@echofiles/echo-pdf 0.4.0 → 0.4.2

package/dist/file-store-do.d.ts

@@ -0,0 +1,36 @@
+import type { StoragePolicy } from "./pdf-types.js";
+import type { DurableObjectNamespace, DurableObjectState, StoredFileMeta, StoredFileRecord } from "./types.js";
+interface StoreStats {
+    readonly fileCount: number;
+    readonly totalBytes: number;
+}
+export declare class FileStoreDO {
+    private readonly state;
+    constructor(state: DurableObjectState);
+    fetch(request: Request): Promise<Response>;
+}
+export declare class DurableObjectFileStore {
+    private readonly namespace;
+    private readonly policy;
+    constructor(namespace: DurableObjectNamespace, policy: StoragePolicy);
+    private stub;
+    put(input: {
+        readonly filename: string;
+        readonly mimeType: string;
+        readonly bytes: Uint8Array;
+    }): Promise<StoredFileMeta>;
+    get(fileId: string): Promise<StoredFileRecord | null>;
+    list(): Promise<ReadonlyArray<StoredFileMeta>>;
+    delete(fileId: string): Promise<boolean>;
+    stats(): Promise<{
+        policy: StoragePolicy;
+        stats: StoreStats;
+    }>;
+    cleanup(): Promise<{
+        policy: StoragePolicy;
+        deletedExpired: number;
+        deletedEvicted: number;
+        stats: StoreStats;
+    }>;
+}
+export {};

