9 Commits

Author SHA1 Message Date
ABAPPLO
bc81e0daf1 merge: 合入 test(批量更新账号资料支持勾选字段) 2026-08-28 14:00:20 +08:00
ABAPPLO
ed13262166 feat: 批量更新账号资料支持勾选字段(含AI打标)
- 更新全部账号ID 升级为 批量更新账号资料:弹窗勾选 账号ID/昵称/粉丝数/IP属地/性别/简介/标签
- 标签勾选时调用 LLM 重新打标(覆盖已有标签),AI_TAGGING_TOOL 未配置自动跳过
- 只写入选中字段,爬取为空的字段跳过不清空;演示账号(xhs- 前缀)不参与
- .env.app.example 为 SEED_DEMO_DATA 补充注释
2026-08-28 14:00:07 +08:00
ABAPPLO
b882a5c7d7 merge: 合入 test(KOC资源列表分页展示) 2026-08-27 19:18:00 +08:00
ABAPPLO
8e4e3de974 feat: KOC资源列表分页展示(每页15个) 2026-08-27 19:17:53 +08:00
ABAPPLO
0815967ca7 merge: 合入 test(一键爬取更新全部账号ID) 2026-08-27 18:55:00 +08:00
ABAPPLO
dfc7dc217d feat: KOC资源页新增一键爬取更新全部账号ID 2026-08-27 18:54:55 +08:00
ABAPPLO
d8a09dcb8b merge: 合入 test(KOC资源卡片手动修改账号ID) 2026-08-27 17:57:04 +08:00
ABAPPLO
5614a13ebb feat: KOC资源卡片支持手动修改账号ID 2026-08-27 17:50:44 +08:00
1567021d57 Merge pull request 'feat: 企微客户群发、LLM 账号打标与生产部署编排(test → main)' (#5) from test into main 2026-08-27 08:33:14 +00:00
6 changed files with 897 additions and 7 deletions

View File

@@ -54,5 +54,6 @@ WECOM_NOTIFY_ENABLED=true
# ── 调度与入口 ────────────────────────────────────────────
# 每天北京时间 09:00 自动执行采集计划,生产保持 true。
ENABLE_SCHEDULER=true
# 是否在空库时写入演示数据,生产保持 false。
SEED_DEMO_DATA=false
HTTP_PORT=80

View File

@@ -74,6 +74,28 @@ type Account = {
last_seen_at: string;
};
type FieldRefreshStatus = {
status: "running" | "done";
fields: string[];
total: number;
processed: number;
updated: number;
failed: number;
skipped: number;
startedAt: string;
finishedAt: string | null;
};
const REFRESH_FIELD_OPTIONS = [
{ key: "publicAccountId", label: "账号ID(小红书号/抖音号)" },
{ key: "nickname", label: "昵称" },
{ key: "followers", label: "粉丝数" },
{ key: "ipLocation", label: "IP属地" },
{ key: "gender", label: "性别" },
{ key: "bio", label: "简介" },
{ key: "tags", label: "标签(AI 重新打标,覆盖已有标签)" },
] as const;
type ResourceImportPreview = {
summary: { total: number; create: number; update: number; error: number };
rows: Array<{
@@ -3160,6 +3182,22 @@ function ScreenshotTaskTable({ distributions }: { distributions: Distribution[]
);
}
const RESOURCE_PAGE_SIZE = 15;
function pageList(current: number, total: number): Array<number | "gap"> {
if (total <= 7) {
return Array.from({ length: total }, (_, index) => index + 1);
}
const items: Array<number | "gap"> = [1];
const start = Math.max(2, current - 1);
const end = Math.min(total - 1, current + 1);
if (start > 2) items.push("gap");
for (let page = start; page <= end; page += 1) items.push(page);
if (end < total - 1) items.push("gap");
items.push(total);
return items;
}
function ResourcesPage({
accounts,
distributions,
@@ -3177,6 +3215,7 @@ function ResourcesPage({
const [platformFilter, setPlatformFilter] = useState("全部平台");
const [ipFilter, setIpFilter] = useState("");
const [sourceFilter, setSourceFilter] = useState("");
const [page, setPage] = useState(1);
const [importOpen, setImportOpen] = useState(false);
const [importMode, setImportMode] = useState<"bulk" | "single">("bulk");
const [importFile, setImportFile] = useState<File | null>(null);
@@ -3196,6 +3235,24 @@ function ResourcesPage({
});
const [singleWorking, setSingleWorking] = useState(false);
const [singleError, setSingleError] = useState("");
const [editAccount, setEditAccount] = useState<Account | null>(null);
const [editAccountId, setEditAccountId] = useState("");
const [editWorking, setEditWorking] = useState(false);
const [editError, setEditError] = useState("");
const [idRefresh, setIdRefresh] = useState<FieldRefreshStatus | null>(null);
