Files
ely-eye/scripts/mcp_write_smoke.sh
T
ZacharyZhang-NY 9fe13fffb3 Make a cartridge a scope you mount, not only a bundle you export
A cartridge used to be an export: materialize it, hash it, and never let it
change behaviour again. The transactive-memory promise needs the opposite —
mounting the home cartridge must make the work cartridge's memories
unfindable, and appending to a cartridge must make a new memory findable
through it.

Cartridge scope (schema v2 -> v3):
- cartridge_atoms (which atoms are in which cartridge) and cartridge_mounts
  (which cartridges this machine has mounted now); both rebuilt on reopen,
  because a mount is a machine-local state, not a fact about the data.
- mount / switch / unmount / append refuse a cartridge that was built for a
  different runtime, naming every differing field instead of a boolean.
- append rewrites atoms.parquet and sparse_index.bm25.zst, recomputes DNA and
  the KV fingerprint, and keeps the cartridges row in one SQLite transaction;
  the files are written first so a failed export cannot leave a half-cartridge
  in the database.
- RetrievalService.search() now ranks inside the mounted scope by default, and
  an explicit cartridge_ids filter overrides the mount. Nothing mounted means
  the whole library, so an empty store behaves as it did before.
- forget() removes the forgotten atoms from every cartridge's membership:
  forgetting must not leave a memory retrievable through a cartridge.

KV fingerprint (kv_fingerprint.py, stored in the manifest):
- model / tokenizer / chat-template hashes, the RoPE block (factor,
  original_max_position_embeddings, mrope_interleaved, mrope_section,
  partial_rotary_factor), the adapter hash, the context length, the engine
  version, and the fingerprint format version.
- A snapshot we cannot read is recorded as unverifiable and listed in
  unverifiable_fields; a blind machine refuses to mount rather than pretending
  the runtime matches, because a wrong RoPE factor silently corrupts retrieval.

Two defects found by running the new tests, not by reading:
- atoms.parquet could not export a memory at all: pyarrow refuses a struct with
  no child field, so metadata={} failed the whole table. The six JSON columns
  are now written as JSON text and decoded by the cartridge reader.
- materialize() inserted cartridge membership before the cartridges row existed,
  so the foreign key rejected every cartridge.

api.py crossed 500 lines, so the cartridge and mount routes moved to
api_cartridges.py. That module must not use `from __future__ import
annotations`: FastAPI then resolves Depends(settings_dep) against module
globals, where the injected dependency is only a parameter of register(), and
every route starts demanding settings as a query parameter.

Verification: 138 pytest passed (26 new: 10 fingerprint, 16 scope), Go
build/vet/gofmt/test clean, and scripts/mcp_write_smoke.sh now proves the whole
loop across two processes - mount narrows search, a memory written after the
upload is invisible until append_to_cartridge, unmount restores the library,
and rewriting the RoPE factor makes mount_cartridge refuse with "rope: factor:
cartridge 4.0 vs local 2.0". ruff clean on every touched file (the 20 remaining
package errors are in files this change did not touch; HEAD had 25). mypy: one
error, the pre-existing np.savez stub artifact at cartridge_assets.py:329
(HEAD:320). Repo pyright: 62 errors, unchanged from the p0-3 ledger.
2026-10-07 16:40:27 -04:00