package/dist/file-store-do.js

@@ -0,0 +1,298 @@
+import { fromBase64, toBase64 } from "./file-utils.js";
+const json = (data, status = 200) => new Response(JSON.stringify(data), {
+    status,
+    headers: { "Content-Type": "application/json; charset=utf-8" },
+});
+const readJson = async (request) => {
+    try {
+        const body = await request.json();
+        if (typeof body === "object" && body !== null && !Array.isArray(body)) {
+            return body;
+        }
+        return {};
+    }
+    catch {
+        return {};
+    }
+};
+const defaultPolicy = () => ({
+    maxFileBytes: 1_200_000,
+    maxTotalBytes: 52_428_800,
+    ttlHours: 24,
+    cleanupBatchSize: 50,
+});
+const parsePolicy = (input) => {
+    const raw = typeof input === "object" && input !== null && !Array.isArray(input)
+        ? input
+        : {};
+    const fallback = defaultPolicy();
+    const maxFileBytes = Number(raw.maxFileBytes ?? fallback.maxFileBytes);
+    const maxTotalBytes = Number(raw.maxTotalBytes ?? fallback.maxTotalBytes);
+    const ttlHours = Number(raw.ttlHours ?? fallback.ttlHours);
+    const cleanupBatchSize = Number(raw.cleanupBatchSize ?? fallback.cleanupBatchSize);
+    return {
+        maxFileBytes: Number.isFinite(maxFileBytes) && maxFileBytes > 0 ? Math.floor(maxFileBytes) : fallback.maxFileBytes,
+        maxTotalBytes: Number.isFinite(maxTotalBytes) && maxTotalBytes > 0 ? Math.floor(maxTotalBytes) : fallback.maxTotalBytes,
+        ttlHours: Number.isFinite(ttlHours) && ttlHours > 0 ? ttlHours : fallback.ttlHours,
+        cleanupBatchSize: Number.isFinite(cleanupBatchSize) && cleanupBatchSize > 0 ? Math.floor(cleanupBatchSize) : fallback.cleanupBatchSize,
+    };
+};
+const toMeta = (value) => ({
+    id: value.id,
+    filename: value.filename,
+    mimeType: value.mimeType,
+    sizeBytes: value.sizeBytes,
+    createdAt: value.createdAt,
+});
+const listStoredValues = async (state) => {
+    const listed = await state.storage.list({ prefix: "file:" });
+    return [...listed.values()];
+};
+const computeStats = (files) => ({
+    fileCount: files.length,
+    totalBytes: files.reduce((sum, file) => sum + file.sizeBytes, 0),
+});
+const isExpired = (createdAt, ttlHours) => {
+    const createdMs = Date.parse(createdAt);
+    if (!Number.isFinite(createdMs))
+        return false;
+    return Date.now() - createdMs > ttlHours * 60 * 60 * 1000;
+};
+const deleteFiles = async (state, files) => {
+    let deleted = 0;
+    for (const file of files) {
+        const ok = await state.storage.delete(`file:${file.id}`);
+        if (ok)
+            deleted += 1;
+    }
+    return deleted;
+};
+export class FileStoreDO {
+    state;
+    constructor(state) {
+        this.state = state;
+    }
+    async fetch(request) {
+        const url = new URL(request.url);
+        if (request.method === "POST" && url.pathname === "/put") {
+            const body = await readJson(request);
+            const policy = parsePolicy(body.policy);
+            const filename = typeof body.filename === "string" ? body.filename : `file-${Date.now()}`;
+            const mimeType = typeof body.mimeType === "string" ? body.mimeType : "application/octet-stream";
+            const bytesBase64 = typeof body.bytesBase64 === "string" ? body.bytesBase64 : "";
+            const bytes = fromBase64(bytesBase64);
+            if (bytes.byteLength > policy.maxFileBytes) {
+                return json({
+                    error: `file too large: ${bytes.byteLength} bytes exceeds maxFileBytes ${policy.maxFileBytes}`,
+                    code: "FILE_TOO_LARGE",
+                    policy,
+                }, 413);
+            }
+            let files = await listStoredValues(this.state);
+            const expired = files.filter((file) => isExpired(file.createdAt, policy.ttlHours));
+            if (expired.length > 0) {
+                await deleteFiles(this.state, expired);
+                files = await listStoredValues(this.state);
+            }
+            let stats = computeStats(files);
+            const projected = stats.totalBytes + bytes.byteLength;
+            if (projected > policy.maxTotalBytes) {
+                const needFree = projected - policy.maxTotalBytes;
+                const candidates = [...files]
+                    .sort((a, b) => Date.parse(a.createdAt) - Date.parse(b.createdAt))
+                    .slice(0, policy.cleanupBatchSize);
+                let freed = 0;
+                const evictList = [];
+                for (const file of candidates) {
+                    evictList.push(file);
+                    freed += file.sizeBytes;
+                    if (freed >= needFree)
+                        break;
+                }
+                if (evictList.length > 0) {
+                    await deleteFiles(this.state, evictList);
+                    files = await listStoredValues(this.state);
+                    stats = computeStats(files);
+                }
+            }
+            if (stats.totalBytes + bytes.byteLength > policy.maxTotalBytes) {
+                return json({
+                    error: `storage quota exceeded: total ${stats.totalBytes} + incoming ${bytes.byteLength} > maxTotalBytes ${policy.maxTotalBytes}`,
+                    code: "STORAGE_QUOTA_EXCEEDED",
+                    policy,
+                    stats,
+                }, 507);
+            }
+            const id = crypto.randomUUID();
+            const value = {
+                id,
+                filename,
+                mimeType,
+                sizeBytes: bytes.byteLength,
+                createdAt: new Date().toISOString(),
+                bytesBase64,
+            };
+            await this.state.storage.put(`file:${id}`, value);
+            return json({ file: toMeta(value), policy });
+        }
+        if (request.method === "GET" && url.pathname === "/get") {
+            const fileId = url.searchParams.get("fileId");
+            if (!fileId)
+                return json({ error: "Missing fileId" }, 400);
+            const value = await this.state.storage.get(`file:${fileId}`);
+            if (!value)
+                return json({ file: null });
+            return json({ file: value });
+        }
+        if (request.method === "GET" && url.pathname === "/list") {
