switch to llms
This commit is contained in:
@@ -9,13 +9,6 @@ export async function executeCodeCommand(args: string[] = []) {
|
||||
// Set environment variables
|
||||
const env = {
|
||||
...process.env,
|
||||
HTTPS_PROXY: undefined,
|
||||
HTTP_PROXY: undefined,
|
||||
ALL_PROXY: undefined,
|
||||
https_proxy: undefined,
|
||||
http_proxy: undefined,
|
||||
all_proxy: undefined,
|
||||
DISABLE_PROMPT_CACHING: "1",
|
||||
ANTHROPIC_AUTH_TOKEN: "test",
|
||||
ANTHROPIC_BASE_URL: `http://127.0.0.1:3456`,
|
||||
API_TIMEOUT_MS: "600000",
|
||||
@@ -29,7 +22,7 @@ export async function executeCodeCommand(args: string[] = []) {
|
||||
const claudeProcess = spawn(claudePath, args, {
|
||||
env,
|
||||
stdio: "inherit",
|
||||
shell: true
|
||||
shell: true,
|
||||
});
|
||||
|
||||
claudeProcess.on("error", (error) => {
|
||||
|
||||
@@ -1,5 +1,3 @@
|
||||
import { HttpsProxyAgent } from "https-proxy-agent";
|
||||
import OpenAI, { ClientOptions } from "openai";
|
||||
import fs from "node:fs/promises";
|
||||
import readline from "node:readline";
|
||||
import {
|
||||
@@ -9,16 +7,6 @@ import {
|
||||
PLUGINS_DIR,
|
||||
} from "../constants";
|
||||
|
||||
export function getOpenAICommonOptions(): ClientOptions {
|
||||
const options: ClientOptions = {};
|
||||
if (process.env.PROXY_URL) {
|
||||
options.httpAgent = new HttpsProxyAgent(process.env.PROXY_URL);
|
||||
} else if (process.env.HTTPS_PROXY) {
|
||||
options.httpAgent = new HttpsProxyAgent(process.env.HTTPS_PROXY);
|
||||
}
|
||||
return options;
|
||||
}
|
||||
|
||||
const ensureDir = async (dir_path: string) => {
|
||||
try {
|
||||
await fs.access(dir_path);
|
||||
@@ -63,9 +51,17 @@ export const readConfigFile = async () => {
|
||||
const baseUrl = await question("Enter OPENAI_BASE_URL: ");
|
||||
const model = await question("Enter OPENAI_MODEL: ");
|
||||
const config = Object.assign({}, DEFAULT_CONFIG, {
|
||||
OPENAI_API_KEY: apiKey,
|
||||
OPENAI_BASE_URL: baseUrl,
|
||||
OPENAI_MODEL: model,
|
||||
Providers: [
|
||||
{
|
||||
name: "openai",
|
||||
api_base_url: baseUrl,
|
||||
api_key: apiKey,
|
||||
models: [model],
|
||||
},
|
||||
],
|
||||
Router: {
|
||||
default: `openai,${model}`,
|
||||
},
|
||||
});
|
||||
await writeConfigFile(config);
|
||||
return config;
|
||||
@@ -82,11 +78,3 @@ export const initConfig = async () => {
|
||||
Object.assign(process.env, config);
|
||||
return config;
|
||||
};
|
||||
|
||||
export const createClient = (options: ClientOptions) => {
|
||||
const client = new OpenAI({
|
||||
...options,
|
||||
...getOpenAICommonOptions(),
|
||||
});
|
||||
return client;
|
||||
};
|
||||
|
||||
87
src/utils/router.ts
Normal file
87
src/utils/router.ts
Normal file
@@ -0,0 +1,87 @@
|
||||
import { MessageCreateParamsBase } from "@anthropic-ai/sdk/resources/messages";
|
||||
import { get_encoding } from "tiktoken";
|
||||
import { log } from "./log";
|
||||
|
||||
const enc = get_encoding("cl100k_base");
|
||||
|
||||
const getUseModel = (req: any, tokenCount: number, config: any) => {
|
||||
if (req.body.model.includes(",")) {
|
||||
return req.body.model;
|
||||
}
|
||||
// if tokenCount is greater than 60K, use the long context model
|
||||
if (tokenCount > 1000 * 60) {