187 lines
7.8 KiB
Bash
Executable File

#!/usr/bin/env bash
# Prove the MCP write tools reach the Python writer over HTTP, end to end.
set -euo pipefail
root="$(cd "$(dirname "${BASH_SOURCE[0]}")/.." && pwd)"
home="$(mktemp -d)"
port="${PORT:-8799}"
export ELY_EYE_HOME="$home"
export ELY_EYE_API="http://127.0.0.1:${port}"
python_bin="$root/.venv/bin/python"
[ -x "$python_bin" ] || python_bin="python3"
# A model snapshot the KV fingerprint can read: without it every mount is refused as
# unverifiable, which is the behaviour, but the smoke test also needs a mount to succeed.
model_dir="$home/model"
mkdir -p "$model_dir"
cat > "$model_dir/config.json" <<'JSON'
{"model_type": "qwen3_5", "max_position_embeddings": 262144, "partial_rotary_factor": 0.25,
"rope_parameters": {"rope_type": "yarn", "factor": 4.0, "mrope_section": [11, 11, 10]},
"chat_template": "{% for message in messages %}{{ message['content'] }}{% endfor %}"}
JSON
printf '{"version": "1.0"}' > "$model_dir/tokenizer.json"
printf '{"metadata": {}, "weight_map": {"model-00001.safetensors": "0"}}' > "$model_dir/model.safetensors.index.json"
export ELY_EYE_MODEL_DIR="$model_dir"
api=""
cleanup() {
if [ -n "$api" ]; then kill "$api" 2>/dev/null || true; fi
rm -rf "$home"
}
trap cleanup EXIT
"$python_bin" -m ely_eye.cli serve --port "$port" >"$home/api.log" 2>&1 &
api=$!
ready=0
for _ in $(seq 1 80); do
if curl -sf "http://127.0.0.1:${port}/api/status" >/dev/null 2>&1; then ready=1; break; fi
sleep 0.5
done
if [ "$ready" -ne 1 ]; then
echo "FAIL: the Ely-Eye API did not start"
cat "$home/api.log"
exit 1
fi
(cd "$root/ely-eye-mcp" && go build -o "$home/ely-eye-mcp" ./cmd/ely-eye-mcp)
printf 'the pantry keeps flour\n' > "$home/pantry.txt"
uploaded="$(curl -sf -X POST "http://127.0.0.1:${port}/api/ingest/upload?cartridge_name=home" \
-F "files=@${home}/pantry.txt;type=text/plain")"
cartridge_id="$(printf '%s' "$uploaded" | "$python_bin" -c 'import json,sys; print(json.load(sys.stdin)["cartridge_id"])')"
if [ -z "$cartridge_id" ] || [ "$cartridge_id" = "None" ]; then
echo "FAIL: the upload did not materialize a cartridge: $uploaded"
exit 1
fi
"$python_bin" - "$home/ely-eye-mcp" "$cartridge_id" "$model_dir" <<'PY'
import json
import pathlib
import subprocess
import sys
binary = sys.argv[1]
cartridge_id = sys.argv[2]
model_dir = pathlib.Path(sys.argv[3])
proc = subprocess.Popen(
[binary, "server"],
stdin=subprocess.PIPE,
stdout=subprocess.PIPE,
stderr=subprocess.DEVNULL,
text=True,
)
def send(message):
proc.stdin.write(json.dumps(message, ensure_ascii=False) + "\n")
proc.stdin.flush()
def call(id_, method, params):
send({"jsonrpc": "2.0", "id": id_, "method": method, "params": params})
while True:
line = proc.stdout.readline()
if not line:
raise SystemExit(f"FAIL: the MCP server closed the pipe at request {id_}")
answer = json.loads(line)
if answer.get("id") == id_:
if "error" in answer:
raise SystemExit(f"FAIL: {method} returned {answer['error']}")
return answer["result"]
def tool(name, arguments):
result = call(3, "tools/call", {"name": name, "arguments": arguments})
if result.get("isError"):
raise SystemExit(f"FAIL: {name} returned an error: {result['content']}")
return result["structuredContent"]
call(1, "initialize", {
"protocolVersion": "2025-06-18",
"capabilities": {},
"clientInfo": {"name": "ely-eye-smoke", "version": "0"},
})
send({"jsonrpc": "2.0", "method": "notifications/initialized"})
names = [item["name"] for item in call(2, "tools/list", {})["tools"]]
for required in ("remember", "supersede_memory", "forget_memory", "mounted_cartridges",
"mount_cartridge", "switch_cartridge", "unmount_cartridge", "append_to_cartridge"):
if required not in names:
raise SystemExit(f"FAIL: {required} is not registered; tools are {names}")
statement = "keys are in the hall drawer"
remembered = tool("remember", {"text": statement, "scenario": "home"})["atom"]