const [idRefreshStarting, setIdRefreshStarting] = useState(false);
const idRefreshNotifiedRef = useRef(true);
const [idRefreshError, setIdRefreshError] = useState("");
const [fieldPickOpen, setFieldPickOpen] = useState(false);
const [fieldChecks, setFieldChecks] = useState<Record<string, boolean>>(
() =>
Object.fromEntries(
REFRESH_FIELD_OPTIONS.map((option) => [
option.key,
option.key === "publicAccountId",
]),
),
);
const resourceAccounts = useMemo(() => {
const deliveriesByAccount = new Map<string, Distribution[]>();
distributions.forEach((distribution) => {
@@ -3288,11 +3345,21 @@ function ResourcesPage({
platformFilter !== "全部平台" ||
Boolean(ipFilter.trim()) ||
Boolean(sourceFilter.trim());
const totalPages = Math.max(
1,
Math.ceil(filteredAccounts.length / RESOURCE_PAGE_SIZE),
);
const currentPage = Math.min(page, totalPages);
const pagedAccounts = filteredAccounts.slice(
(currentPage - 1) * RESOURCE_PAGE_SIZE,
currentPage * RESOURCE_PAGE_SIZE,
);
const clearFilters = () => {
setQuery("");
setPlatformFilter("全部平台");
setIpFilter("");
setSourceFilter("");
setPage(1);
};
const closeImport = () => {
if (importWorking || singleWorking) return;
@@ -3403,6 +3470,126 @@ function ResourcesPage({
setImportWorking(false);
}
};
const closeAccountEdit = () => {
if (editWorking) return;
setEditAccount(null);
setEditAccountId("");
setEditError("");
};
const submitAccountEdit = async () => {
if (!editAccount) return;
const value = editAccountId.trim();
if (!value) {
setEditError("请填写新的账号ID");
return;
}
if (!/^[\p{L}\p{N}._-]{2,80}$/u.test(value)) {
setEditError("账号ID需为 2-80 位字母、数字或 . _ - 组合");
return;
}
try {
setEditWorking(true);
setEditError("");
const response = await fetch("/api/action", {
method: "POST",
headers: { "Content-Type": "application/json" },
body: JSON.stringify({
action: "set_public_account_ids",
items: [{ accountId: editAccount.id, publicAccountId: value }],
}),
});
const result = await readApiResponse<{ error?: string }>(response, "保存失败");
if (!response.ok) throw new Error(result.error || "保存失败");
const label = editAccount.nickname || editAccount.public_account_id || "该账号";
setEditAccount(null);
setEditAccountId("");
await onImported(`已更新 ${label} 的账号ID`);
} catch (reason) {
setEditError(reason instanceof Error ? reason.message : "保存失败");
} finally {
setEditWorking(false);
}
};
const runningIdRefresh = idRefresh?.status === "running";
const fetchIdRefreshStatus = useCallback(async () => {
const response = await fetch("/api/action", {
method: "POST",
headers: { "Content-Type": "application/json" },
body: JSON.stringify({ action: "account_field_refresh_status" }),
});
const result = await readApiResponse<{ refresh: FieldRefreshStatus | null }>(
response,
"查询失败",
);
if (!response.ok) throw new Error(result.error || "查询失败");
return result.refresh;
}, []);
const applyIdRefreshStatus = useCallback(
(status: FieldRefreshStatus) => {
if (status.status === "done" && !idRefreshNotifiedRef.current) {
idRefreshNotifiedRef.current = true;
const summary = `账号资料更新完成:成功 ${status.updated} 个${
status.skipped ? `,未识别 ${status.skipped} 个` : ""
}${status.failed ? `,失败 ${status.failed} 个` : ""}`;
void onImported(summary);
}
setIdRefresh(status);
},
[onImported],
);
useEffect(() => {
void fetchIdRefreshStatus()
.then((status) => {
if (status?.status === "running") {
idRefreshNotifiedRef.current = false;
applyIdRefreshStatus(status);
}
})
.catch(() => undefined);
}, [applyIdRefreshStatus, fetchIdRefreshStatus]);
useEffect(() => {
if (!runningIdRefresh) return;