+            const listed = await this.state.storage.list({ prefix: "file:" });
+            const files = [...listed.values()].map(toMeta);
+            return json({ files });
+        }
+        if (request.method === "POST" && url.pathname === "/delete") {
+            const body = await readJson(request);
+            const fileId = typeof body.fileId === "string" ? body.fileId : "";
+            if (!fileId)
+                return json({ error: "Missing fileId" }, 400);
+            const key = `file:${fileId}`;
+            const existing = await this.state.storage.get(key);
+            if (!existing)
+                return json({ deleted: false });
+            await this.state.storage.delete(key);
+            return json({ deleted: true });
+        }
+        if (request.method === "GET" && url.pathname === "/stats") {
+            let policyInput;
+            const encoded = url.searchParams.get("policy");
+            if (encoded) {
+                try {
+                    policyInput = JSON.parse(encoded);
+                }
+                catch {
+                    policyInput = undefined;
+                }
+            }
+            const policy = parsePolicy(policyInput);
+            const files = await listStoredValues(this.state);
+            const stats = computeStats(files);
+            return json({ policy, stats });
+        }
+        if (request.method === "POST" && url.pathname === "/cleanup") {
+            const body = await readJson(request);
+            const policy = parsePolicy(body.policy);
+            const files = await listStoredValues(this.state);
+            const expired = files.filter((file) => isExpired(file.createdAt, policy.ttlHours));
+            const deletedExpired = await deleteFiles(this.state, expired);
+            const afterExpired = await listStoredValues(this.state);
+            let stats = computeStats(afterExpired);
+            let deletedEvicted = 0;
+            if (stats.totalBytes > policy.maxTotalBytes) {
+                const sorted = [...afterExpired].sort((a, b) => Date.parse(a.createdAt) - Date.parse(b.createdAt));
+                const evictList = [];
+                for (const file of sorted) {
+                    evictList.push(file);
+                    const projected = stats.totalBytes - evictList.reduce((sum, item) => sum + item.sizeBytes, 0);
+                    if (projected <= policy.maxTotalBytes)
+                        break;
+                    if (evictList.length >= policy.cleanupBatchSize)
+                        break;
+                }
+                deletedEvicted = await deleteFiles(this.state, evictList);
+                stats = computeStats(await listStoredValues(this.state));
+            }
+            return json({
+                policy,
+                deletedExpired,
+                deletedEvicted,
+                stats,
+            });
+        }
+        return json({ error: "Not found" }, 404);
+    }
+}
+export class DurableObjectFileStore {
+    namespace;
+    policy;
+    constructor(namespace, policy) {
+        this.namespace = namespace;
+        this.policy = policy;
+    }
+    stub() {
+        return this.namespace.get(this.namespace.idFromName("echo-pdf-file-store"));
+    }
+    async put(input) {
+        const response = await this.stub().fetch("https://do/put", {
+            method: "POST",
+            headers: { "Content-Type": "application/json" },
+            body: JSON.stringify({
+                filename: input.filename,
+                mimeType: input.mimeType,
+                bytesBase64: toBase64(input.bytes),
+                policy: this.policy,
+            }),
+        });
+        const payload = (await response.json());
+        if (!response.ok || !payload.file) {
+            const details = payload;
+            const error = new Error(payload.error ?? "DO put failed");
+            error.status = response.status;
+            error.code = typeof details.code === "string" ? details.code : undefined;
+            error.details = { policy: details.policy, stats: details.stats };
+            throw error;
+        }
+        return payload.file;
+    }
+    async get(fileId) {
+        const response = await this.stub().fetch(`https://do/get?fileId=${encodeURIComponent(fileId)}`);
+        const payload = (await response.json());
+        if (!response.ok)
+            throw new Error("DO get failed");
+        if (!payload.file)
+            return null;
+        return {
+            id: payload.file.id,
+            filename: payload.file.filename,
+            mimeType: payload.file.mimeType,
+            sizeBytes: payload.file.sizeBytes,
+            createdAt: payload.file.createdAt,
+            bytes: fromBase64(payload.file.bytesBase64),
+        };
+    }
+    async list() {
+        const response = await this.stub().fetch("https://do/list");
+        const payload = (await response.json());
+        if (!response.ok)
+            throw new Error("DO list failed");
+        return payload.files ?? [];
+    }
+    async delete(fileId) {
+        const response = await this.stub().fetch("https://do/delete", {
+            method: "POST",
+            headers: { "Content-Type": "application/json" },
+            body: JSON.stringify({ fileId }),
+        });
+        const payload = (await response.json());
+        if (!response.ok)
+            throw new Error("DO delete failed");
+        return payload.deleted === true;
+    }
+    async stats() {
+        const policyEncoded = encodeURIComponent(JSON.stringify(this.policy));
+        const response = await this.stub().fetch(`https://do/stats?policy=${policyEncoded}`);
+        const payload = (await response.json());
+        if (!response.ok)
+            throw new Error("DO stats failed");
+        return payload;
+    }
+    async cleanup() {
+        const response = await this.stub().fetch("https://do/cleanup", {
+            method: "POST",
+            headers: { "Content-Type": "application/json" },
+            body: JSON.stringify({ policy: this.policy }),
+        });
+        const payload = (await response.json());
+        if (!response.ok)
+            throw new Error("DO cleanup failed");
+        return payload;
+    }
+}

