fixes google ai language filter and updates readme

README.md

@@ -1,14 +1,20 @@
 # OAI Reverse Proxy
-
 Reverse proxy server for various LLM APIs.
 
 ### Table of Contents
-- [What is this?](#what-is-this)
-- [Features](#features)
-- [Usage Instructions](#usage-instructions)
-  - [Self-hosting](#self-hosting)
-  - [Huggingface (outdated, not advised)](#huggingface-outdated-not-advised)
-- [Local Development](#local-development)
+<!-- TOC -->
+* [OAI Reverse Proxy](#oai-reverse-proxy)
+    * [Table of Contents](#table-of-contents)
+  * [What is this?](#what-is-this)
+  * [Features](#features)
+  * [Usage Instructions](#usage-instructions)
+    * [Personal Use (single-user)](#personal-use-single-user)
+      * [Updating](#updating)
+      * [Local Development](#local-development)
+    * [Self-hosting](#self-hosting)
+  * [Building](#building)
+  * [Forking](#forking)
+<!-- TOC -->
 
 ## What is this?
 This project allows you to run a reverse proxy server for various LLM APIs.
@@ -27,37 +33,42 @@ This project allows you to run a reverse proxy server for various LLM APIs.
   - [x] Simple role-based permissions
   - [x] Per-model token quotas
   - [x] Temporary user accounts
-- [x] Prompt and completion logging
+- [x] Event audit logging
+- [x] Optional full logging of prompts and completions
 - [x] Abuse detection and prevention
+  - [x] IP address and user token model invocation rate limits
+  - [x] IP blacklists
+  - [x] Proof-of-work challenge for access by anonymous users
 
 ---
 
 ## Usage Instructions
 If you'd like to run your own instance of this server, you'll need to deploy it somewhere and configure it with your API keys. A few easy options are provided below, though you can also deploy it to any other service you'd like if you know what you're doing and the service supports Node.js.
 
+### Personal Use (single-user)
+If you just want to run the proxy server to use yourself without hosting it for others:
+1. Install [Node.js](https://nodejs.org/en/download/) >= 18.0.0
+2. Clone this repository
+3. Create a `.env` file in the root of the project and add your API keys. See the [.env.example](./.env.example) file for an example.
+4. Install dependencies with `npm install`
+5. Run `npm run build`
+6. Run `npm start`
+
+#### Updating
+You must re-run `npm install` and `npm run build` whenever you pull new changes from the repository.
+
+#### Local Development
+Use `npm run start:dev` to run the proxy in development mode with watch mode enabled. Use `npm run type-check` to run the type checker across the project.
+
 ### Self-hosting
-[See here for instructions on how to self-host the application on your own VPS or local machine.](./docs/self-hosting.md)
+[See here for instructions on how to self-host the application on your own VPS or local machine and expose it to the internet for others to use.](./docs/self-hosting.md)
 
 **Ensure you set the `TRUSTED_PROXIES` environment variable according to your deployment.** Refer to [.env.example](./.env.example) and [config.ts](./src/config.ts) for more information.
 
-### Huggingface (outdated, not advised)
-[See here for instructions on how to deploy to a Huggingface Space.](./docs/deploy-huggingface.md)
-
-## Local Development
-To run the proxy locally for development or testing, install Node.js >= 18.0.0 and follow the steps below.
-
-1. Clone the repo
-2. Install dependencies with `npm install`
-3. Create a `.env` file in the root of the project and add your API keys. See the [.env.example](./.env.example) file for an example.
-4. Start the server in development mode with `npm run start:dev`.
-
-You can also use `npm run start:dev:tsc` to enable project-wide type checking at the cost of slower startup times. `npm run type-check` can be used to run type checking without starting the server.
-
 ## Building
-To build the project, run `npm run build`. This will compile the TypeScript code to JavaScript and output it to the `build` directory.
+To build the project, run `npm run build`. This will compile the TypeScript code to JavaScript and output it to the `build` directory. You should run this whenever you pull new changes from the repository.
 
 Note that if you are trying to build the server on a very memory-constrained (<= 1GB) VPS, you may need to run the build with `NODE_OPTIONS=--max_old_space_size=2048 npm run build` to avoid running out of memory during the build process, assuming you have swap enabled.  The application itself should run fine on a 512MB VPS for most reasonable traffic levels.
 
 ## Forking
-
 If you are forking the repository on GitGud, you may wish to disable GitLab CI/CD or you will be spammed with emails about failed builds due not having any CI runners. You can do this by going to *Settings > General > Visibility, project features, permissions* and then disabling the "CI/CD" feature.

package.json

@@ -10,7 +10,6 @@
     "start": "node --trace-deprecation --trace-warnings build/server.js",
     "start:dev": "nodemon --watch src --exec ts-node --transpile-only src/server.ts",
     "start:debug": "ts-node --inspect --transpile-only src/server.ts",
-    "start:replit": "tsc && node build/server.js",
     "start:watch": "nodemon --require source-map-support/register build/server.js",
     "type-check": "tsc --noEmit"
   },

src/proxy/middleware/request/preprocessors/language-filter.ts

@@ -1,4 +1,5 @@
 import { Request } from "express";
+import { z } from "zod";
 import { config } from "../../../../config";
 import { assertNever } from "../../../../shared/utils";
 import { RequestPreprocessor } from "../index";
@@ -8,6 +9,7 @@ import {
   OpenAIChatMessage,
   flattenAnthropicMessages,
 } from "../../../../shared/api-schemas";
+import { GoogleAIV1GenerateContentSchema } from "../../../../shared/api-schemas/google-ai";
 
 const rejectedClients = new Map<string, number>();
 
@@ -50,6 +52,10 @@ export const languageFilter: RequestPreprocessor = async (req) => {
   }
 };
 
+/* 
+TODO: this is not type safe and does not raise errors if request body zod schema
+is changed.
+*/
 function getPromptFromRequest(req: Request) {
   const service = req.outboundApi;
   const body = req.body;
@@ -75,8 +81,13 @@ function getPromptFromRequest(req: Request) {
     case "openai-image":
     case "mistral-text":
       return body.prompt;
-    case "google-ai":
-      return body.prompt.text;
+    case "google-ai": {
+      const b = body as z.infer<typeof GoogleAIV1GenerateContentSchema>;
+      return [
+        b.systemInstruction?.parts.map((p) => p.text),
+        ...b.contents.flatMap((c) => c.parts.map((p) => p.text)),
+      ].join("\n");
+    }
     default:
       assertNever(service);
   }