const timer = window.setInterval(() => {
void fetchIdRefreshStatus()
.then((status) => {
if (status) applyIdRefreshStatus(status);
})
.catch(() => undefined);
}, 3000);
return () => window.clearInterval(timer);
}, [applyIdRefreshStatus, fetchIdRefreshStatus, runningIdRefresh]);
const closeFieldPick = () => {
if (idRefreshStarting) return;
setFieldPickOpen(false);
};
const startIdRefresh = async () => {
if (runningIdRefresh || idRefreshStarting) return;
const fields = REFRESH_FIELD_OPTIONS.filter(
(option) => fieldChecks[option.key],
).map((option) => option.key);
if (fields.length === 0) return;
try {
setIdRefreshStarting(true);
setIdRefreshError("");
const response = await fetch("/api/action", {
method: "POST",
headers: { "Content-Type": "application/json" },
body: JSON.stringify({ action: "refresh_account_fields", fields }),
});
const result = await readApiResponse<{ refresh: FieldRefreshStatus }>(
response,
"启动失败",
);
if (!response.ok) throw new Error(result.error || "启动失败");
idRefreshNotifiedRef.current = false;
setIdRefresh(result.refresh);
setFieldPickOpen(false);
} catch (reason) {
setIdRefreshError(reason instanceof Error ? reason.message : "启动失败");
} finally {
setIdRefreshStarting(false);
}
};
return (
<div className="stack">
<section className="resource-intro">
@@ -3432,7 +3619,10 @@ function ResourcesPage({
<div className="filter-chips">
<button
className={platformFilter === "全部平台" ? "active" : ""}
onClick={() => setPlatformFilter("全部平台")}
onClick={() => {
setPlatformFilter("全部平台");
setPage(1);
}}
>
全部平台
</button>
@@ -3440,7 +3630,10 @@ function ResourcesPage({
<button
className={platformFilter === platform ? "active" : ""}
key={platform}
onClick={() => setPlatformFilter(platform)}
onClick={() => {
setPlatformFilter(platform);
setPage(1);
}}
>
<PlatformBadge platform={platform} compact />
</button>
@@ -3453,7 +3646,10 @@ function ResourcesPage({
<span aria-hidden="true">⌕</span>
<input
aria-label="搜索账号名称、账号ID或当前联系人"
onChange={(event) => setQuery(event.target.value)}
onChange={(event) => {
setQuery(event.target.value);
setPage(1);
}}
placeholder="搜索账号名称 / 账号ID / 当前联系人 / 标签"
type="search"
value={query}
@@ -3464,7 +3660,10 @@ function ResourcesPage({
<input
aria-label="模糊搜索IP地区"
list="resource-ip-options"
onChange={(event) => setIpFilter(event.target.value)}
onChange={(event) => {
setIpFilter(event.target.value);
setPage(1);
}}
placeholder="搜索IP地区"
type="search"
value={ipFilter}
@@ -3478,7 +3677,10 @@ function ResourcesPage({
<input
aria-label="模糊搜索合作来源"
list="resource-source-options"
onChange={(event) => setSourceFilter(event.target.value)}
onChange={(event) => {
setSourceFilter(event.target.value);
setPage(1);
}}
placeholder="搜索合作来源"
type="search"
value={sourceFilter}
@@ -3497,7 +3699,19 @@ function ResourcesPage({
清空筛选
</button>
<div className="resource-toolbar-summary">
<span>共 <b>{filteredAccounts.length}</b> 个结果</span>
<span>共 <b>{filteredAccounts.length}</b> 个结果{totalPages > 1 ? ` · 第 ${currentPage}/${totalPages} 页` : ""}</span>
<button
className="export-data-button"
disabled={runningIdRefresh || idRefreshStarting}
title="爬取全部账号主页,按选择的字段批量更新账号资料"
onClick={() => setFieldPickOpen(true)}
>
{idRefreshStarting
? "正在启动…"
: runningIdRefresh
? `更新中 ${idRefresh?.processed ?? 0}/${idRefresh?.total ?? 0}`
: "批量更新账号资料"}
</button>
<button
className="export-data-button"
disabled={exporting || filteredAccounts.length === 0}
@@ -3508,9 +3722,10 @@ function ResourcesPage({
{exporting ? "正在导出…" : "导出筛选结果"}
</button>
</div>