|
||||
log("Using long context model due to token count:", tokenCount);
|
||||
return config.Router!.longContext;
|
||||
}
|
||||
// If the model is claude-3-5-haiku, use the background model
|
||||
if (req.body.model?.startsWith("claude-3-5-haiku")) {
|
||||
log("Using background model for ", req.body.model);
|
||||
return config.Router!.background;
|
||||
}
|
||||
// if exits thinking, use the think model
|
||||
if (req.body.thinking) {
|
||||
log("Using think model for ", req.body.thinking);
|
||||
return config.Router!.think;
|
||||
}
|
||||
return config.Router!.default;
|
||||
};
|
||||
|
||||
export const router = async (req: any, res: any, config: any) => {
|
||||
const { messages, system = [], tools }: MessageCreateParamsBase = req.body;
|
||||
try {
|
||||
let tokenCount = 0;
|
||||
if (Array.isArray(messages)) {
|
||||
messages.forEach((message) => {
|
||||
if (typeof message.content === "string") {
|
||||
tokenCount += enc.encode(message.content).length;
|
||||
} else if (Array.isArray(message.content)) {
|
||||
message.content.forEach((contentPart) => {
|
||||
if (contentPart.type === "text") {
|
||||
tokenCount += enc.encode(contentPart.text).length;
|
||||
} else if (contentPart.type === "tool_use") {
|
||||
tokenCount += enc.encode(
|
||||
JSON.stringify(contentPart.input)
|
||||
).length;
|
||||
} else if (contentPart.type === "tool_result") {
|
||||
tokenCount += enc.encode(
|
||||
typeof contentPart.content === "string"
|
||||
? contentPart.content
|
||||
: JSON.stringify(contentPart.content)
|
||||
).length;
|
||||
}
|
||||
});
|
||||
}
|
||||
});
|
||||
}
|
||||
if (typeof system === "string") {
|
||||
tokenCount += enc.encode(system).length;
|
||||
} else if (Array.isArray(system)) {
|
||||
system.forEach((item) => {
|
||||
if (item.type !== "text") return;
|
||||
if (typeof item.text === "string") {
|
||||
tokenCount += enc.encode(item.text).length;
|
||||
} else if (Array.isArray(item.text)) {
|
||||
item.text.forEach((textPart) => {
|
||||
tokenCount += enc.encode(textPart || "").length;
|
||||
});
|
||||
}
|
||||
});
|
||||
}
|
||||
if (tools) {
|
||||
tools.forEach((tool) => {
|
||||
if (tool.description) {
|
||||
tokenCount += enc.encode(tool.name + tool.description).length;
|
||||
}
|
||||
if (tool.input_schema) {
|
||||
tokenCount += enc.encode(JSON.stringify(tool.input_schema)).length;
|
||||
}
|
||||
});
|
||||
}
|
||||
const model = getUseModel(req, tokenCount, config);
|
||||
req.body.model = model;
|
||||
} catch (error: any) {
|
||||
log("Error in router middleware:", error.message);
|
||||
req.body.model = config.Router!.default;
|
||||
}
|
||||
return;
|
||||
};
|
||||
@@ -15,7 +15,7 @@ export function showStatus() {
|
||||
console.log('');
|
||||
console.log('🚀 Ready to use! Run the following commands:');
|
||||
console.log(' ccr code # Start coding with Claude');
|
||||
console.log(' ccr close # Stop the service');
|
||||
console.log(' ccr stop # Stop the service');
|
||||
} else {
|
||||
console.log('❌ Status: Not Running');
|
||||
console.log('');
|
||||
|
||||
@@ -1,343 +0,0 @@
|
||||
import { Response } from "express";
|
||||
import { OpenAI } from "openai";
|
||||
import { log } from "./log";
|
