mirror of
https://github.com/HKUDS/nanobot.git
synced 2026-04-30 06:45:55 +00:00
47f5795708
ContextBuilder._build_user_content now only handles images (its original responsibility). Document text extraction (PDF, DOCX, XLSX, PPTX) is performed by the new _extract_documents() helper in server.py, called before process_direct(). This keeps the core context builder free of format-specific dependencies and makes the API boundary the single place where uploaded files are pre-processed. Tests updated to reflect the new responsibility boundary. Made-with: Cursor
59 lines
2.2 KiB
Python
59 lines
2.2 KiB
Python
"""Tests for context builder media handling.
|
|
|
|
The ContextBuilder._build_user_content method should ONLY handle images.
|
|
Document text extraction is the responsibility of the API layer.
|
|
"""
|
|
|
|
from __future__ import annotations
|
|
|
|
from pathlib import Path
|
|
|
|
from nanobot.agent.context import ContextBuilder
|
|
|
|
|
|
def _make_builder(tmp_path: Path) -> ContextBuilder:
|
|
"""Create a minimal ContextBuilder for testing."""
|
|
return ContextBuilder(workspace=tmp_path, timezone="UTC")
|
|
|
|
|
|
def test_build_user_content_with_no_media_returns_string(tmp_path: Path) -> None:
|
|
builder = _make_builder(tmp_path)
|
|
result = builder._build_user_content("hello", None)
|
|
assert result == "hello"
|
|
|
|
|
|
def test_build_user_content_with_image_returns_list(tmp_path: Path) -> None:
|
|
"""Image files should produce base64 content blocks."""
|
|
builder = _make_builder(tmp_path)
|
|
png = tmp_path / "test.png"
|
|
png.write_bytes(b"\x89PNG\r\n\x1a\n" + b"\x00" * 100)
|
|
result = builder._build_user_content("describe this", [str(png)])
|
|
assert isinstance(result, list)
|
|
types = [b["type"] for b in result]
|
|
assert "image_url" in types
|
|
assert "text" in types
|
|
|
|
|
|
def test_build_user_content_ignores_non_image_files(tmp_path: Path) -> None:
|
|
"""Non-image files should be silently skipped — extraction is not context builder's job."""
|
|
builder = _make_builder(tmp_path)
|
|
txt = tmp_path / "notes.txt"
|
|
txt.write_text("some text", encoding="utf-8")
|
|
result = builder._build_user_content("summarize", [str(txt)])
|
|
assert result == "summarize"
|
|
|
|
|
|
def test_build_user_content_mixed_image_and_non_image(tmp_path: Path) -> None:
|
|
"""Only images should be included; non-image files are skipped."""
|
|
builder = _make_builder(tmp_path)
|
|
png = tmp_path / "chart.png"
|
|
png.write_bytes(b"\x89PNG\r\n\x1a\n" + b"\x00" * 100)
|
|
txt = tmp_path / "report.txt"
|
|
txt.write_text("report text", encoding="utf-8")
|
|
|
|
result = builder._build_user_content("analyze", [str(png), str(txt)])
|
|
assert isinstance(result, list)
|
|
assert any(b["type"] == "image_url" for b in result)
|
|
text_parts = [b.get("text", "") for b in result if b.get("type") == "text"]
|
|
assert all("report text" not in t for t in text_parts)
|