if not remembered["holder_id"]:
raise SystemExit(f"FAIL: the server stamped no holder: {remembered}")
atom_id = remembered["atom_id"]
hits = tool("search_atoms", {"query": statement})["hits"]
if not hits or hits[0]["atom"]["atom_id"] != atom_id:
raise SystemExit(f"FAIL: the new memory is not retrievable: {hits}")
if hits[0]["atom"].get("scenario") != "home":
raise SystemExit(f"FAIL: the search hit lost the scenario: {hits[0]['atom']}")
chinese = tool("remember", {"text": "我们的钥匙放在玄关抽屉里", "scenario": "home"})["atom"]
short = tool("search_atoms", {"query": "钥匙"})["hits"]
if not any(hit["atom"]["atom_id"] == chinese["atom_id"] for hit in short):
raise SystemExit(f"FAIL: a two-character Chinese query missed the memory written in Chinese: {short}")
if tool("search_atoms", {"query": "keys", "scenario": "travel"})["hits"]:
raise SystemExit("FAIL: the scenario filter let another scenario answer")
replacement = "keys are on the coffee table"
newer = tool("supersede_memory", {"atom_id": atom_id, "text": replacement})["atom"]
if newer["supersedes"] != [atom_id]:
raise SystemExit(f"FAIL: the replacement does not point back: {newer}")
after = tool("search_atoms", {"query": statement})["hits"]
if any(hit["atom"]["atom_id"] == atom_id for hit in after):
raise SystemExit(f"FAIL: the superseded memory is still retrievable: {after}")
report = tool("forget_memory", {"atom_id": atom_id})["report"]
if report["atoms_deleted"] != 1:
raise SystemExit(f"FAIL: forget deleted {report['atoms_deleted']} atoms, want 1")
gone = call(3, "tools/call", {"name": "fetch_atom", "arguments": {"atom_id": atom_id}})
if not gone.get("isError") or "not found" not in gone["content"][0]["text"]:
raise SystemExit(f"FAIL: the forgotten memory is still readable: {gone}")
state = tool("mounted_cartridges", {})["state"]
if state["mounted"] or state["scoped"]:
raise SystemExit(f"FAIL: a fresh store reports a scope: {state}")
outside = tool("remember", {"text": "the spare key hides in the garage", "scenario": "work"})["atom"]
manifest = tool("mount_cartridge", {"cartridge_id": cartridge_id})["manifest"]
if manifest["cartridge_id"] != cartridge_id or not manifest.get("dna"):
raise SystemExit(f"FAIL: mount returned the wrong header: {manifest}")
state = tool("mounted_cartridges", {})["state"]
if not state["scoped"] or state["mounted"] != [cartridge_id]:
raise SystemExit(f"FAIL: the mount did not take effect: {state}")
if tool("search_atoms", {"query": "garage"})["hits"]:
raise SystemExit("FAIL: a mounted cartridge let a memory outside it answer")
grown = tool("append_to_cartridge", {"cartridge_id": cartridge_id, "atom_ids": [outside["atom_id"]]})["manifest"]
if grown["dna"] == manifest["dna"]:
raise SystemExit("FAIL: append rewrote nothing; the DNA is unchanged")
if not any(hit["atom"]["atom_id"] == outside["atom_id"] for hit in tool("search_atoms", {"query": "garage"})["hits"]):
raise SystemExit("FAIL: an appended memory is still not reachable")
tool("unmount_cartridge", {"cartridge_id": cartridge_id})
if tool("mounted_cartridges", {})["state"]["scoped"]:
raise SystemExit("FAIL: unmount left the store scoped")
(model_dir / "config.json").write_text(json.dumps({
"model_type": "qwen3_5",
"max_position_embeddings": 262144,
"partial_rotary_factor": 0.25,
"rope_parameters": {"rope_type": "yarn", "factor": 2.0},
"chat_template": "{% for message in messages %}{{ message['content'] }}{% endfor %}",
}), encoding="utf-8")
refused = call(3, "tools/call", {"name": "mount_cartridge", "arguments": {"cartridge_id": cartridge_id}})
if not refused.get("isError") or "rope" not in refused["content"][0]["text"]:
raise SystemExit(f"FAIL: a cartridge built for RoPE factor 4.0 mounted on a factor 2.0 machine: {refused}")
print("PASS: remember, supersede, search, forget and cartridge mounting work through MCP")
PY