{idRefreshError && <div className="form-error">{idRefreshError}</div>}
</div>
<div className="resource-grid">
{filteredAccounts.map(({ account, sources, partnerManagedOnly }) => {
{pagedAccounts.map(({ account, sources, partnerManagedOnly }) => {
const profileLink = resourceProfileLink(account);
const accountTags = (account.tags || "")
.split(/[,,、;;|]/)
@@ -3548,6 +3763,18 @@ function ResourcesPage({
<div className="resource-account-number">
<span>{accountIdLabel}</span>
<strong>{account.public_account_id || "待识别"}</strong>
<button
type="button"
className="account-id-edit-button"
title="手动修改账号ID"
onClick={() => {
setEditAccount(account);
setEditAccountId(account.public_account_id || "");
setEditError("");
}}
>
修改
</button>
</div>
<div className="resource-platform-line">
<PlatformBadge platform={account.platform} compact />
@@ -3600,7 +3827,149 @@ function ResourcesPage({
</button>
</div>
)}
{totalPages > 1 && (
<div className="resource-pagination">
<button
disabled={currentPage === 1}
onClick={() => setPage(currentPage - 1)}
>
上一页
</button>
{pageList(currentPage, totalPages).map((item, index) =>
item === "gap" ? (
<span className="gap" key={`gap-${index}`}>…</span>
) : (
<button
className={item === currentPage ? "active" : ""}
key={item}
onClick={() => setPage(item)}
>
{item}
</button>
),
)}
<button
disabled={currentPage === totalPages}
onClick={() => setPage(currentPage + 1)}
>
下一页
</button>
</div>
)}
</section>
{fieldPickOpen && (
<div className="modal-backdrop" role="presentation" onMouseDown={closeFieldPick}>
<div
className="modal-card compact"
role="dialog"
aria-modal="true"
aria-labelledby="field-pick-title"
onMouseDown={(event) => event.stopPropagation()}
>
<div className="modal-heading">
<div>
<p className="eyebrow">批量更新</p>
<h2 id="field-pick-title">选择要更新的字段</h2>
</div>
<button type="button" onClick={closeFieldPick} aria-label="关闭">×</button>
</div>
<p className="field-pick-note">
将逐个爬取全部账号主页(小红书 + 抖音),只写入选中的字段;
爬取为空的字段自动跳过,不会清空已有数据。账号较多时需要几分钟。
</p>
<div className="field-pick-list">
{REFRESH_FIELD_OPTIONS.map((option) => (
<label className="field-pick-item" key={option.key}>
<input
checked={Boolean(fieldChecks[option.key])}
onChange={(event) =>
setFieldChecks((checks) => ({
...checks,
[option.key]: event.target.checked,
}))
}
type="checkbox"
/>
<span>{option.label}</span>
</label>
))}
</div>
{idRefreshError && <div className="form-error">{idRefreshError}</div>}
<div className="modal-actions">
<button type="button" className="ghost-button" onClick={closeFieldPick}>取消</button>
<button
type="button"
className="primary-button"
disabled={
idRefreshStarting ||
runningIdRefresh ||
!REFRESH_FIELD_OPTIONS.some((option) => fieldChecks[option.key])
}
onClick={() => void startIdRefresh()}
>
{idRefreshStarting
? "正在启动…"
: runningIdRefresh
? `更新中 ${idRefresh?.processed ?? 0}/${idRefresh?.total ?? 0}`
: "开始更新"}
</button>
</div>
</div>
</div>
)}
{editAccount && (
<div className="modal-backdrop" role="presentation" onMouseDown={closeAccountEdit}>
<div
className="modal-card compact"
role="dialog"
aria-modal="true"
aria-labelledby="account-id-edit-title"
onMouseDown={(event) => event.stopPropagation()}
>
<div className="modal-heading">
<div>
<p className="eyebrow">{editAccount.nickname || "账号"}</p>
<h2 id="account-id-edit-title">修改账号ID</h2>
</div>
<button type="button" onClick={closeAccountEdit} aria-label="关闭">×</button>
</div>
<div className="account-id-edit-form">
<label>
<span>
{editAccount.platform === "小红书"
? "小红书号"
: editAccount.platform === "抖音"
