chore: fork @llblab/[email protected].0 as @evandrodevbr/oh-my-pi-telegram
Fork of upstream llblab/[email protected].0 (SHA afe09c5) repackaged as a drop-in extension for can1357/oh-my-pi (omp >= 16.5.2). Changes vs upstream: - Rename package: @llblab/pi-telegram -> @evandrodevbr/oh-my-pi-telegram - Bump version: 0.22.0 -> 0.22.0-evandro.1 - Update repository/homepage/bugs URLs to evandrodevbr - Add omp field in package.json (displayName + compatibleWith) - README: prepend 'Oh My Pi (omp) Compatibility' section with install commands, verified-components table, upstream lineage, and rationale - CHANGELOG: add '0.22.0-evandro.1' entry under Unreleased - docs/: rename @llblab/pi-telegram imports to @evandrodevbr/oh-my-pi-telegram - tests/public-api.test.ts: rename package import paths Gates re-run on the fork: - npm run typecheck -> tsc --noEmit (success) - npm test -> 1229 pass / 0 fail / 1 platform-only skip - npm run audit -> found 0 vulnerabilities - npm run pack:check -> evandrodevbr-oh-my-pi-telegram-0.22.0-evandro.1.tgz Verified: - omp install --link ./pi-telegram -> Linked OK - omp plugin list -> @evandrodevbr/[email protected].0-evandro.1 Upstream lineage: badlogic/pi-telegram -> llblab/pi-telegram (0.22.0) -> evandrodevbr/oh-my-pi-telegram (this fork)
This commit is contained in:
Evandro Fonseca Junior committed
commit
d1d04001d0
155 files changed
+107848
No files matched your search
@@ -0,0 +1,486 @@
|
||||
/**
|
||||
* Regression tests for Telegram model control helpers
|
||||
* Exercises model selection state/resolution plus in-flight model-switch orchestration
|
||||
*/
|
||||
|
||||
import assert from "node:assert/strict";
|
||||
import test from "node:test";
|
||||
|
||||
import {
|
||||
buildTelegramModelSwitchContinuationText,
|
||||
buildTelegramModelSwitchContinuationTurn,
|
||||
canRestartTelegramTurnForModelSwitch,
|
||||
createCurrentModelRuntime,
|
||||
createCurrentModelStore,
|
||||
createCurrentModelUpdateRuntime,
|
||||
createPendingModelSwitchStore,
|
||||
createTelegramModelSwitchContinuationQueue,
|
||||
createTelegramModelSwitchContinuationQueueRuntime,
|
||||
createTelegramModelSwitchContinuationTurnBuilder,
|
||||
createTelegramModelSwitchController,
|
||||
createTelegramModelSwitchControllerRuntime,
|
||||
getCanonicalModelId,
|
||||
isThinkingLevel,
|
||||
type MenuModel,
|
||||
modelsMatch,
|
||||
parseTelegramCliScopedModelPatterns,
|
||||
parseTelegramScopedModelPatternList,
|
||||
resolveScopedModelPatterns,
|
||||
restartTelegramModelSwitchContinuation,
|
||||
shouldTriggerPendingTelegramModelSwitchAbort,
|
||||
sortScopedModels,
|
||||
THINKING_LEVELS,
|
||||
} from "../lib/model.ts";
|
||||
import type { PendingTelegramTurn } from "../lib/queue.ts";
|
||||
|
||||
function createModelTestModel(
|
||||
provider = "openai",
|
||||
id = "gpt-5",
|
||||
extra: Partial<MenuModel> = {},
|
||||
): MenuModel {
|
||||
return { provider, id, ...extra };
|
||||
}
|
||||
|
||||
function createModelTestSelection(
|
||||
model = createModelTestModel(),
|
||||
thinkingLevel?: "high",
|
||||
