],
|
|
filename=sample_text_file.name
|
|
)
|
|
|
|
assert result["status"] == "success"
|
|
assert result["chunks"] > 0
|
|
|
|
def test_pdf_file_ingestion(self, sample_kb, clean_db, sample_pdf_file):
|
|
"""Test ingesting a PDF file (requires llama_index)."""
|
|
pytest.importorskip("llama_index", reason="llama_index required for ingestion")
|
|
|
|
from app.core.ingestion import ingest_file
|
|
|
|
with patch('app.core.ingestion.OllamaEmbedding') as mock_embed:
|
|
mock_instance = MagicMock()
|
|
mock_instance.get_text_embedding.return_value = [0.1] * 768
|
|
mock_embed.return_value = mock_instance
|
|
|
|
result = ingest_file(
|
|
file_path=str(sample_pdf_file),
|
|
collection_name=sample_kb["name"],
|
|
filename=sample_pdf_file.name
|
|
)
|
|
|
|
assert result["status"] == "success"
|
|
assert result["chunks"] > 0
|
|
|
|
def test_markdown_file_ingestion(self, sample_kb, clean_db, sample_markdown_file):
|
|
"""Test ingesting a Markdown file (requires llama_index)."""
|
|
pytest.importorskip("llama_index", reason="llama_index required for ingestion")
|
|
|
|
from app.core.ingestion import ingest_file
|
|
|
|
with patch('app.core.ingestion.OllamaEmbedding') as mock_embed:
|
|
mock_instance = MagicMock()
|
|
mock_instance.get_text_embedding.return_value = [0.1] * 768
|
|
mock_embed.return_value = mock_instance
|
|
|
|
result = ingest_file(
|
|
file_path=str(sample_markdown_file),
|
|
collection_name=sample_kb["name"],
|
|
filename=sample_markdown_file.name
|
|
)
|
|
|
|
assert result["status"] == "success"
|
|
assert result["chunks"] > 0
|
|
|
|
|
|
@pytest.mark.unit
|
|
class TestDocumentChunking:
|
|
"""Test document chunking strategies."""
|
|
|
|
def test_fixed_size_chunking(self):
|
|
"""Test fixed-size chunking."""
|
|
text = "This is a test. " * 100 # Long text
|
|
chunk_size = 100
|
|
overlap = 20
|
|
|
|
chunks = []
|
|
start = 0
|
|
while start < len(text):
|
|
end = start + chunk_size
|
|
chunks.append(text[start:end])
|
|
start = end - overlap
|
|
|
|
assert len(chunks) > 1
|
|
assert all(len(chunk) <= chunk_size for chunk in chunks)
|
|
|
|
def test_sentence_chunking(self):
|
|
"""Test sentence-based chunking."""
|
|
text = "First sentence. Second sentence. Third sentence. Fourth sentence."
|
|
sentences = text.split('. ')
|
|
|
|
assert len(sentences) == 4
|
|
assert all(len(s) > 0 for s in sentences)
|
|
|
|
|
|
@pytest.mark.unit
|
|
class TestFileTypeDetection:
|
|
"""Test file type detection."""
|
|
|
|
def test_detect_text_file(self, sample_text_file):
|
|
"""Test detecting text file type."""
|
|
assert sample_text_file.suffix == ".txt"
|
|
|
|
def test_detect_pdf_file(self, sample_pdf_file):
|
|
"""Test detecting PDF file type."""
|
|
assert sample_pdf_file.suffix == ".pdf"
|
|
|
|
def test_detect_markdown_file(self, sample_markdown_file):
|
|
"""Test detecting Markdown file type."""
|
|
assert sample_markdown_file.suffix == ".md"
|