Self-evolving Context Database for AI Agents. Unify Agent Memory, Knowledge RAG and Skills.
Python agent-contextagent-memoryagent-pluginsagentic-ragcontext-databasedsh-pluginknowledge-basememorymemory-managementself-evolvingvector-database
import os
import tempfile
import uuid
class TestResourceEndToEnd:
def test_add_directory_resource(self, api_client):
with tempfile.TemporaryDirectory() as temp_dir:
unique_kw = f"dir_e2e_{uuid.uuid4().hex[:8]}"
sub_dir = os.path.join(temp_dir, unique_kw)
os.makedirs(sub_dir)
for i in range(3):
with open(os.path.join(sub_dir, f"file_{i}.md"), "w") as f:
f.write(f"File {i} content about {unique_kw}\n")
add_resp = api_client.add_resource(path=sub_dir, wait=True)
assert add_resp.status_code == 200, (
f"add directory should return 200, got {add_resp.status_code}: {add_resp.text[:200]}"
)
result = add_resp.json().get("result", {})
root_uri = result.get("root_uri", "")
ls_resp = api_client.fs_ls(root_uri, recursive=True)
assert ls_resp.status_code == 200
children = ls_resp.json().get("result", [])
assert len(children) >= 3, (
f"directory resource should have at least 3 children, got {len(children)}"
)
def test_add_resource_with_reason_and_instruction(self, api_client):
with tempfile.TemporaryDirectory() as temp_dir:
test_file = os.path.join(temp_dir, f"reason_instr_{uuid.uuid4().hex[:8]}.md")
with open(test_file, "w") as f:
f.write("Content for reason and instruction test")
add_resp = api_client.add_resource(
path=test_file,
reason="Testing reason parameter for resource tracking",
instruction="Focus on key concepts and summarize concisely",
wait=True,
)
assert add_resp.status_code == 200, (
f"add_resource with reason+instruction should return 200, got {add_resp.status_code}: {add_resp.text[:200]}"
)
result = add_resp.json().get("result", {})
assert "root_uri" in result, "result should contain root_uri"
def test_add_resource_to_and_parent_conflict(self, api_client):
with tempfile.TemporaryDirectory() as temp_dir:
test_file = os.path.join(temp_dir, "conflict_test.txt")
with open(test_file, "w") as f:
f.write("Content for to+parent conflict test")
add_resp = api_client.add_resource(
path=test_file,
to="viking://resources/to_conflict",
parent="viking://resources/parent_conflict",
wait=False,
)
assert add_resp.status_code == 400, (
f"to+parent conflict should return error, got {add_resp.status_code}: {add_resp.text[:200]}"
)
if add_resp.status_code == 200:
raise AssertionError("to+parent should conflict and NOT return 200")
def test_add_same_resource_twice(self, api_client):
with tempfile.TemporaryDirectory() as temp_dir:
test_file = os.path.join(temp_dir, f"dup_{uuid.uuid4().hex[:8]}.md")
with open(test_file, "w") as f:
f.write("Duplicate resource test content")
add1 = api_client.add_resource(path=test_file, wait=True)
if add1.status_code != 200:
return
root_uri_1 = add1.json().get("result", {}).get("root_uri", "")
add2 = api_client.add_resource(path=test_file, wait=True)
assert add2.status_code == 200, (
f"adding same resource again should return 200 or 409, got {add2.status_code}: {add2.text[:200]}"
)
if add2.status_code == 200:
root_uri_2 = add2.json().get("result", {}).get("root_uri", "")
assert root_uri_1 != root_uri_2, (
"duplicate add should create a different resource (different root_uri)"
)
def test_add_resource_with_to_creates_at_target(self, api_client):
target_uri = f"viking://resources/target_{uuid.uuid4().hex[:8]}"
with tempfile.TemporaryDirectory() as temp_dir:
test_file = os.path.join(temp_dir, "targeted.txt")
with open(test_file, "w") as f:
f.write("Content placed at target URI")
add_resp = api_client.add_resource(path=test_file, to=target_uri, wait=True)
if add_resp.status_code != 200:
return
stat_resp = api_client.fs_stat(target_uri)
assert stat_resp.status_code == 200, (
"target URI should exist in filesystem after add_resource with to param"
)
def test_resource_searchable_by_semantic_content(self, api_client):
with tempfile.TemporaryDirectory() as temp_dir:
test_file = os.path.join(temp_dir, f"semantic_{uuid.uuid4().hex[:8]}.md")
with open(test_file, "w") as f:
f.write(
"# Neural Networks\n\n"
"Neural networks are computing systems inspired by biological neural networks. "
"They consist of layers of interconnected nodes that process information. "
"Deep learning uses multiple hidden layers to learn hierarchical representations."
)
add_resp = api_client.add_resource(path=test_file, wait=True)
if add_resp.status_code != 200:
return
find_resp = api_client.find(query="artificial intelligence deep learning", limit=5)
assert find_resp.status_code == 200
resources = find_resp.json().get("result", {}).get("resources", [])
if len(resources) > 0:
top_result = resources[0]
assert "score" in top_result, "search result should have score"
assert top_result["score"] > 0, "score should be positive for relevant result"
def test_resource_glob_finds_files(self, api_client):
with tempfile.TemporaryDirectory() as temp_dir:
unique_name = f"glob_find_{uuid.uuid4().hex[:8]}"
test_file = os.path.join(temp_dir, f"{unique_name}.md")
with open(test_file, "w") as f:
f.write("Glob test content")
add_resp = api_client.add_resource(path=test_file, wait=True)
if add_resp.status_code != 200:
return
root_uri = add_resp.json().get("result", {}).get("root_uri", "")
glob_resp = api_client.glob(pattern="**/*.md", uri=root_uri)
if glob_resp.status_code == 200:
matches = glob_resp.json().get("result", {}).get("matches", [])
assert len(matches) >= 1, (
f"glob *.md should find at least one file under resource, got {len(matches)}"
)