): { model: MenuModel; thinkingLevel?: "high" } {
|
||||
return thinkingLevel === undefined ? { model } : { model, thinkingLevel };
|
||||
}
|
||||
|
||||
function createModelTestTurn(
|
||||
overrides: Partial<PendingTelegramTurn> = {},
|
||||
): PendingTelegramTurn {
|
||||
return {
|
||||
kind: "prompt",
|
||||
chatId: 7,
|
||||
replyToMessageId: 8,
|
||||
sourceMessageIds: [8],
|
||||
queueOrder: 0,
|
||||
queueLane: "default",
|
||||
laneOrder: 0,
|
||||
queuedAttachments: [],
|
||||
content: [],
|
||||
historyText: "active",
|
||||
statusSummary: "active",
|
||||
...overrides,
|
||||
};
|
||||
}
|
||||
|
||||
test("Current model store prefers explicit Telegram model over context fallback", () => {
|
||||
const fallback = { provider: "fallback", id: "f", name: "Fallback" };
|
||||
const selected = { provider: "selected", id: "s", name: "Selected" };
|
||||
const store = createCurrentModelStore<
|
||||
{ model?: typeof fallback },
|
||||
typeof fallback
|
||||
>((ctx) => ctx.model);
|
||||
assert.equal(store.get({ model: fallback }), fallback);
|
||||
assert.equal(store.getStored(), undefined);
|
||||
store.set(selected);
|
||||
assert.equal(store.get({ model: fallback }), selected);
|
||||
assert.equal(store.getStored(), selected);
|
||||
store.set(undefined);
|
||||
assert.equal(store.get({ model: fallback }), fallback);
|
||||
});
|
||||
|
||||
test("Current model runtime combines store fallback and status updates", () => {
|
||||
const fallback = { provider: "fallback", id: "f", name: "Fallback" };
|
||||
const selected = { provider: "selected", id: "s", name: "Selected" };
|
||||
const statuses: string[] = [];
|
||||
const runtime = createCurrentModelRuntime<
|
||||
{ id: string; model?: typeof fallback },
|
||||
typeof fallback
|
||||
>({
|
||||
getContextModel: (ctx) => ctx.model,
|
||||
updateStatus: (ctx) => {
|
||||
statuses.push(ctx.id);
|
||||
},
|
||||
});
|
||||
assert.equal(runtime.get({ id: "fallback", model: fallback }), fallback);
|
||||
runtime.setCurrentModel(selected, { id: "direct" });
|
||||
assert.equal(runtime.get({ id: "fallback", model: fallback }), selected);
|
||||
runtime.onModelSelect({ model: undefined }, { id: "event" });
|
||||
assert.equal(runtime.get({ id: "fallback", model: fallback }), fallback);
|
||||
assert.deepEqual(statuses, ["direct", "event"]);
|
||||
});
|
||||
|
||||
test("Current model update runtime stores selections and refreshes status", () => {
|
||||
const selected = { provider: "selected", id: "s", name: "Selected" };
|
||||
let stored: typeof selected | undefined;
|
||||
const statuses: string[] = [];
|
||||
const runtime = createCurrentModelUpdateRuntime<
|
||||
{ id: string },
|
||||
typeof selected
|
||||
>({
|
||||
setCurrentModel: (model) => {
|
||||
stored = model;
|
||||
},
|
||||
updateStatus: (ctx) => {
|
||||
statuses.push(ctx.id);
|
||||
},
|
||||
});
|
||||
runtime.setCurrentModel(selected, { id: "direct" });
|
||||
assert.equal(stored, selected);
|
||||
runtime.onModelSelect({ model: selected }, { id: "event" });
|
||||
assert.deepEqual(statuses, ["direct", "event"]);
|