||||
|
||||
interface ContentBlock {
|
||||
type: string;
|
||||
id?: string;
|
||||
name?: string;
|
||||
input?: any;
|
||||
text?: string;
|
||||
}
|
||||
|
||||
interface MessageEvent {
|
||||
type: string;
|
||||
message?: {
|
||||
id: string;
|
||||
type: string;
|
||||
role: string;
|
||||
content: any[];
|
||||
model: string;
|
||||
stop_reason: string | null;
|
||||
stop_sequence: string | null;
|
||||
usage: {
|
||||
input_tokens: number;
|
||||
output_tokens: number;
|
||||
};
|
||||
};
|
||||
delta?: {
|
||||
stop_reason?: string;
|
||||
stop_sequence?: string | null;
|
||||
content?: ContentBlock[];
|
||||
type?: string;
|
||||
text?: string;
|
||||
partial_json?: string;
|
||||
};
|
||||
index?: number;
|
||||
content_block?: ContentBlock;
|
||||
usage?: {
|
||||
input_tokens: number;
|
||||
output_tokens: number;
|
||||
};
|
||||
}
|
||||
|
||||
export async function streamOpenAIResponse(
|
||||
res: Response,
|
||||
completion: any,
|
||||
model: string,
|
||||
body: any
|
||||
) {
|
||||
const write = (data: string) => {
|
||||
log("response: ", data);
|
||||
res.write(data);
|
||||
};
|
||||
const messageId = "msg_" + Date.now();
|
||||
if (!body.stream) {
|
||||
let content: any = [];
|
||||
if (completion.choices[0].message.content) {
|
||||
content = [ { text: completion.choices[0].message.content, type: "text" } ];
|
||||
}
|
||||
else if (completion.choices[0].message.tool_calls) {
|
||||
content = completion.choices[0].message.tool_calls.map((item: any) => {
|
||||
return {
|
||||
type: 'tool_use',
|
||||
id: item.id,
|
||||
name: item.function?.name,
|
||||
input: item.function?.arguments ? JSON.parse(item.function.arguments) : {},
|
||||
};
|
||||
});
|
||||
}
|
||||
|
||||
const result = {
|
||||
id: messageId,
|
||||
type: "message",
|
||||
role: "assistant",
|
||||
// @ts-ignore
|
||||
content: content,
|
||||
stop_reason: completion.choices[0].finish_reason === 'tool_calls' ? "tool_use" : "end_turn",
|
||||
stop_sequence: null,
|
||||
};
|
||||
try {
|
||||
res.json(result);
|
||||
res.end();
|
||||
return;
|
||||
} catch (error) {
|
||||
log("Error sending response:", error);
|
||||
res.status(500).send("Internal Server Error");
|
||||
}
|
||||
}
|
||||
|
||||
let contentBlockIndex = 0;
|
||||
let currentContentBlocks: ContentBlock[] = [];
|
||||
|
||||
// Send message_start event
|
||||
const messageStart: MessageEvent = {
|
||||
type: "message_start",
|
||||
message: {
|
||||
id: messageId,
|
||||
type: "message",
|
||||
role: "assistant",
|
||||
content: [],
|
||||
model,
|
||||
stop_reason: null,
|
||||
stop_sequence: null,
|
||||
usage: { input_tokens: 1, output_tokens: 1 },
|
||||
},
|
||||
};
|
||||
write(`event: message_start\ndata: ${JSON.stringify(messageStart)}\n\n`);
|
||||
|
||||
let isToolUse = false;
|
||||
let toolUseJson = "";
|
||||
let hasStartedTextBlock = false;
|
||||
let currentToolCallId: string | null = null;
|
||||
let toolCallJsonMap = new Map<string, string>();
|
||||
|
||||
try {
|
||||
for await (const chunk of completion) {
|
||||
log("Processing chunk:", chunk);
|
||||
const delta = chunk.choices[0].delta;
|
||||
|
||||
if (delta.tool_calls && delta.tool_calls.length > 0) {
|
||||