? "抖音号"
: "账号ID"}
</span>
<input
autoFocus
type="text"
value={editAccountId}
placeholder="2-80 位字母、数字或 . _ -"
onChange={(event) => setEditAccountId(event.target.value)}
onKeyDown={(event) => {
if (event.key === "Enter") void submitAccountEdit();
}}
/>
<small>只更新账号ID,昵称、粉丝数等其他资料保持不变</small>
</label>
{editError && <div className="form-error">{editError}</div>}
</div>
<div className="modal-actions">
<button type="button" className="ghost-button" onClick={closeAccountEdit}>取消</button>
<button
type="button"
className="primary-button"
disabled={editWorking || !editAccountId.trim()}
onClick={() => void submitAccountEdit()}
>
{editWorking ? "保存中…" : "保存"}
</button>
</div>
</div>
</div>
)}
{importOpen && (
<div className="modal-backdrop" role="presentation" onMouseDown={closeImport}>
<div

View File

@@ -54,6 +54,10 @@ import {
} from "../../../lib/wecom-group-push-service";
import { isManagerRequest } from "../../../lib/user-auth";
import { extractPublishUrl } from "../../../lib/publish-url";
import {
getAccountFieldRefreshStatus,
startAccountFieldRefresh,
} from "../../../lib/account-field-refresh-service";
type ActionBody = {
action?: string;
@@ -683,6 +687,15 @@ export async function POST(request: Request) {
const taskId = String(body.taskId ?? "").trim();
const pushes = await listGroupPushes(db, taskId || undefined);
return Response.json({ pushes });
} else if (body.action === "refresh_account_fields") {
const status = startAccountFieldRefresh(
db,
resolveCollectionMcpConfig(env as unknown as CollectionMcpBindings),
body.fields,
);
return Response.json({ refresh: status });
} else if (body.action === "account_field_refresh_status") {
return Response.json({ refresh: getAccountFieldRefreshStatus() });
} else if (body.action === "wecom_group_chats") {
const groups = await listGroupChatRows(db);
return Response.json({ groups });

View File

@@ -2577,6 +2577,49 @@ a {
font-size: 11px;
}
.resource-pagination {
display: flex;
align-items: center;
justify-content: center;
gap: 6px;
padding-top: 18px;
}
.resource-pagination button {
min-width: 30px;
height: 30px;
padding: 0 10px;
border: 1px solid transparent;
border-radius: 8px;
color: #66736f;
background: transparent;
font-size: 11px;
font-weight: 600;
cursor: pointer;
}
.resource-pagination button:hover:not(:disabled) {
background: #f0f4f2;
}
.resource-pagination button.active {
color: #34443f;
background: white;
border-color: #dce2de;
box-shadow: 0 1px 3px rgb(18 31 28 / 0.08);
}
.resource-pagination button:disabled {
color: #b7c1bd;
cursor: default;
}
.resource-pagination .gap {
color: #9aa4a0;
font-size: 11px;
padding: 0 2px;
}
.resource-import-modal {
width: min(780px, calc(100vw - 32px));
max-height: min(86vh, 820px);
@@ -2972,6 +3015,76 @@ a {
white-space: nowrap;
}
.account-id-edit-button {
flex: 0 0 auto;
padding: 1px 6px;
border: 1px solid #dce2de;
border-radius: 999px;
color: #8a978f;
background: transparent;
font-size: 9px;
line-height: 1.5;
cursor: pointer;
}
.account-id-edit-button:hover {
border-color: #70ad95;
color: #40806a;
}
.account-id-edit-form {
display: flex;
flex-direction: column;
gap: 4px;
}
.account-id-edit-form small {
color: #9aa4a0;
font-size: 10px;
}
.field-pick-note {
margin: 0 0 14px;
color: #8a978f;
font-size: 11px;
line-height: 1.7;
}
.field-pick-list {
display: flex;
flex-direction: column;
gap: 6px;
margin-bottom: 12px;
}
.field-pick-item {
flex-direction: row;
align-items: center;
gap: 10px;
margin-bottom: 0;
padding: 9px 11px;
border: 1px solid #eef1ef;
