feat: extract models from command-code dist, load via models.json
- Add scripts/extract-models.ts to parse command-code npm dist file - Generate models.json (21 models, 15 pricing entries) with contextWindow and maxOutputTokens pre-filled; no nulls or hardcoded fallbacks in index.ts - Rewrite index.ts to load model list and costs from models.json - Cap gateway model maxOutputTokens at 65536 (API limit for Baseten/Vercel) - Add 'Update models' section to README documenting the generation flow - Add npm run extract-models script
This commit is contained in:
Tao Yang
@@ -9,158 +9,100 @@
|
||||
* 3. Place API key in `~/.commandcode/auth.json` or `~/.pi/agent/auth.json`
|
||||
* as {"apiKey": "user_..."} or {"commandcode": "user_..."}
|
||||
*
|
||||
* Models: deepseek-v4-pro, deepseek-v4-flash, claude-sonnet-4-6, claude-opus-4-7, etc.
|
||||
* Models are sourced from models.json, which is extracted from the command-code
|
||||
* npm package dist file. Run `npx tsx scripts/extract-models.ts` to refresh.
|
||||
*/
|
||||
|
||||
import { calculateCost, createAssistantMessageEventStream } from "@mariozechner/pi-ai"
|
||||
import type { ExtensionAPI } from "@mariozechner/pi-coding-agent"
|
||||
import { readFileSync } from "node:fs";
|
||||
|
||||
import { createStreamCommandCode, DEFAULT_API_BASE } from "./src/core.ts"
|
||||
import { getApiKey, login, refreshToken } from "./src/oauth.ts"
|
||||
import { calculateCost, createAssistantMessageEventStream } from "@mariozechner/pi-ai";
|
||||
import type { ExtensionAPI } from "@mariozechner/pi-coding-agent";
|
||||
|
||||
const API_BASE = process.env.COMMANDCODE_API_BASE ?? DEFAULT_API_BASE
|
||||
import { createStreamCommandCode, DEFAULT_API_BASE } from "./src/core.ts";
|
||||
import { getApiKey, login, refreshToken } from "./src/oauth.ts";
|
||||
|
||||
const API_BASE = process.env.COMMANDCODE_API_BASE ?? DEFAULT_API_BASE;
|
||||
|
||||
// ---------------------------------------------------------------------------
|
||||
// Model definitions
|
||||
// Load model definitions from models.json
|
||||
// ---------------------------------------------------------------------------
|
||||
|
||||
const MODELS = [
|
||||
// Premium (Anthropic)
|
||||
{
|
||||
id: "claude-opus-4-7",
|
||||
name: "Claude Opus 4.7 (CC)",
|
||||
reasoning: true,
|
||||
contextWindow: 200_000,
|
||||
maxTokens: 32_000,
|
||||
},
|
||||
{
|
||||
id: "claude-opus-4-6",
|
||||
name: "Claude Opus 4.6 (CC)",
|
||||
reasoning: true,
|
||||
contextWindow: 200_000,
|
||||
maxTokens: 32_000,
|
||||
},
|
||||
{
|
||||
id: "claude-sonnet-4-6",
|
||||
name: "Claude Sonnet 4.6 (CC)",
|
||||
reasoning: true,
|
||||
contextWindow: 200_000,
|
||||
maxTokens: 16_384,
|
||||
},
|
||||
{
|
||||
id: "claude-haiku-4-5-20251001",
|
||||
name: "Claude Haiku 4.5 (CC)",
|
||||
reasoning: true,
|
||||
contextWindow: 200_000,
|
||||
maxTokens: 8_192,
|
||||
},
|
||||
// Premium (OpenAI)
|
||||
{
|
||||
id: "gpt-5.5",
|
||||
name: "GPT-5.5 (CC)",
|
||||
reasoning: true,
|
||||
contextWindow: 256_000,
|
||||
maxTokens: 128_000,
|
||||
},
|
||||
{
|
||||
id: "gpt-5.4",
|
||||
name: "GPT-5.4 (CC)",
|
||||
reasoning: true,
|
||||
contextWindow: 256_000,
|
||||
maxTokens: 128_000,
|
||||
},
|
||||
{
|
||||
id: "gpt-5.3-codex",
|
||||
name: "GPT-5.3 Codex (CC)",
|
||||
reasoning: true,
|
||||
contextWindow: 256_000,
|
||||
maxTokens: 128_000,
|
||||
},
|
||||
{
|
||||
id: "gpt-5.4-mini",
|
||||
name: "GPT-5.4 Mini (CC)",
|
||||
reasoning: false,
|
||||
contextWindow: 256_000,
|
||||
maxTokens: 128_000,
|
||||
},
|
||||
// Open-source
|
||||
{
|
||||
id: "deepseek/deepseek-v4-pro",
|
||||
name: "DeepSeek V4 Pro (CC)",
|
||||
reasoning: true,
|
||||
contextWindow: 1_000_000,
|
||||
maxTokens: 384_000,
|
||||
},
|
||||
{
|
||||
id: "deepseek/deepseek-v4-flash",
|
||||
name: "DeepSeek V4 Flash (CC)",
|
||||
reasoning: true,
|
||||
contextWindow: 1_000_000,
|
||||
maxTokens: 384_000,
|
||||
},
|
||||
{
|
||||
id: "moonshotai/Kimi-K2.6",
|
||||
name: "Kimi K2.6 (CC)",
|
||||
reasoning: true,
|
||||
contextWindow: 262_144,
|
||||
maxTokens: 131_072,
|
||||
},
|
||||
{
|
||||
id: "moonshotai/Kimi-K2.5",
|
||||
name: "Kimi K2.5 (CC)",
|
||||
reasoning: true,
|
||||
contextWindow: 262_144,
|
||||