package/dist/file-utils.d.ts

@@ -0,0 +1,6 @@
+import type { ReturnMode, StoredFileRecord } from "./types.js";
+export declare const fromBase64: (value: string) => Uint8Array;
+export declare const toBase64: (bytes: Uint8Array) => string;
+export declare const toDataUrl: (bytes: Uint8Array, mimeType: string) => string;
+export declare const normalizeReturnMode: (value: unknown) => ReturnMode;
+export declare const toInlineFilePayload: (file: StoredFileRecord, includeBase64: boolean) => Record<string, unknown>;

package/dist/file-utils.js

@@ -0,0 +1,36 @@
+export const fromBase64 = (value) => {
+    const raw = atob(value.replace(/^data:.*;base64,/, ""));
+    const out = new Uint8Array(raw.length);
+    for (let i = 0; i < raw.length; i++) {
+        out[i] = raw.charCodeAt(i);
+    }
+    return out;
+};
+export const toBase64 = (bytes) => {
+    let binary = "";
+    const chunkSize = 0x8000;
+    for (let i = 0; i < bytes.length; i += chunkSize) {
+        const chunk = bytes.subarray(i, i + chunkSize);
+        binary += String.fromCharCode(...chunk);
+    }
+    return btoa(binary);
+};
+export const toDataUrl = (bytes, mimeType) => `data:${mimeType};base64,${toBase64(bytes)}`;
+export const normalizeReturnMode = (value) => {
+    if (value === "file_id" || value === "url" || value === "inline") {
+        return value;
+    }
+    return "inline";
+};
+export const toInlineFilePayload = (file, includeBase64) => ({
+    file: {
+        id: file.id,
+        filename: file.filename,
+        mimeType: file.mimeType,
+        sizeBytes: file.sizeBytes,
+        createdAt: file.createdAt,
+    },
+    dataUrl: file.mimeType.startsWith("image/") ? toDataUrl(file.bytes, file.mimeType) : undefined,
+    base64: includeBase64 ? toBase64(file.bytes) : undefined,
+    text: file.mimeType.startsWith("text/") ? new TextDecoder().decode(file.bytes) : undefined,
+});

package/dist/http-error.d.ts

@@ -0,0 +1,9 @@
+export declare class HttpError extends Error {
+    readonly status: number;
+    readonly code: string;
+    readonly details?: unknown;
+    constructor(status: number, code: string, message: string, details?: unknown);
+}
+export declare const badRequest: (code: string, message: string, details?: unknown) => HttpError;
+export declare const notFound: (code: string, message: string, details?: unknown) => HttpError;
+export declare const unprocessable: (code: string, message: string, details?: unknown) => HttpError;

package/dist/http-error.js

@@ -0,0 +1,14 @@
+export class HttpError extends Error {
+    status;
+    code;
+    details;
+    constructor(status, code, message, details) {
+        super(message);
+        this.status = status;
+        this.code = code;
+        this.details = details;
+    }
+}
+export const badRequest = (code, message, details) => new HttpError(400, code, message, details);
+export const notFound = (code, message, details) => new HttpError(404, code, message, details);
+export const unprocessable = (code, message, details) => new HttpError(422, code, message, details);

package/dist/index.d.ts

@@ -0,0 +1 @@
+export * from "./core/index.js";

package/dist/index.js

@@ -0,0 +1 @@
+export * from "./core/index.js";

package/dist/mcp-server.d.ts

@@ -0,0 +1,3 @@
+import type { Env, FileStore } from "./types.js";
+import type { EchoPdfConfig } from "./pdf-types.js";
+export declare const handleMcpRequest: (request: Request, env: Env, config: EchoPdfConfig, fileStore: FileStore) => Promise<Response>;