border-radius: 9px;
cursor: pointer;
}
.field-pick-item:hover {
border-color: #d8e2dd;
}
.field-pick-item input {
width: 15px;
height: 15px;
padding: 0;
accent-color: #70ad95;
cursor: pointer;
}
.field-pick-item span {
color: #4a5854;
font-size: 12px;
}
.resource-platform-line {
display: flex;
min-width: 0;

View File

@@ -0,0 +1,253 @@
import { runInBackground } from "./background";
import {
generateAccountTagsFromMcp,
resolveProfileDetailsFromMcp,
type AccountTaggingEnv,
type CollectionMcpConfig,
} from "./mcp-collection-client";
import { getRuntimeEnv } from "./runtime-env";
import type { DatabaseClient } from "./database";
export const ACCOUNT_REFRESH_FIELDS = [
"publicAccountId",
"nickname",
"followers",
"ipLocation",
"gender",
"bio",
"tags",
] as const;
export type AccountRefreshField = (typeof ACCOUNT_REFRESH_FIELDS)[number];
export const ACCOUNT_REFRESH_FIELD_LABELS: Record<AccountRefreshField, string> = {
publicAccountId: "账号ID(小红书号/抖音号)",
nickname: "昵称",
followers: "粉丝数",
ipLocation: "IP属地",
gender: "性别",
bio: "简介",
tags: "标签(AI 重新打标,覆盖已有标签)",
};
const FIELD_COLUMNS: Record<AccountRefreshField, string> = {
publicAccountId: "public_account_id",
nickname: "nickname",
followers: "followers",
ipLocation: "ip_location",
gender: "gender",
bio: "bio",
tags: "tags",
};
export function sanitizeRefreshFields(value: unknown): AccountRefreshField[] {
const requested = Array.isArray(value) ? value : [];
const fields = ACCOUNT_REFRESH_FIELDS.filter((field) =>
requested.includes(field),
);
return fields.length > 0 ? fields : ["publicAccountId"];
}
export type AccountFieldRefreshStatus = {
status: "running" | "done";
fields: AccountRefreshField[];
total: number;
processed: number;
updated: number;
failed: number;
skipped: number;
startedAt: string;
finishedAt: string | null;
};
type RefreshTargetRow = {
id: string;
platform: string;
platform_uid: string;
profile_url: string;
};
export type ScrapedProfile = {
nickname: string | null;
redId: string | null;
followers: number | null;
ipLocation: string | null;
gender: string;
bio: string;
recentNoteTitles: string[];
providerTags: string[];
};
export type ProfileScraper = (
profileUrl: string,
platform: "小红书" | "抖音",
) => Promise<ScrapedProfile>;
export type TagGenerator = (input: {
platform: string;
nickname: string;
bio: string;
gender: string;
ipLocation: string;
followers: number | null;
providerTags: string[];
recentNoteTitles: string[];
}) => Promise<string[]>;
function scrapedFieldValues(
details: ScrapedProfile,
): Partial<Record<AccountRefreshField, string | number>> {
const values: Partial<Record<AccountRefreshField, string | number>> = {};
const redId = details.redId?.trim();
if (redId) values.publicAccountId = redId;
const nickname = details.nickname?.trim();
if (nickname) values.nickname = nickname;
if (details.followers !== null && Number.isFinite(details.followers)) {
values.followers = details.followers;
}
const ipLocation = details.ipLocation?.trim();
if (ipLocation) values.ipLocation = ipLocation;
if (details.gender === "男" || details.gender === "女") {
values.gender = details.gender;
}
const bio = details.bio?.trim();
if (bio) values.bio = bio;
return values;
}
let currentJob: AccountFieldRefreshStatus | null = null;
export function getAccountFieldRefreshStatus(): AccountFieldRefreshStatus | null {
return currentJob ? { ...currentJob } : null;
}
export function startAccountFieldRefresh(
db: DatabaseClient,
config: CollectionMcpConfig,
requestedFields: unknown,
