Skip to content

armgit / ODS

Public Imported from GitHub
0

ODS V3 Pre-Release: Public testing and refinement ahead of the official V3 launch. Turn your PC, Mac, or Linux box into a private AI server.

8.0 KB
Raw Blame
#!/usr/bin/env bash
# Regression checks for Linux cloud mode. Cloud/external LLM installs must not
# require a local llama-server container or local-mode dependency overlays.

set -euo pipefail

ROOT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")/.." && pwd)"
cd "$ROOT_DIR"

pass() { printf '[PASS] %s\n' "$1"; }
fail() { printf '[FAIL] %s\n' "$1" >&2; exit 1; }

contains() {
    local haystack="$1" needle="$2" label="$3"
    [[ "$haystack" == *"$needle"* ]] && pass "$label" || fail "$label"
}

rejects() {
    local haystack="$1" needle="$2" label="$3"
    [[ "$haystack" != *"$needle"* ]] && pass "$label" || fail "$label"
}

PY="${ODS_PYTHON_CMD:-}"
if [[ -z "$PY" ]]; then
    if command -v python3 >/dev/null 2>&1; then
        PY=python3
    elif command -v python >/dev/null 2>&1; then
        PY=python
    else
        fail "python is required"
    fi
fi

flags="$(ODS_PYTHON_CMD="$PY" ./scripts/resolve-compose-stack.sh \
    --script-dir "$ROOT_DIR" \
    --tier CLOUD \
    --gpu-backend cpu \
    --gpu-count 0 \
    --ods-mode cloud)"
flags="${flags//\\//}"

contains "$flags" "docker-compose.base.yml" "cloud mode keeps base stack"
contains "$flags" "docker-compose.cloud.yml" "cloud mode layers cloud overlay"
contains "$flags" "extensions/services/litellm/compose.yaml" "cloud mode includes LiteLLM gateway"
rejects "$flags" "docker-compose.cpu.yml" "cloud mode does not include CPU llama-server overlay"
rejects "$flags" "compose.local.yaml" "cloud mode does not include local dependency overlays"

# Installed user extensions can retain older backend-named overlays. Continue,
# for example, used compose.nvidia.yaml solely to wait for the local
# llama-server. Cloud mode must keep its route-agnostic base service while
# omitting that local-only readiness edge; local mode must still include it.
user_fixture="$(mktemp -d)"
trap 'rm -rf "$user_fixture"' EXIT
mkdir -p \
    "$user_fixture/scripts" \
    "$user_fixture/config" \
    "$user_fixture/extensions/services" \
    "$user_fixture/data/user-extensions/continue"
cp scripts/resolve-compose-stack.sh "$user_fixture/scripts/"
cp docker-compose.base.yml docker-compose.cloud.yml docker-compose.nvidia.yml "$user_fixture/"
cp config/core-service-ids.json "$user_fixture/config/"
cp extensions/library/services/continue/manifest.yaml \
    extensions/library/services/continue/compose.yaml \
    extensions/library/services/continue/compose.nvidia.yaml \
    "$user_fixture/data/user-extensions/continue/"

cloud_user_flags="$(ODS_PYTHON_CMD="$PY" "$user_fixture/scripts/resolve-compose-stack.sh" \
    --script-dir "$user_fixture" \
    --tier CLOUD \
    --gpu-backend nvidia \
    --gpu-count 1 \
    --ods-mode cloud)"
cloud_user_flags="${cloud_user_flags//\\//}"
contains "$cloud_user_flags" "data/user-extensions/continue/compose.yaml" \
    "cloud mode retains route-agnostic user extension base"
rejects "$cloud_user_flags" "data/user-extensions/continue/compose.nvidia.yaml" \
    "cloud mode omits user overlay that requires local inference"

local_user_flags="$(ODS_PYTHON_CMD="$PY" "$user_fixture/scripts/resolve-compose-stack.sh" \
    --script-dir "$user_fixture" \
    --tier 1 \
    --gpu-backend nvidia \
    --gpu-count 1 \
    --ods-mode local)"
local_user_flags="${local_user_flags//\\//}"
contains "$local_user_flags" "data/user-extensions/continue/compose.nvidia.yaml" \
    "local mode retains user overlay that requires local inference"

native_flags="$(NATIVE_LLM_BASE_URL=http://localhost:8080 ODS_PYTHON_CMD="$PY" ./scripts/resolve-compose-stack.sh \
    --script-dir "$ROOT_DIR" \
    --tier 1 \
    --gpu-backend cpu \
    --gpu-count 0 \
    --ods-mode local)"
native_flags="${native_flags//\\//}"

contains "$native_flags" "docker-compose.base.yml" "host-native llama-server keeps base stack"
rejects "$native_flags" "docker-compose.cloud.yml" "host-native llama-server retains model-router instead of cloud profile gate"
contains "$native_flags" "docker-compose.host-native-llm.yml" "host-native llama-server layers dedicated overlay"
rejects "$native_flags" "docker-compose.cpu.yml" "host-native llama-server does not include CPU llama-server overlay"
rejects "$native_flags" "compose.local.yaml" "host-native llama-server adds no local-inference dependencies"

if grep -q 'profiles:' docker-compose.cloud.yml && grep -q 'local-inference' docker-compose.cloud.yml; then
    pass "cloud overlay profiles local llama-server out of default startup"
else
    fail "cloud overlay must profile local llama-server out of default startup"
fi