maxTokens: 131_072,
|
||||
},
|
||||
{
|
||||
id: "zai-org/GLM-5.1",
|
||||
name: "GLM-5.1 (CC)",
|
||||
reasoning: true,
|
||||
contextWindow: 200_000,
|
||||
maxTokens: 131_072,
|
||||
},
|
||||
{
|
||||
id: "zai-org/GLM-5",
|
||||
name: "GLM-5 (CC)",
|
||||
reasoning: true,
|
||||
contextWindow: 200_000,
|
||||
maxTokens: 131_072,
|
||||
},
|
||||
{
|
||||
id: "MiniMaxAI/MiniMax-M2.7",
|
||||
name: "MiniMax M2.7 (CC)",
|
||||
reasoning: true,
|
||||
contextWindow: 1_048_576,
|
||||
maxTokens: 131_072,
|
||||
},
|
||||
{
|
||||
id: "MiniMaxAI/MiniMax-M2.5",
|
||||
name: "MiniMax M2.5 (CC)",
|
||||
reasoning: true,
|
||||
contextWindow: 1_048_576,
|
||||
maxTokens: 131_072,
|
||||
},
|
||||
{
|
||||
id: "Qwen/Qwen3.6-Max-Preview",
|
||||
name: "Qwen 3.6 Max (CC)",
|
||||
reasoning: true,
|
||||
contextWindow: 1_000_000,
|
||||
maxTokens: 131_072,
|
||||
},
|
||||
{
|
||||
id: "Qwen/Qwen3.6-Plus",
|
||||
name: "Qwen 3.6 Plus (CC)",
|
||||
reasoning: true,
|
||||
contextWindow: 1_000_000,
|
||||
maxTokens: 131_072,
|
||||
},
|
||||
]
|
||||
interface ModelsJson {
|
||||
providers: Record<string, string>;
|
||||
models: Array<{
|
||||
key: string;
|
||||
id: string;
|
||||
provider: string;
|
||||
spec: string;
|
||||
label: string;
|
||||
name: string;
|
||||
description: string;
|
||||
reasoning: boolean;
|
||||
reasoningEfforts: string[] | null;
|
||||
contextWindow: number;
|
||||
maxOutputTokens: number;
|
||||
vendorLabel: string | null;
|
||||
}>;
|
||||
pricing: Array<{
|
||||
provider: string;
|
||||
id: string;
|
||||
category: string;
|
||||
promptCost: number;
|
||||
completionCost: number;
|
||||
cacheWrite5mCost: number;
|
||||
cacheWrite1hCost: number;
|
||||
cacheHitCost: number;
|
||||
}>;
|
||||
}
|
||||
|
||||
const modelsJson: ModelsJson = JSON.parse(
|
||||
readFileSync(new URL("./models.json", import.meta.url), "utf8"),
|
||||
);
|
||||
|
||||
// ---------------------------------------------------------------------------
|
||||
// Build cost lookup (model id -> pricing)
|
||||
// ---------------------------------------------------------------------------
|
||||
|
||||
const costByModelId = new Map<string, ModelsJson["pricing"][number]>();
|
||||
for (const p of modelsJson.pricing) {
|
||||
// Pricing id is like "anthropic:claude-sonnet-4-6"
|
||||
const colonIdx = p.id.indexOf(":");
|
||||
if (colonIdx > 0) {
|
||||
costByModelId.set(p.id.substring(colonIdx + 1), p);
|
||||
}
|
||||
costByModelId.set(p.id, p);
|
||||
}
|
||||
|
||||
// ---------------------------------------------------------------------------
|
||||
// Build pi model list (all defaults come from models.json)
|
||||
// ---------------------------------------------------------------------------
|
||||
|
||||
const MODELS = modelsJson.models.map((m) => {
|
||||
const cost = costByModelId.get(m.id);
|
||||
return {
|
||||
id: m.id,
|
||||
name: `${m.name} (CC)`,
|
||||
reasoning: m.reasoning,
|
||||
contextWindow: m.contextWindow,
|
||||
maxTokens: m.maxOutputTokens,
|
||||
cost: {
|
||||
input: cost?.promptCost ?? 0,
|
||||
output: cost?.completionCost ?? 0,
|
||||
cacheRead: cost?.cacheHitCost ?? 0,
|
||||
cacheWrite: Math.max(cost?.cacheWrite5mCost ?? 0, cost?.cacheWrite1hCost ?? 0),
|
||||
},
|
||||
};
|
||||
});
|
||||
|
||||
// ---------------------------------------------------------------------------
|
||||
// Stream factory
|
||||
// ---------------------------------------------------------------------------
|
||||
|
||||
const streamCommandCode = createStreamCommandCode({
|
||||
createStream: createAssistantMessageEventStream,
|
||||
calculateCost,
|
||||
apiBase: API_BASE,
|
||||
})
|
||||
});
|
||||
|
||||
// ---------------------------------------------------------------------------
|
||||
// Extension entry point
|
||||
@@ -188,10 +130,10 @@ export default function (pi: ExtensionAPI) {
|
||||
id: model.id,
|
||||
name: model.name,
|
||||
reasoning: model.reasoning,
|
||||
input: ["text"],
|
||||
cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },
|
||||
input: ["text"] as const,
|
||||
cost: model.cost,
|
||||
contextWindow: model.contextWindow,
|
||||
maxTokens: model.maxTokens,
|
||||
})),
|
||||
})
|
||||
});
|
||||
}
|
||||
|
||||
Reference in New Issue
Block a user