||||
});
|
||||
|
||||
test("Model helpers match models, detect thinking levels, and expose constants", () => {
|
||||
assert.deepEqual(THINKING_LEVELS, [
|
||||
"off",
|
||||
"minimal",
|
||||
"low",
|
||||
"medium",
|
||||
"high",
|
||||
"xhigh",
|
||||
"max",
|
||||
]);
|
||||
assert.equal(
|
||||
modelsMatch(createModelTestModel(), createModelTestModel()),
|
||||
true,
|
||||
);
|
||||
assert.equal(
|
||||
modelsMatch(
|
||||
createModelTestModel(),
|
||||
createModelTestModel("anthropic", "gpt-5"),
|
||||
),
|
||||
false,
|
||||
);
|
||||
assert.equal(getCanonicalModelId(createModelTestModel()), "openai/gpt-5");
|
||||
assert.equal(isThinkingLevel("high"), true);
|
||||
assert.equal(isThinkingLevel("impossible"), false);
|
||||
});
|
||||
|
||||
test("Model helpers parse scoped model pattern lists and CLI flags", () => {
|
||||
assert.deepEqual(parseTelegramScopedModelPatternList("a, b:high, "), [
|
||||
"a",
|
||||
"b:high",
|
||||
]);
|
||||
assert.deepEqual(
|
||||
parseTelegramCliScopedModelPatterns(["--models", "a,b:high"]),
|
||||
["a", "b:high"],
|
||||
);
|
||||
assert.deepEqual(parseTelegramCliScopedModelPatterns(["--models=a, b "]), [
|
||||
"a",
|
||||
"b",
|
||||
]);
|
||||
assert.equal(
|
||||
parseTelegramCliScopedModelPatterns(["--models", " "]),
|
||||
undefined,
|
||||
);
|
||||
assert.equal(
|
||||
parseTelegramCliScopedModelPatterns(["--other", "a"]),
|
||||
undefined,
|
||||
);
|
||||
});
|
||||
|
||||
test("Model helpers resolve scoped model patterns and sort current models first", () => {
|
||||
const models: MenuModel[] = [
|
||||
createModelTestModel("openai", "gpt-5", { name: "GPT 5" }),
|
||||
createModelTestModel("openai", "gpt-5-latest", { name: "GPT 5 Latest" }),
|
||||
createModelTestModel("anthropic", "claude-sonnet-20250101", {
|
||||
name: "Claude Sonnet",
|
||||
}),
|
||||
];
|
||||
const resolved = resolveScopedModelPatterns(
|
||||
["gpt-5:high", "anthropic/*:low"],
|
||||
models,
|
||||
);
|
||||
assert.deepEqual(
|
||||
resolved.map((entry) => ({
|
||||
id: entry.model.id,
|
||||
thinking: entry.thinkingLevel,
|
||||
})),
|
||||
[
|
||||
{ id: "gpt-5", thinking: "high" },
|
||||
{ id: "claude-sonnet-20250101", thinking: "low" },
|
||||
],
|
||||
);
|
||||
const sorted = sortScopedModels(resolved, models[0]);
|
||||
assert.equal(sorted[0]?.model.id, "gpt-5");
|
||||
});
|
||||
|
||||
test("Pending model-switch store owns selection state helpers", () => {
|
||||
const store = createPendingModelSwitchStore<{ id: string }>();
|
||||
assert.equal(store.has(), false);
|
||||
assert.equal(store.get(), undefined);
|
||||
store.set({ id: "next" });
|
||||
assert.equal(store.has(), true);
|
||||
assert.deepEqual(store.get(), { id: "next" });
|
||||
store.clear();
|
||||
assert.equal(store.has(), false);
|
||||
});
|
||||
|
||||
test("In-flight model switch is allowed only for active Telegram turns with abort support", () => {
|
||||
assert.equal(
|
||||
canRestartTelegramTurnForModelSwitch({
|
||||
isIdle: false,
|
||||
hasActiveTelegramTurn: true,
|
||||
hasAbortHandler: true,
|
||||
}),
|
||||