): AccountFieldRefreshStatus {
if (currentJob?.status === "running") {
return { ...currentJob };
}
const fields = sanitizeRefreshFields(requestedFields);
const job: AccountFieldRefreshStatus = {
status: "running",
fields,
total: 0,
processed: 0,
updated: 0,
failed: 0,
skipped: 0,
startedAt: new Date().toISOString(),
finishedAt: null,
};
currentJob = job;
const scraper: ProfileScraper = (profileUrl, platform) =>
resolveProfileDetailsFromMcp(profileUrl, platform, config);
const tagger: TagGenerator = (input) =>
generateAccountTagsFromMcp(
input,
config,
getRuntimeEnv() as AccountTaggingEnv,
);
runInBackground(
executeAccountFieldRefresh(job, db, scraper, tagger),
"account field refresh",
);
return { ...job };
}
async function executeAccountFieldRefresh(
job: AccountFieldRefreshStatus,
db: DatabaseClient,
scraper: ProfileScraper,
tagger: TagGenerator,
) {
const rows = await db
.prepare(
`SELECT id, platform, platform_uid, profile_url
FROM accounts
WHERE profile_url IS NOT NULL AND profile_url != ''
AND platform IN ('小红书', '抖音')`,
)
.all<RefreshTargetRow>();
const targets = rows.results.filter(
(row) => !row.platform_uid.startsWith("xhs-"),
);
job.total = targets.length;
const fields = job.fields;
const wantTags = fields.includes("tags");
let cursor = 0;
const worker = async () => {
while (cursor < targets.length) {
const row = targets[cursor];
cursor += 1;
try {
const details = await scraper(
row.profile_url,
row.platform === "抖音" ? "抖音" : "小红书",
);
const values = scrapedFieldValues(details);
if (wantTags) {
const tags = await tagger({
platform: row.platform,
nickname: details.nickname?.trim() || "",
bio: details.bio,
gender: details.gender,
ipLocation: details.ipLocation ?? "",
followers: details.followers,
providerTags: details.providerTags,
recentNoteTitles: details.recentNoteTitles,
}).catch(() => []);
if (tags.length > 0) values.tags = tags.join(",");
}
const updates = fields.filter(
(field) => values[field] !== undefined,
);
if (updates.length === 0) {
job.skipped += 1;
} else {
const assignments = updates
.map((field) => `${FIELD_COLUMNS[field]} = ?`)
.join(", ");
await db
.prepare(`UPDATE accounts SET ${assignments} WHERE id = ?`)
.bind(
...updates.map((field) => values[field]),
row.id,
)
.run();
job.updated += 1;
}
} catch {
job.failed += 1;
}
job.processed += 1;
}
};
await Promise.all(Array.from({ length: Math.min(4, targets.length) }, worker));
job.status = "done";
job.finishedAt = new Date().toISOString();
}
export async function runAccountFieldRefresh(
db: DatabaseClient,
scraper: ProfileScraper,
tagger: TagGenerator,
fields: AccountRefreshField[],
): Promise<AccountFieldRefreshStatus> {
const job: AccountFieldRefreshStatus = {
status: "running",
fields: sanitizeRefreshFields(fields),
total: 0,
processed: 0,
updated: 0,
failed: 0,
skipped: 0,
startedAt: new Date().toISOString(),
finishedAt: null,
};
await executeAccountFieldRefresh(job, db, scraper, tagger);
return job;
}

View File

@@ -0,0 +1,141 @@
import assert from "node:assert/strict";
import test from "node:test";
import {
runAccountFieldRefresh,
sanitizeRefreshFields,
} from "../lib/account-field-refresh-service.ts";
function fakeDb(accounts) {
const updates = [];
return {
db: {
prepare: (sql) => {
if (sql.startsWith("SELECT")) {
return {
all: async () => ({
results: accounts.map((account) => ({
id: account.id,
platform: account.platform,
platform_uid: account.platform_uid,
profile_url: account.profile_url,
})),
success: true,
meta: { changes: 0, last_row_id: 0 },
}),
};
}
