Skip to content

fix(chat-swarm-runtime): submit prompts through CDP input #188

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
James3014 wants to merge 1 commit into
base: main
Choose a base branch
from
Open

fix(chat-swarm-runtime): submit prompts through CDP input #188

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
26 changes: 20 additions & 6 deletions src/chat-swarm-runtime.test.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -214,7 +214,7 @@ test("CDP prompt delivery uses exact-target Input commands for visible editable
(driver as any).evaluate = async (_target: unknown, candidate: string) => {
expressions.push(candidate);
if (candidate.includes("const editable=")) return { ok: true, kind: "editable" };
if (candidate.includes("return {ready:Boolean")) return { ready: true };
if (candidate.includes("const button=")) return { ready: true, x: 320, y: 48 };
return { ok: true };
};
(driver as any).pageCommand = async (
Expand Down Expand Up @@ -242,21 +242,31 @@ test("CDP prompt delivery uses exact-target Input commands for visible editable
{ method: "Input.dispatchKeyEvent", params: { type: "keyDown", key: "Backspace", code: "Backspace", windowsVirtualKeyCode: 8 } },
{ method: "Input.dispatchKeyEvent", params: { type: "keyUp", key: "Backspace", code: "Backspace", windowsVirtualKeyCode: 8 } },
{ method: "Input.insertText", params: { text: "probe" } },
{ method: "Input.dispatchMouseEvent", params: { type: "mousePressed", x: 320, y: 48, button: "left", clickCount: 1 } },
{ method: "Input.dispatchMouseEvent", params: { type: "mouseReleased", x: 320, y: 48, button: "left", clickCount: 1 } },
]);
assert.match(expressions.at(-1)!, /button\.click\(\)/);
assert.doesNotMatch(expressions.join("\n"), /button\.click\(\)/);
assert.match(expressions.at(-1)!, /getBoundingClientRect/);
});

test("CDP prompt delivery keeps textarea fallback on the DOM value/input path", async () => {
const driver = cdpDriverForSelectorTest();
const expressions: string[] = [];
let pageCommands = 0;
const commands: Array<{ method: string; params: Record<string, unknown> }> = [];
(driver as any).evaluate = async (_target: unknown, candidate: string) => {
expressions.push(candidate);
if (candidate.includes("const editable=")) return { ok: true, kind: "textarea" };
if (candidate.includes("return {ready:Boolean")) return { ready: true };
if (candidate.includes("const button=")) return { ready: true, x: 160, y: 72 };
return { ok: true };
};
(driver as any).pageCommand = async () => { pageCommands += 1; };
(driver as any).pageCommand = async (
_target: unknown,
method: string,
params: Record<string, unknown>,
) => {
commands.push({ method, params });
return {};
};
await (driver as any).sendPromptToTarget(
{ id: "target-1", url: "https://chatgpt.com/" },
"probe",
Expand All @@ -265,7 +275,11 @@ test("CDP prompt delivery keeps textarea fallback on the DOM value/input path",
assert.match(expressions[0]!, /HTMLTextAreaElement\.prototype/);
assert.match(expressions[0]!, /setter\.call\(textarea,prompt\)/);
assert.match(expressions[0]!, /textarea\.dispatchEvent\(new Event\('input'/);
assert.equal(pageCommands, 0);
assert.deepEqual(commands, [
{ method: "Input.dispatchMouseEvent", params: { type: "mousePressed", x: 160, y: 72, button: "left", clickCount: 1 } },
{ method: "Input.dispatchMouseEvent", params: { type: "mouseReleased", x: 160, y: 72, button: "left", clickCount: 1 } },
]);
assert.doesNotMatch(expressions.join("\n"), /button\.click\(\)/);
});

test("CDP managed conversation waits for the initialization turn to become idle before exposing identity", async () => {
Expand Down
44 changes: 31 additions & 13 deletions src/chat-swarm-runtime.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2255,28 +2255,46 @@ export class CdpMacWebDriver implements MacWebDriver {
await this.pageCommand(target, "Input.insertText", { text: prompt });
}

let sendReady = false;
let sendPoint: { x: number; y: number } | undefined;
while (Date.now() < Date.parse(deadlineAt)) {
const readiness = await this.evaluate<{ ready: boolean; reason?: string }>(
const readiness = await this.evaluate<{
ready: boolean;
reason?: string;
x?: number;
y?: number;
}>(
target,
`(() => { const expectedUrl=${expectedUrl}; if(expectedUrl && location.href !== expectedUrl) return {ready:false,reason:'CHATGPT_CONVERSATION_IDENTITY_DRIFT'}; const button=document.querySelector('[data-testid="send-button"]') || [...document.querySelectorAll('button')].find(b => /send/i.test((b.getAttribute('aria-label')||b.textContent||''))); return {ready:Boolean(button && !button.disabled)}; })()`,
`(() => { const expectedUrl=${expectedUrl}; if(expectedUrl && location.href !== expectedUrl) return {ready:false,reason:'CHATGPT_CONVERSATION_IDENTITY_DRIFT'}; const button=document.querySelector('[data-testid="send-button"]') || [...document.querySelectorAll('button')].find(b => /send/i.test((b.getAttribute('aria-label')||b.textContent||''))); if(!button || button.disabled) return {ready:false}; const rect=button.getBoundingClientRect(); const style=getComputedStyle(button); if(rect.width <= 0 || rect.height <= 0 || style.display === 'none' || style.visibility === 'hidden') return {ready:false}; return {ready:true,x:rect.left + rect.width / 2,y:rect.top + rect.height / 2}; })()`,
);
if (readiness?.reason) throw new Error(readiness.reason);
if (readiness?.ready) {
sendReady = true;
if (
readiness?.ready &&
typeof readiness.x === "number" &&
Number.isFinite(readiness.x) &&
typeof readiness.y === "number" &&
Number.isFinite(readiness.y)
) {
sendPoint = { x: readiness.x, y: readiness.y };
break;
}
await new Promise((resolvePromise) => setTimeout(resolvePromise, 50));
}
if (!sendReady) throw new Error("send_button_missing_or_disabled");
if (!sendPoint) throw new Error("send_button_missing_or_disabled");

const sent = await this.evaluate<{ ok: boolean; reason?: string }>(
target,
`(() => { const expectedUrl=${expectedUrl}; if(expectedUrl && location.href !== expectedUrl) return {ok:false,reason:'CHATGPT_CONVERSATION_IDENTITY_DRIFT'}; const button=document.querySelector('[data-testid="send-button"]') || [...document.querySelectorAll('button')].find(b => /send/i.test((b.getAttribute('aria-label')||b.textContent||''))); if(!button || button.disabled) return {ok:false,reason:'send_button_missing_or_disabled'}; button.click(); return {ok:true}; })()`,
);
if (!sent?.ok) {
throw new Error(sent?.reason ?? "ChatGPT prompt delivery failed");
}
await this.pageCommand(target, "Input.dispatchMouseEvent", {
type: "mousePressed",
x: sendPoint.x,
y: sendPoint.y,
button: "left",
clickCount: 1,
});
await this.pageCommand(target, "Input.dispatchMouseEvent", {
type: "mouseReleased",
x: sendPoint.x,
y: sendPoint.y,
button: "left",
clickCount: 1,
});
if (Date.now() > Date.parse(deadlineAt)) {
throw new Error("prompt delivery exceeded deadline");
}
Expand Down
Loading