true,
|
||||
);
|
||||
assert.equal(
|
||||
canRestartTelegramTurnForModelSwitch({
|
||||
isIdle: true,
|
||||
hasActiveTelegramTurn: true,
|
||||
hasAbortHandler: true,
|
||||
}),
|
||||
false,
|
||||
);
|
||||
assert.equal(
|
||||
canRestartTelegramTurnForModelSwitch({
|
||||
isIdle: false,
|
||||
hasActiveTelegramTurn: false,
|
||||
hasAbortHandler: true,
|
||||
}),
|
||||
false,
|
||||
);
|
||||
assert.equal(
|
||||
canRestartTelegramTurnForModelSwitch({
|
||||
isIdle: false,
|
||||
hasActiveTelegramTurn: true,
|
||||
hasAbortHandler: false,
|
||||
}),
|
||||
false,
|
||||
);
|
||||
});
|
||||
|
||||
test("Pending model switch abort waits until no tool executions remain", () => {
|
||||
assert.equal(
|
||||
shouldTriggerPendingTelegramModelSwitchAbort({
|
||||
hasPendingModelSwitch: true,
|
||||
hasActiveTelegramTurn: true,
|
||||
hasAbortHandler: true,
|
||||
activeToolExecutions: 0,
|
||||
}),
|
||||
true,
|
||||
);
|
||||
assert.equal(
|
||||
shouldTriggerPendingTelegramModelSwitchAbort({
|
||||
hasPendingModelSwitch: true,
|
||||
hasActiveTelegramTurn: true,
|
||||
hasAbortHandler: true,
|
||||
activeToolExecutions: 1,
|
||||
}),
|
||||
false,
|
||||
);
|
||||
assert.equal(
|
||||
shouldTriggerPendingTelegramModelSwitchAbort({
|
||||
hasPendingModelSwitch: false,
|
||||
hasActiveTelegramTurn: true,
|
||||
hasAbortHandler: true,
|
||||
activeToolExecutions: 0,
|
||||
}),
|
||||
false,
|
||||
);
|
||||
});
|
||||
|
||||
test("Model-switch continuation restart queues before abort when state is present", () => {
|
||||
const events: string[] = [];
|
||||
assert.equal(
|
||||
restartTelegramModelSwitchContinuation({
|
||||
activeTurn: { id: 1 },
|
||||
abort: () => {
|
||||
events.push("abort");
|
||||
},
|
||||
selection: createModelTestSelection(),
|
||||
queueContinuation: (turn, selection) => {
|
||||
events.push(`queue:${turn.id}:${selection.model.id}`);
|
||||
},
|
||||
}),
|
||||
true,
|
||||
);
|
||||
assert.deepEqual(events, ["queue:1:gpt-5", "abort"]);
|
||||
assert.equal(
|
||||
restartTelegramModelSwitchContinuation({
|
||||
activeTurn: undefined,
|
||||
abort: () => {},
|
||||
selection: createModelTestSelection(),
|
||||
queueContinuation: () => {
|
||||
events.push("unexpected");
|
||||
},
|
||||
}),
|
||||
false,
|
||||
);
|
||||
});
|
||||
|
||||
test("Model-switch controller centralizes pending abort and continuation queueing", () => {
|
||||
const events: string[] = [];
|
||||
let pendingSelection: { model: { provider: string; id: string } } | undefined;
|
||||
const activeTurn = createModelTestTurn({
|
||||
chatId: 1,
|
||||
replyToMessageId: 2,
|
||||
sourceMessageIds: [2],
|
||||
queueOrder: 3,
|
||||
laneOrder: 3,
|
||||
historyText: "previous",
|
||||
statusSummary: "previous",
|
||||
});
|
||||
const controller = createTelegramModelSwitchController({
|
||||
isIdle: () => false,
|
||||
getPendingModelSwitch: () => pendingSelection,
|
||||
setPendingModelSwitch: (selection) => {
|
||||
pendingSelection = selection;
|
||||
},
|
||||
getActiveTurn: () => activeTurn,
|
||||
getAbortHandler: () => {
|
||||