"$PY" - <<'PY'
from pathlib import Path
import sys
import yaml

services = yaml.safe_load(Path("docker-compose.cloud.yml").read_text(encoding="utf-8"))["services"]
for name in ("llama-server", "model-router"):
    service = services.get(name, {})
    if "local-inference" not in service.get("profiles", []) or service.get("restart") != "no":
        print(f"[FAIL] cloud mode must profile {name} out with its local dependency chain", file=sys.stderr)
        sys.exit(1)
if "pixel-model-relay" in services:
    print("[FAIL] cloud overlay must not disable the enabled Pixel relay", file=sys.stderr)
    sys.exit(1)
print("[PASS] cloud mode profiles local inference out and retains Pixel's external gateway route")

native = yaml.safe_load(Path("docker-compose.host-native-llm.yml").read_text(encoding="utf-8"))["services"]
if native.get("llama-server", {}).get("profiles") != ["local-inference"]:
    print("[FAIL] host-native llama-server must disable only the in-stack llama-server", file=sys.stderr)
    sys.exit(1)
if "profiles" in native.get("model-router", {}):
    print("[FAIL] host-native llama-server must leave model-router enabled", file=sys.stderr)
    sys.exit(1)
print("[PASS] host-native llama-server disables the in-stack llama-server and keeps model-router")
PY

if grep -Fq -- '--ods-mode "${ODS_MODE:-local}"' installers/lib/compose-select.sh \
    && grep -Fq -- '--ods-mode "${ODS_MODE:-local}"' installers/phases/03-features.sh \
    && grep -Fq -- '--ods-mode "${ODS_MODE:-local}"' installers/phases/11-services.sh \
    && grep -Fq -- '--ods-mode "${ODS_MODE:-local}"' ods-cli; then
    pass "installer and CLI pass ods mode to compose resolver"
else
    fail "all installer/CLI resolver calls must pass --ods-mode"
fi

if grep -q 'ODS_MODE:-local.*cloud' installers/phases/12-health.sh \
    && grep -Fq 'LiteLLM' installers/phases/12-health.sh \
    && grep -Fq 'skipping local llama-server pre-warm' installers/phases/12-health.sh; then
    pass "cloud health path skips local llama-server"
else
    fail "cloud health path must skip local llama-server"
fi

if grep -Fq 'image: ${HERMES_AGENT_IMAGE:-nousresearch/hermes-agent:v2026.9.24@sha256:fca358f12efd65bfaaca05884166f15c0e2788375ca30d77061ac1ebc96452b7}' extensions/services/hermes/compose.yaml \
    && grep -Fq '${HERMES_AGENT_IMAGE:-nousresearch/hermes-agent:v2026.9.24@sha256:fca358f12efd65bfaaca05884166f15c0e2788375ca30d77061ac1ebc96452b7}|HERMES' installers/phases/08-images.sh \
    && grep -Fq 'HERMES_AGENT_IMAGE_FALLBACK' installers/phases/08-images.sh \
    && ! grep -R -q 'nousresearch/hermes-agent:sha-' extensions/services/hermes installers/phases config/dependency-lock.json; then
    pass "Hermes image default is resolvable and overrideable for cloud installs"
else
    fail "Hermes image default must not rely on removed sha-* Docker tags"
fi

"$PY" - "$ROOT_DIR" <<'PY'
from pathlib import Path
import sys

root = Path(sys.argv[1])
text = (root / "installers/phases/11-services.sh").read_text(encoding="utf-8")
model_config = text.index('mkdir -p "$INSTALL_DIR/config/llama-server"')
hermes_block = text.index('if [[ "${ENABLE_HERMES:-false}" == "true" ]]; then')
soul_block = text.index('_soul_output="$INSTALL_DIR/data/persona/SOUL.md"')
if model_config < hermes_block < soul_block:
    # Make sure the local-model block was closed before Hermes/SOUL rendering begins.
    between = text[model_config:hermes_block]
    if '\n    fi\n' in between:
        print("[PASS] SOUL.md render is outside local-model-only block")
        sys.exit(0)
print("[FAIL] SOUL.md render must run for cloud installs too", file=sys.stderr)
sys.exit(1)
PY

echo "[PASS] linux cloud mode contracts"