for (const toolCall of delta.tool_calls) {
|
||||
const toolCallId = toolCall.id;
|
||||
|
||||
// Check if this is a new tool call by ID
|
||||
if (toolCallId && toolCallId !== currentToolCallId) {
|
||||
// End previous tool call if one was active
|
||||
if (isToolUse && currentToolCallId) {
|
||||
const contentBlockStop: MessageEvent = {
|
||||
type: "content_block_stop",
|
||||
index: contentBlockIndex,
|
||||
};
|
||||
write(
|
||||
`event: content_block_stop\ndata: ${JSON.stringify(
|
||||
contentBlockStop
|
||||
)}\n\n`
|
||||
);
|
||||
}
|
||||
|
||||
// Start new tool call block
|
||||
isToolUse = true;
|
||||
currentToolCallId = toolCallId;
|
||||
contentBlockIndex++;
|
||||
toolCallJsonMap.set(toolCallId, ""); // Initialize JSON accumulator for this tool call
|
||||
|
||||
const toolBlock: ContentBlock = {
|
||||
type: "tool_use",
|
||||
id: toolCallId,
|
||||
name: toolCall.function?.name,
|
||||
input: {},
|
||||
};
|
||||
|
||||
const toolBlockStart: MessageEvent = {
|
||||
type: "content_block_start",
|
||||
index: contentBlockIndex,
|
||||
content_block: toolBlock,
|
||||
};
|
||||
|
||||
currentContentBlocks.push(toolBlock);
|
||||
|
||||
write(
|
||||
`event: content_block_start\ndata: ${JSON.stringify(
|
||||
toolBlockStart
|
||||
)}\n\n`
|
||||
);
|
||||
}
|
||||
|
||||
// Stream tool call JSON
|
||||
if (toolCall.function?.arguments && currentToolCallId) {
|
||||
const jsonDelta: MessageEvent = {
|
||||
type: "content_block_delta",
|
||||
index: contentBlockIndex,
|
||||
delta: {
|
||||
type: "input_json_delta",
|
||||
partial_json: toolCall.function.arguments,
|
||||
},
|
||||
};
|
||||
|
||||
// Accumulate JSON for this specific tool call
|
||||
const currentJson = toolCallJsonMap.get(currentToolCallId) || "";
|
||||
toolCallJsonMap.set(currentToolCallId, currentJson + toolCall.function.arguments);
|
||||
toolUseJson = toolCallJsonMap.get(currentToolCallId) || "";
|
||||
|
||||
try {
|
||||
const parsedJson = JSON.parse(toolUseJson);
|
||||
currentContentBlocks[contentBlockIndex].input = parsedJson;
|
||||
} catch (e) {
|
||||
log("JSON parsing error (continuing to accumulate):", e);
|
||||
// JSON not yet complete, continue accumulating
|
||||
}
|
||||
|
||||
write(
|
||||
`event: content_block_delta\ndata: ${JSON.stringify(jsonDelta)}\n\n`
|
||||
);
|
||||
}
|
||||
}
|
||||
} else if (delta.content) {
|
||||
// Handle regular text content
|
||||
if (isToolUse) {
|
||||
log("Tool call ended here:", delta);
|
||||
// End previous tool call block
|
||||
const contentBlockStop: MessageEvent = {
|
||||
type: "content_block_stop",
|
||||
index: contentBlockIndex,
|
||||
};
|
||||
|
||||
write(
|
||||
`event: content_block_stop\ndata: ${JSON.stringify(
|
||||
contentBlockStop
|
||||
)}\n\n`
|
||||
);
|
||||
contentBlockIndex++;
|
||||
isToolUse = false;
|
||||
currentToolCallId = null;
|
||||
toolUseJson = ""; // Reset for safety
|
||||
}
|
||||
|
||||
if (!delta.content) continue;
|
||||
|
||||
// If text block not yet started, send content_block_start
|
||||
if (!hasStartedTextBlock) {
|
||||
const textBlock: ContentBlock = {
|
||||
type: "text",
|
||||
text: "",
|
||||
};
|
||||
|
||||
const textBlockStart: MessageEvent = {
|
||||