return () => {
|
||||
events.push("abort");
|
||||
};
|
||||
},
|
||||
hasAbortHandler: () => true,
|
||||
getActiveToolExecutions: () => 0,
|
||||
queueContinuation: (turn, selection) => {
|
||||
events.push(`queue:${turn.replyToMessageId}:${selection.model.id}`);
|
||||
},
|
||||
updateStatus: () => {
|
||||
events.push("status");
|
||||
},
|
||||
});
|
||||
assert.equal(controller.canOfferInFlightSwitch({}), true);
|
||||
controller.stagePendingSwitch(createModelTestSelection(), {});
|
||||
assert.deepEqual(events, ["status"]);
|
||||
assert.equal(controller.triggerPendingAbort({}), true);
|
||||
assert.deepEqual(events, ["status", "queue:2:gpt-5", "abort"]);
|
||||
assert.equal(pendingSelection, undefined);
|
||||
});
|
||||
|
||||
test("Model-switch controller runtime binds continuation queue construction", () => {
|
||||
const appended: PendingTelegramTurn[] = [];
|
||||
let pending: { model: { provider: string; id: string } } | undefined =
|
||||
createModelTestSelection();
|
||||
let aborted = false;
|
||||
const activeTurn = createModelTestTurn();
|
||||
const controller = createTelegramModelSwitchControllerRuntime({
|
||||
isIdle: () => false,
|
||||
getPendingModelSwitch: () => pending,
|
||||
setPendingModelSwitch: (selection) => {
|
||||
pending = selection;
|
||||
},
|
||||
getActiveTurn: () => activeTurn,
|
||||
getAbortHandler: () => {
|
||||
return () => {
|
||||
aborted = true;
|
||||
};
|
||||
},
|
||||
hasAbortHandler: () => true,
|
||||
getActiveToolExecutions: () => 0,
|
||||
telegramPrefix: "[telegram]",
|
||||
allocateItemOrder: () => 5,
|
||||
allocateControlOrder: () => 6,
|
||||
appendQueuedItem: (item) => {
|
||||
appended.push(item);
|
||||
},
|
||||
updateStatus: () => {},
|
||||
});
|
||||
assert.equal(controller.triggerPendingAbort(undefined), true);
|
||||
assert.equal(aborted, true);
|
||||
assert.equal(pending, undefined);
|
||||
assert.equal(appended[0]?.queueOrder, 5);
|
||||
assert.equal(appended[0]?.queueLane, "control");
|
||||
});
|
||||
|
||||
test("Model-switch continuation turn stays control-lane and resume-oriented", () => {
|
||||
const createContinuationTurn =
|
||||
createTelegramModelSwitchContinuationTurnBuilder({
|
||||
allocateItemOrder: () => 5,
|
||||
allocateControlOrder: () => 6,
|
||||
});
|
||||
const builtTurn = createContinuationTurn({
|
||||
turn: { chatId: 1, replyToMessageId: 2 },
|
||||
selection: createModelTestSelection(),
|
||||
});
|
||||
assert.equal(builtTurn.queueOrder, 5);
|
||||
assert.equal(builtTurn.laneOrder, 6);
|
||||
const sourceTarget = { chatId: 1, threadId: 42 };
|
||||
const turn = buildTelegramModelSwitchContinuationTurn({
|
||||
turn: { chatId: 1, replyToMessageId: 2, target: sourceTarget },
|
||||
selection: createModelTestSelection(createModelTestModel(), "high"),
|
||||
telegramPrefix: "[telegram]",
|
||||
queueOrder: 3,
|
||||
laneOrder: 4,
|
||||
});
|
||||
assert.equal(turn.kind, "prompt");
|
||||
assert.equal(turn.queueLane, "control");
|
||||
assert.equal(turn.queueOrder, 3);
|
||||
assert.equal(turn.laneOrder, 4);
|
||||
assert.equal(turn.chatId, 1);
