tests/test_units.py

import io
import os
import importlib

def test_chunk_text_basic():
    app = importlib.import_module("app")
    text = "a" * 2500
    chunks = app.chunk_text(text, max_chars=1000)
    assert len(chunks) == 3
    assert "".join(chunks) == text

def test_build_prompt_includes_contexts():
    app = importlib.import_module("app")
    ctx = [{"id": 1, "title": "t", "content": "c1"}, {"id": 2, "title": "t2", "content": "c2"}]
    p = app.build_prompt("q", ctx)
    assert "c1" in p and "c2" in p and "q" in p

def test_extract_pdf_text_handles_bytes():
    app = importlib.import_module("app")
    # Minimal PDF content (may not contain extractable text, should not raise)
    minimal_pdf = b"%PDF-1.4\n1 0 obj\n<<>>\nendobj\nxref\n0 1\n0000000000 65535 f \ntrailer\n<<>>\nstartxref\n0\n%%EOF\n"
    text = app.extract_pdf_text(minimal_pdf)
    assert isinstance(text, str)