type: "content_block_start",
|
||||
index: contentBlockIndex,
|
||||
content_block: textBlock,
|
||||
};
|
||||
|
||||
currentContentBlocks.push(textBlock);
|
||||
|
||||
write(
|
||||
`event: content_block_start\ndata: ${JSON.stringify(
|
||||
textBlockStart
|
||||
)}\n\n`
|
||||
);
|
||||
hasStartedTextBlock = true;
|
||||
}
|
||||
|
||||
// Send regular text content
|
||||
const contentDelta: MessageEvent = {
|
||||
type: "content_block_delta",
|
||||
index: contentBlockIndex,
|
||||
delta: {
|
||||
type: "text_delta",
|
||||
text: delta.content,
|
||||
},
|
||||
};
|
||||
|
||||
// Update content block text
|
||||
if (currentContentBlocks[contentBlockIndex]) {
|
||||
currentContentBlocks[contentBlockIndex].text += delta.content;
|
||||
}
|
||||
|
||||
write(
|
||||
`event: content_block_delta\ndata: ${JSON.stringify(
|
||||
contentDelta
|
||||
)}\n\n`
|
||||
);
|
||||
}
|
||||
}
|
||||
} catch (e: any) {
|
||||
// If text block not yet started, send content_block_start
|
||||
if (!hasStartedTextBlock) {
|
||||
const textBlock: ContentBlock = {
|
||||
type: "text",
|
||||
text: "",
|
||||
};
|
||||
|
||||
const textBlockStart: MessageEvent = {
|
||||
type: "content_block_start",
|
||||
index: contentBlockIndex,
|
||||
content_block: textBlock,
|
||||
};
|
||||
|
||||
currentContentBlocks.push(textBlock);
|
||||
|
||||
write(
|
||||
`event: content_block_start\ndata: ${JSON.stringify(
|
||||
textBlockStart
|
||||
)}\n\n`
|
||||
);
|
||||
hasStartedTextBlock = true;
|
||||
}
|
||||
|
||||
// Send regular text content
|
||||
const contentDelta: MessageEvent = {
|
||||
type: "content_block_delta",
|
||||
index: contentBlockIndex,
|
||||
delta: {
|
||||
type: "text_delta",
|
||||
text: JSON.stringify(e),
|
||||
},
|
||||
};
|
||||
|
||||
// Update content block text
|
||||
if (currentContentBlocks[contentBlockIndex]) {
|
||||
currentContentBlocks[contentBlockIndex].text += JSON.stringify(e);
|
||||
}
|
||||
|
||||
write(
|
||||
`event: content_block_delta\ndata: ${JSON.stringify(contentDelta)}\n\n`
|
||||
);
|
||||
}
|
||||
|
||||
// Close last content block if any is open
|
||||
if (isToolUse || hasStartedTextBlock) {
|
||||
const contentBlockStop: MessageEvent = {
|
||||
type: "content_block_stop",
|
||||
index: contentBlockIndex,
|
||||
};
|
||||
|
||||
write(
|
||||
`event: content_block_stop\ndata: ${JSON.stringify(contentBlockStop)}\n\n`
|
||||
);
|
||||
}
|
||||
|
||||
// Send message_delta event with appropriate stop_reason
|
||||
const messageDelta: MessageEvent = {
|
||||
type: "message_delta",
|
||||
delta: {
|
||||
stop_reason: isToolUse ? "tool_use" : "end_turn",
|
||||
stop_sequence: null,
|
||||
content: currentContentBlocks,
|
||||
},
|
||||
usage: { input_tokens: 100, output_tokens: 150 },
|
||||
};
|
||||
if (!isToolUse) {
|
||||
log("body: ", body, "messageDelta: ", messageDelta);
|
||||
}
|
||||
|
||||
write(`event: message_delta\ndata: ${JSON.stringify(messageDelta)}\n\n`);
|
||||
|
||||
// Send message_stop event
|
||||
const messageStop: MessageEvent = {
|
||||
type: "message_stop",
|
||||
};
|
||||
|
||||
write(`event: message_stop\ndata: ${JSON.stringify(messageStop)}\n\n`);
|
||||
res.end();
|
||||
}
|
||||
Reference in New Issue
Block a user