|
||||
assert.equal(turn.replyToMessageId, 2);
|
||||
assert.deepEqual(turn.target, { chatId: 1, threadId: 42 });
|
||||
assert.notEqual(turn.target, sourceTarget);
|
||||
assert.deepEqual(turn.sourceMessageIds, []);
|
||||
assert.equal(
|
||||
turn.historyText,
|
||||
"Continue interrupted Telegram request on openai/gpt-5",
|
||||
);
|
||||
assert.equal(turn.statusSummary, "↻ continue on gpt-5");
|
||||
assert.match(String(turn.content[0]?.type), /text/);
|
||||
assert.match(
|
||||
String((turn.content[0] as { text?: string } | undefined)?.text ?? ""),
|
||||
/Continue the interrupted previous Telegram request/,
|
||||
);
|
||||
});
|
||||
|
||||
test("Model-switch continuation queue runtime builds and appends control-lane turns", () => {
|
||||
const appended: PendingTelegramTurn[] = [];
|
||||
const queueContinuation = createTelegramModelSwitchContinuationQueueRuntime({
|
||||
allocateItemOrder: () => 3,
|
||||
allocateControlOrder: () => 4,
|
||||
appendQueuedItem: (item) => {
|
||||
appended.push(item);
|
||||
},
|
||||
});
|
||||
queueContinuation(
|
||||
createModelTestTurn(),
|
||||
createModelTestSelection(),
|
||||
undefined,
|
||||
);
|
||||
assert.equal(appended[0]?.queueOrder, 3);
|
||||
assert.equal(appended[0]?.laneOrder, 4);
|
||||
assert.equal(appended[0]?.queueLane, "control");
|
||||
assert.match(
|
||||
String((appended[0]?.content[0] as { text?: string } | undefined)?.text),
|
||||
/^\[telegram\] Continue/,
|
||||
);
|
||||
});
|
||||
|
||||
test("Model-switch continuation queue creates and appends continuation turns", () => {
|
||||
const appended: Array<{ item: PendingTelegramTurn; ctx: { id: string } }> =
|
||||
[];
|
||||
const queueContinuation = createTelegramModelSwitchContinuationQueue({
|
||||
createContinuationTurn: ({ turn, selection }) => ({
|
||||
...turn,
|
||||
kind: "prompt" as const,
|
||||
queueOrder: 1,
|
||||
queueLane: "control" as const,
|
||||
laneOrder: 2,
|
||||
sourceMessageIds: [],
|
||||
queuedAttachments: [],
|
||||
content: [{ type: "text" as const, text: selection.model.id }],
|
||||
historyText: "continue",
|
||||
statusSummary: "continue",
|
||||
}),
|
||||
appendQueuedItem: (item, ctx: { id: string }) => {
|
||||
appended.push({ item, ctx });
|
||||
},
|
||||
});
|
||||
queueContinuation(createModelTestTurn(), createModelTestSelection(), {
|
||||
id: "ctx",
|
||||
});
|
||||
assert.equal(appended[0]?.item.content[0]?.type, "text");
|
||||
assert.equal(appended[0]?.item.content[0]?.text, "gpt-5");
|
||||
assert.deepEqual(appended[0]?.ctx, { id: "ctx" });
|
||||
});
|
||||
|
||||
test("Continuation prompt stays Telegram-scoped and resume-oriented", () => {
|
||||
const text = buildTelegramModelSwitchContinuationText(
|
||||
"[telegram]",
|
||||
createModelTestModel(),
|
||||
"high",
|
||||
);
|
||||
assert.match(text, /^\[telegram\]/);
|
||||
assert.match(text, /Continue the interrupted previous Telegram request/);
|
||||
assert.match(text, /openai\/gpt-5/);
|
||||
assert.match(text, /thinking level \(high\)/);
|
||||
});
|
||||
Reference in new issue
Block a user