Buckets:

download
raw
3.84 kB
"""CRUD-oriented py-cml example with typed writes and filtered reads."""
from __future__ import annotations
from uuid import UUID
from _shared import build_cml_config, explain_connection_failure, print_header
from cml import CognitiveMemoryLayer
from cml.models import MemoryType
EXAMPLE_META = {
"name": "basic_usage",
"kind": "python",
"summary": "Typed writes plus read, update, stats, and forget flows.",
"requires_api": True,
"requires_api_key": True,
"requires_base_url": True,
"requires_admin_key": False,
"requires_embedded": False,
"requires_openai": False,
"requires_anthropic": False,
"interactive": False,
"timeout_sec": 60,
}
def main() -> int:
print_header("CML Basic Usage")
session_id = "examples-basic-usage"
try:
with CognitiveMemoryLayer(config=build_cml_config()) as memory:
writes = [
(
"The user is Alice and works as a software engineer.",
MemoryType.SEMANTIC_FACT,
["identity"],
),
(
"The user prefers Python for backend work.",
MemoryType.PREFERENCE,
["preferences"],
),
(
"The user is allergic to shellfish and should not receive seafood recommendations.",
MemoryType.CONSTRAINT,
["medical", "constraints"],
),
(
"On 2026-02-15 the user said they want to learn Rust this year.",
MemoryType.EPISODIC_EVENT,
["timeline"],
),
(
"The user might be interested in machine learning.",
MemoryType.HYPOTHESIS,
["inference"],
),
]
for text, memory_type, tags in writes:
response = memory.write(
text,
session_id=session_id,
context_tags=tags,
memory_type=memory_type,
)
print(f"Stored {memory_type.value}: success={response.success}")
filtered = memory.read(
"What programming languages does the user like?",
memory_types=[MemoryType.PREFERENCE, MemoryType.CONSTRAINT],
)
print(f"\nFiltered read returned {filtered.total_count} memories")
for item in filtered.memories:
print(f" - [{item.type}] {item.text}")
llm_ready = memory.read("Summarize the user", response_format="llm_context")
print(f"\nLLM context snippet: {(llm_ready.context or '')[:200]}")
if llm_ready.constraints:
print("Constraints:")
for constraint in llm_ready.constraints[:3]:
print(f" - {constraint.text}")
hypothesis = memory.read("machine learning")
if hypothesis.memories:
memory_id = hypothesis.memories[0].id
if not isinstance(memory_id, UUID):
memory_id = UUID(str(memory_id))
updated = memory.update(memory_id=memory_id, feedback="correct")
print(f"\nUpdated hypothesis version -> {updated.version}")
stats = memory.stats()
print(f"\nStats by type: {stats.by_type}")
forgotten = memory.forget(query="learn Rust", action="archive")
print(f"Archived {forgotten.affected_count} matching memories")
return 0
except Exception as exc:
print(f"Example failed: {exc}")
print(explain_connection_failure())
return 1
if __name__ == "__main__":
raise SystemExit(main())

Xet Storage Details

Size:
3.84 kB
·
Xet hash:
835bf1742d3b9905e269ab7fe4959b6fed5067cf769ff3444d9ea50b5c8fa8e7

Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.