package/dist/mcp-server.js

@@ -0,0 +1,127 @@
+import { checkHeaderAuth } from "./auth.js";
+import { buildMcpContent, buildToolOutputEnvelope } from "./response-schema.js";
+import { callTool, listToolSchemas } from "./tool-registry.js";
+const ok = (id, result) => new Response(JSON.stringify({
+    jsonrpc: "2.0",
+    id: id ?? null,
+    result,
+}), { headers: { "Content-Type": "application/json" } });
+const err = (id, code, message, data) => new Response(JSON.stringify({
+    jsonrpc: "2.0",
+    id: id ?? null,
+    error: data ? { code, message, data } : { code, message },
+}), { status: 400, headers: { "Content-Type": "application/json" } });
+const asObj = (v) => typeof v === "object" && v !== null && !Array.isArray(v) ? v : {};
+const resolvePublicBaseUrl = (request, configured) => typeof configured === "string" && configured.length > 0 ? configured : request.url;
+const prepareMcpToolArgs = (toolName, args) => {
+    if (toolName === "pdf_extract_pages") {
+        const mode = typeof args.returnMode === "string" ? args.returnMode : "";
+        if (!mode) {
+            return { ...args, returnMode: "url" };
+        }
+    }
+    return args;
+};
+export const handleMcpRequest = async (request, env, config, fileStore) => {
+    const auth = checkHeaderAuth(request, env, {
+        authHeader: config.mcp.authHeader,
+        authEnv: config.mcp.authEnv,
+        allowMissingSecret: env.ECHO_PDF_ALLOW_MISSING_AUTH_SECRET === "1",
+        misconfiguredCode: "AUTH_MISCONFIGURED",
+        unauthorizedCode: "UNAUTHORIZED",
+        contextName: "MCP",
+    });
+    if (!auth.ok) {
+        return new Response(JSON.stringify({ error: auth.message, code: auth.code }), {
+            status: auth.status,
+            headers: { "Content-Type": "application/json" },
+        });
+    }
+    let body;
+    try {
+        body = (await request.json());
+    }
+    catch {
+        return err(null, -32700, "Parse error");
+    }
+    if (typeof body !== "object" || body === null) {
+        return err(null, -32600, "Invalid Request");
+    }
+    if (body.jsonrpc !== "2.0") {
+        return err(body.id ?? null, -32600, "Invalid Request: jsonrpc must be '2.0'");
+    }
+    const method = body.method ?? "";
+    const id = body.id ?? null;
+    if (typeof method !== "string" || method.length === 0) {
+        return err(id, -32600, "Invalid Request: method is required");
+    }
+    if (method.startsWith("notifications/")) {
+        return new Response(null, { status: 204 });
+    }
+    const params = asObj(body.params);
+    if (method === "initialize") {
+        return ok(id, {
+            protocolVersion: "2024-11-05",
+            serverInfo: {
+                name: config.mcp.serverName,
+                version: config.mcp.version,
+            },
+            capabilities: {
+                tools: {},
+            },
+        });
+    }
+    if (method === "tools/list") {
+        return ok(id, { tools: listToolSchemas().map((tool) => ({
+                name: tool.name,
+                description: tool.description,
+                inputSchema: tool.inputSchema,
+            })) });
+    }
+    if (method !== "tools/call") {
+        return err(id, -32601, `Unsupported method: ${method}`);
+    }
+    const toolName = typeof params.name === "string" ? params.name : "";
+    const args = prepareMcpToolArgs(toolName, asObj(params.arguments));
+    if (!toolName) {
+        return err(id, -32602, "Invalid params: name is required", {
+            code: "INVALID_PARAMS",
+            status: 400,
+        });
+    }
+    try {
+        const result = await callTool(toolName, args, {
+            config,
+            env,
+            fileStore,
+        });
+        const envelope = buildToolOutputEnvelope(result, resolvePublicBaseUrl(request, config.service.publicBaseUrl));
+        return ok(id, { content: buildMcpContent(envelope) });
+    }
+    catch (error) {
+        const message = error instanceof Error ? error.message : String(error);
+        const status = error?.status;
+        const stableStatus = typeof status === "number" && Number.isFinite(status) ? status : 500;
+        const code = error?.code;
+        const details = error?.details;
+        if (message.startsWith("Unknown tool:")) {
+            return err(id, -32601, message, {
+                code: typeof code === "string" ? code : "TOOL_NOT_FOUND",
+                status: 404,
+                details,
+            });
+        }
+        if (stableStatus >= 400 && stableStatus < 500) {
+            return err(id, -32602, message, {
+                code: typeof code === "string" ? code : "INVALID_PARAMS",
+                status: stableStatus,
+                details,
+            });
+        }
+        return err(id, -32000, message, {
+            code: typeof code === "string" ? code : "INTERNAL_ERROR",
+            status: stableStatus,
+            details,
+        });
+    }
+};

package/dist/pdf-agent.d.ts

@@ -0,0 +1,18 @@
+import type { Env, FileStore } from "./types.js";
+import type { AgentTraceEvent, EchoPdfConfig, PdfOperationRequest } from "./pdf-types.js";
+interface RuntimeOptions {
+    readonly trace?: (event: AgentTraceEvent) => void;
+    readonly fileStore: FileStore;
+}
+export declare const ingestPdfFromPayload: (config: EchoPdfConfig, input: {
+    readonly fileId?: string;
+    readonly url?: string;
+    readonly base64?: string;
+    readonly filename?: string;
+}, opts: RuntimeOptions) => Promise<{
+    id: string;
+    filename: string;
+    bytes: Uint8Array;
+}>;
+export declare const runPdfAgent: (config: EchoPdfConfig, env: Env, request: PdfOperationRequest, opts: RuntimeOptions) => Promise<unknown>;
+export {};