Add message counting and date-bounded scrapes

This commit is contained in:
Fisto
2026-09-16 12:06:34 +02:00
parent f14f067413
commit fae754fba0
21 changed files with 410 additions and 22 deletions
+30 -3
View File
@@ -778,10 +778,12 @@ function buildServer(): McpServer {
downloadDir: z
.string()
.optional()
.describe("Absolute destination for downloads; previews stay local")
.describe("Absolute destination for downloads; previews stay local"),
after: z.string().optional().describe("From-date ISO; the walk stops below it"),
before: z.string().optional().describe("To-date ISO; the walk starts just below it")
}
},
async ({ channelId, guildId, mode, backend, confirm, downloadDir }) =>
async ({ channelId, guildId, mode, backend, confirm, downloadDir, after, before }) =>
jsonText(
await media.startScrape({
channelId,
@@ -789,7 +791,32 @@ function buildServer(): McpServer {
mode,
backend,
confirm,
downloadDir: downloadDir ?? null
downloadDir: downloadDir ?? null,
after: after ?? null,
before: before ?? null
})
)
);
server.registerTool(
"count_media_messages",
{
description:
"Discord search total for a channel or date range. Guild channels with slopbot only; DMs and group chats are not available.",
inputSchema: {
channelId: z.string(),
guildId: z.string().optional(),
after: z.string().optional().describe("From-date ISO"),
before: z.string().optional().describe("To-date ISO")
}
},
async ({ channelId, guildId, after, before }) =>
jsonText(
await media.countMessages({
channelId,
guildId: guildId ?? null,
after: after ?? null,
before: before ?? null
})
)
);
+54 -1
View File
@@ -14,6 +14,7 @@ import {
ScrapeEngine,
ScrapeStore,
scrapeDir,
snowflakeFromIso,
resolveDbRoot,
type MediaKind,
type MediaRow,
@@ -22,12 +23,14 @@ import {
} from "../../media-scrape/src/index.js";
import { PuppetBackend } from "../../media-scrape/src/puppet-backend.js";
import {
countMessages,
loadSlopbotConfig,
slopbotFetcher,
slopbotStatus,
SlopbotBackend,
startSlopbot,
type SlopbotFetcher,
type SlopbotRest,
type SlopbotStatus
} from "../../media-scrape/slopbot/index.js";
import type { CoreClient } from "./core-client.js";
@@ -51,6 +54,10 @@ export interface StartScrapeInput {
confirm?: boolean;
/** Absolute destination for downloaded files. Null keeps them under the scrape dir. */
downloadDir?: string | null;
/** From-date as ISO; the walk stops below it. */
after?: string | null;
/** To-date as ISO; the walk starts just below it. */
before?: string | null;
}
export interface MediaErrorEntry {
@@ -295,6 +302,8 @@ export class MediaService {
try {
const mode = input.mode ?? "links";
const afterId = input.after ? snowflakeFromIso(input.after) : null;
const beforeId = input.before ? snowflakeFromIso(input.before) : null;
let backend: ScrapeBackend;
if (input.backend === "puppet") {
@@ -314,12 +323,28 @@ export class MediaService {
: this.puppetBackend();
}
// Best effort: one search request gives the total for the range.
let estimatedMessages: number | null = null;
if (backend.name === "slopbot" && input.guildId && this.bot) {
estimatedMessages = await countMessages(this.bot.rest as SlopbotRest, {
channelId: input.channelId,
guildId: input.guildId,
afterId,
beforeId
});
}
return this.engine.start({
channelId: input.channelId,
guildId: input.guildId ?? null,
mode,
backend,
downloadDir: input.downloadDir ?? null
downloadDir: input.downloadDir ?? null,
afterId,
beforeId,
fromDate: input.after ?? null,
toDate: input.before ?? null,
estimatedMessages
});
} catch (error) {
this.recordError("scrape", error instanceof Error ? error.message : String(error));
@@ -373,6 +398,34 @@ export class MediaService {
};
}
/** Discord search total for a channel or date range. Slopbot (guild) channels only. */
async countMessages(input: {
channelId: string;
guildId?: string | null;
after?: string | null;
before?: string | null;
}): Promise<{ count: number | null; reason?: string }> {
if (!input.guildId) {
return { count: null, reason: "not available for DMs or group chats" };
}
const available = await this.ensureSlopbot();
if (!available || !this.bot) {
return { count: null, reason: "slopbot is not connected" };
}
const count = await countMessages(this.bot.rest as SlopbotRest, {
channelId: input.channelId,
guildId: input.guildId,
afterId: input.after ? snowflakeFromIso(input.after) : null,
beforeId: input.before ? snowflakeFromIso(input.before) : null
});
return count === null
? { count: null, reason: "search unavailable or rate limited" }
: { count };
}
private recordError(source: string, message: string): void {
this.errors.unshift({ at: new Date().toISOString(), source, message });
if (this.errors.length > 50) this.errors.length = 50;
+23 -1
View File
@@ -832,6 +832,24 @@ export function startWebUi(deps: WebUiDeps): WebUiHandle {
})
);
app.post(
"/api/media/count",
wrap(async (req, res) => {
const channelId = String(req.body?.channelId ?? "");
if (!channelId) throw new HttpError(400, "channelId is required");
res.json({
ok: true,
result: await deps.media.countMessages({
channelId,
guildId: req.body?.guildId ? String(req.body.guildId) : null,
after: req.body?.after ? String(req.body.after) : null,
before: req.body?.before ? String(req.body.before) : null
})
});
})
);
app.get(
"/api/media/errors",
wrap(async (_req, res) => {
@@ -859,6 +877,8 @@ export function startWebUi(deps: WebUiDeps): WebUiHandle {
const backend =
requested === "slopbot" || requested === "puppet" ? requested : "auto";
const downloadDir = req.body?.downloadDir ? String(req.body.downloadDir) : null;
const after = req.body?.after ? String(req.body.after) : null;
const before = req.body?.before ? String(req.body.before) : null;
const scrape = await deps.media.startScrape({
channelId,
@@ -866,7 +886,9 @@ export function startWebUi(deps: WebUiDeps): WebUiHandle {
mode,
confirm,
backend,
downloadDir
downloadDir,
after,
before
});
deps.getPuppet()?.appendAction({
type: "media",