assert.match(sql, /^UPDATE accounts SET [a-z_, ?= ]+ WHERE id = \?$/);
return {
bind: (...args) => ({
run: async () => {
updates.push({ sql, args });
return { success: true, meta: { changes: 1, last_row_id: 0 } };
},
}),
};
},
},
updates,
};
}
const profile = (overrides = {}) => ({
nickname: "昵称A",
redId: "id-a",
followers: 123,
ipLocation: "上海",
gender: "女",
bio: "简介A",
recentNoteTitles: [],
providerTags: [],
...overrides,
});
const noTag = async () => [];
test("writes only the selected fields as dynamic columns", async () => {
const accounts = [
{ id: "a1", platform: "小红书", platform_uid: "uid-1", profile_url: "https://www.xiaohongshu.com/user/profile/a1" },
];
const { db, updates } = fakeDb(accounts);
const status = await runAccountFieldRefresh(
db,
async () => profile(),
noTag,
["nickname", "ipLocation"],
);
assert.equal(status.status, "done");
assert.equal(status.updated, 1);
assert.equal(status.fields.join(","), "nickname,ipLocation");
assert.equal(updates.length, 1);
assert.match(updates[0].sql, /SET nickname = \?, ip_location = \?/);
assert.doesNotMatch(updates[0].sql, /public_account_id|followers|gender|bio|tags/);
assert.deepEqual(updates[0].args, ["昵称A", "上海", "a1"]);
});
test("tags field uses the tagger output and joins with commas", async () => {
const accounts = [
{ id: "a1", platform: "小红书", platform_uid: "uid-1", profile_url: "https://www.xiaohongshu.com/user/profile/a1" },
];
const { db, updates } = fakeDb(accounts);
const status = await runAccountFieldRefresh(
db,
async () => profile({ redId: "" }),
async () => ["美妆", "学生党"],
["tags"],
);
assert.equal(status.updated, 1);
assert.match(updates[0].sql, /SET tags = \?/);
assert.deepEqual(updates[0].args, ["美妆,学生党", "a1"]);
});
test("empty scrape values and empty tagger result are skipped without writing", async () => {
const accounts = [
{ id: "a1", platform: "小红书", platform_uid: "uid-1", profile_url: "https://www.xiaohongshu.com/user/profile/a1" },
];
const { db, updates } = fakeDb(accounts);
const status = await runAccountFieldRefresh(
db,
async () => profile({ redId: " ", nickname: null, followers: null, ipLocation: "", gender: "", bio: "" }),
noTag,
["publicAccountId", "nickname", "followers", "ipLocation", "gender", "bio", "tags"],
);
assert.equal(status.updated, 0);
assert.equal(status.skipped, 1);
assert.equal(updates.length, 0);
});
test("scrape failures count as failed and demo accounts are excluded", async () => {
const accounts = [
{ id: "boom", platform: "小红书", platform_uid: "uid-1", profile_url: "https://www.xiaohongshu.com/user/profile/boom" },
{ id: "demo", platform: "小红书", platform_uid: "xhs-8af3", profile_url: "https://www.xiaohongshu.com/user/profile/demo" },
{ id: "ok", platform: "抖音", platform_uid: "uid-2", profile_url: "https://www.douyin.com/user/ok" },
];
const { db, updates } = fakeDb(accounts);
const status = await runAccountFieldRefresh(
db,
async (url) => {
if (url.includes("boom")) throw new Error("mcp down");
return profile({ redId: "ok-2" });
},
noTag,
["publicAccountId"],
);
assert.equal(status.total, 2);
assert.equal(status.updated, 1);
assert.equal(status.failed, 1);
assert.deepEqual(updates[0].args.slice(-1), ["ok"]);
});
test("sanitizeRefreshFields filters unknown values and falls back to account id", () => {
assert.deepEqual(sanitizeRefreshFields(["nickname", "hacked", "tags"]), [
"nickname",
"tags",
]);
assert.deepEqual(sanitizeRefreshFields([]), ["publicAccountId"]);
assert.deepEqual(sanitizeRefreshFields("bad"), ["publicAccountId"]);
});