pi-usereq 0.4.0

package/docs/pi.dev/mom-docs/v86.md

@@ -0,0 +1,319 @@
+# v86 Sandbox Evaluation
+
+v86 is an x86 emulator written in JavaScript/WebAssembly that can run Linux in the browser or Node.js. This document details our evaluation for using it as a sandboxed execution environment.
+
+## Overview
+
+- **What it is**: x86 PC emulator (32-bit, Pentium 4 level)
+- **How it works**: Translates machine code to WebAssembly at runtime
+- **Guest OS**: Alpine Linux 3.21 (32-bit x86)
+- **Available packages**: Node.js 22, Python 3.12, git, curl, etc. (full Alpine repos)
+
+## Key Findings
+
+### What Works
+
+| Feature | Status | Notes |
+|---------|--------|-------|
+| Outbound TCP | ✅ | HTTP, HTTPS, TLS all work |
+| Outbound UDP | ✅ | DNS queries work |
+| WebSocket client | ✅ | Can connect to external WebSocket servers |
+| File I/O | ✅ | 9p filesystem for host<->guest file exchange |
+| State save/restore | ✅ | ~80-100MB state files, instant resume |
+| Package persistence | ✅ | Installed packages persist in saved state |
+| npm install | ✅ | Works (outbound HTTPS) |
+| git clone | ✅ | Works (outbound HTTPS) |
+
+### What Doesn't Work
+
+| Feature | Status | Notes |
+|---------|--------|-------|
+| Inbound connections | ❌ | VM is behind NAT (10.0.2.x), needs port forwarding |
+| ICMP ping | ❌ | Userspace network stack limitation |
+| 64-bit | ❌ | v86 only emulates 32-bit x86 |
+
+## Architecture
+
+```
+┌─────────────────────────────────────────────────────────┐
+│                      Host (Node.js)                      │
+│                                                          │
+│  ┌──────────────┐     ┌─────────────────────────────┐   │
+│  │ rootlessRelay│◄───►│           v86               │   │
+│  │  (WebSocket) │     │  ┌─────────────────────┐    │   │
+│  │              │     │  │   Alpine Linux      │    │   │
+│  │  - DHCP      │     │  │   - Node.js 22      │    │   │
+│  │  - DNS proxy │     │  │   - Python 3.12     │    │   │
+│  │  - NAT       │     │  │   - etc.            │    │   │
+│  └──────────────┘     │  └─────────────────────┘    │   │
+│         │             │            │                 │   │
+│         │             │     9p filesystem            │   │
+│         ▼             │            │                 │   │
+│    Internet           │            ▼                 │   │
+│                       │     Host filesystem          │   │
+│                       └─────────────────────────────┘   │
+└─────────────────────────────────────────────────────────┘
+```
+
+## Components & Sizes
+
+| Component | Size | Purpose |
+|-----------|------|---------|
+| v86.wasm | ~2 MB | x86 emulator |
+| libv86.mjs | ~330 KB | JavaScript runtime |
+| seabios.bin | ~128 KB | BIOS |
+| vgabios.bin | ~36 KB | VGA BIOS |
+| Alpine rootfs | ~57 MB | Compressed filesystem (loaded on-demand) |
+| alpine-fs.json | ~160 KB | Filesystem index |
+| rootlessRelay | ~75 KB | Network relay |
+| **Total** | **~60 MB** | Without saved state |
+| Saved state | ~80-100 MB | Optional, for instant resume |
+
+## Installation
+
+```bash
+npm install v86 ws
+```
+
+## Building the Alpine Image
+
+v86 provides Docker tooling to build the Alpine image:
+
+```bash
+git clone https://github.com/copy/v86.git
+cd v86/tools/docker/alpine
+
+# Edit Dockerfile to add packages:
+# ENV ADDPKGS=nodejs,npm,python3,git,curl
+
+./build.sh
+```
+
+This creates:
+- `images/alpine-fs.json` - Filesystem index
+- `images/alpine-rootfs-flat/` - Compressed file chunks
+
+## Network Relay Setup
+
+v86 needs a network relay for TCP/UDP connectivity. We use rootlessRelay:
+
+```bash
+git clone https://github.com/obegron/rootlessRelay.git
+cd rootlessRelay
+npm install
+```
+
+### Required Patches for Host Access
+
+To allow the VM to connect to host services via the gateway IP (10.0.2.2), apply these patches to `relay.js`:
+
+**Patch 1: Disable reverse TCP handling for gateway (line ~684)**
+```javascript
+// Change:
+if (protocol === 6 && dstIP === GATEWAY_IP) {
+  this.handleReverseTCP(ipPacket);
+  return;
+}
+
+// To:
+if (false && protocol === 6 && dstIP === GATEWAY_IP) { // PATCHED
+  this.handleReverseTCP(ipPacket);
+  return;
+}
+```
+
+**Patch 2: Redirect gateway TCP to localhost (line ~792)**
+```javascript
+// Change:
+const socket = net.connect(dstPort, dstIP, () => {
+
+// To:
+const actualDstIP = dstIP === GATEWAY_IP ? "127.0.0.1" : dstIP;
+const socket = net.connect(dstPort, actualDstIP, () => {
+```
+
+**Patch 3: Redirect gateway UDP to localhost (lines ~1431 and ~1449)**
+```javascript
+// Change:
+this.udpSocket.send(payload, dstPort, dstIP, (err) => {
+
+// To:
+const actualUdpDstIP = dstIP === GATEWAY_IP ? "127.0.0.1" : dstIP;
+this.udpSocket.send(payload, dstPort, actualUdpDstIP, (err) => {
+```
+
+### Starting the Relay
+
+```bash
+ENABLE_WSS=false LOG_LEVEL=1 node relay.js
+# Listens on ws://127.0.0.1:8086/
+```
+
+## Basic Usage
+
+```javascript
+import { V86 } from "v86";
+import path from "node:path";
+
+const emulator = new V86({
+    wasm_path: path.join(__dirname, "node_modules/v86/build/v86.wasm"),
+    bios: { url: path.join(__dirname, "bios/seabios.bin") },
+    vga_bios: { url: path.join(__dirname, "bios/vgabios.bin") },
+    filesystem: {
+        basefs: path.join(__dirname, "images/alpine-fs.json"),
+        baseurl: path.join(__dirname, "images/alpine-rootfs-flat/"),
+    },
+    autostart: true,
+    memory_size: 512 * 1024 * 1024,
+    bzimage_initrd_from_filesystem: true,
+    cmdline: "rw root=host9p rootfstype=9p rootflags=trans=virtio,cache=loose modules=virtio_pci tsc=reliable console=ttyS0",
+    net_device: {
+        type: "virtio",
+        relay_url: "ws://127.0.0.1:8086/",
+    },
+});
+
+// Capture output
+emulator.add_listener("serial0-output-byte", (byte) => {
+    process.stdout.write(String.fromCharCode(byte));
+});
+
+// Send commands
+emulator.serial0_send("echo hello\n");
+```
+
+## Communication Methods
+
+### 1. Serial Console (stdin/stdout)
+
+```javascript
+// Send command
+emulator.serial0_send("ls -la\n");
+
+// Receive output
+let output = "";
+emulator.add_listener("serial0-output-byte", (byte) => {
+    output += String.fromCharCode(byte);
+});
+```
+
+### 2. 9p Filesystem (file I/O)
+
+```javascript
+// Write file to VM
+const data = new TextEncoder().encode("#!/bin/sh\necho hello\n");
+await emulator.create_file("/tmp/script.sh", data);
+
+// Read file from VM
+const result = await emulator.read_file("/tmp/output.txt");
+console.log(new TextDecoder().decode(result));
+```
+
+### 3. Network (TCP to host services)
+
+From inside the VM, connect to `10.0.2.2:PORT` to reach `localhost:PORT` on the host (requires patched relay).
+
+```bash
+# Inside VM
+wget http://10.0.2.2:8080/  # Connects to host's localhost:8080
+```
+
+## State Save/Restore
+
+```javascript
+// Save state (includes all installed packages, files, etc.)
+const state = await emulator.save_state();
+fs.writeFileSync("vm-state.bin", Buffer.from(state));
+
+// Restore state (instant resume, ~2 seconds)
+const stateBuffer = fs.readFileSync("vm-state.bin");
+await emulator.restore_state(stateBuffer.buffer);
+```
+
+## Network Setup Inside VM
+
+After boot, run these commands to enable networking:
+
+```bash
+modprobe virtio-net
+ip link set eth0 up
+udhcpc -i eth0
+```
+
+Or as a one-liner:
+```bash
+modprobe virtio-net && ip link set eth0 up && udhcpc -i eth0
+```
+
+The VM will get IP `10.0.2.15` (or similar) via DHCP from the relay.
+
+## Performance
+
+| Metric | Value |
+|--------|-------|
+| Cold boot | ~20-25 seconds |
+| State restore | ~2-3 seconds |
+| Memory usage | ~512 MB (configurable) |
+
+## Typical Workflow for Mom
+
+1. **First run**:
+   - Start rootlessRelay
+   - Boot v86 with Alpine (~25s)
+   - Setup network
+   - Install needed packages (`apk add nodejs npm python3 git`)
+   - Save state
+
+2. **Subsequent runs**:
+   - Start rootlessRelay
+   - Restore saved state (~2s)
+   - Ready to execute commands
+
+3. **Command execution**:
+   - Send commands via `serial0_send()`
+   - Capture output via `serial0-output-byte` listener
+   - Exchange files via 9p filesystem
+
+## Alternative: fetch Backend (No Relay Needed)
+
+For HTTP-only networking, v86 has a built-in `fetch` backend:
+
+```javascript
+net_device: {
+    type: "virtio",
+    relay_url: "fetch",
+}
+```
+
+This uses the browser/Node.js `fetch()` API for HTTP requests. Limitations:
+- Only HTTP/HTTPS (no raw TCP/UDP)
+- No WebSocket
+- Host access via `http://<port>.external` (e.g., `http://8080.external`)
+
+## Files Reference
+
+After building, you need these files:
+
+```
+project/
+├── node_modules/v86/build/
+│   ├── v86.wasm
+│   └── libv86.mjs
+├── bios/
+│   ├── seabios.bin
+│   └── vgabios.bin
+├── images/
+│   ├── alpine-fs.json
+│   └── alpine-rootfs-flat/
+│       └── *.bin.zst (many files)
+└── rootlessRelay/
+    └── relay.js (patched)
+```
+
+## Resources
+
+- [v86 GitHub](https://github.com/copy/v86)
+- [v86 Networking Docs](https://github.com/copy/v86/blob/master/docs/networking.md)
+- [v86 Alpine Setup](https://github.com/copy/v86/tree/master/tools/docker/alpine)
+- [rootlessRelay](https://github.com/obegron/rootlessRelay)
+- [v86 npm package](https://www.npmjs.com/package/v86)

package/docs/pi.dev/pods-docs/gml-4.5.md

@@ -0,0 +1,189 @@
+# GLM-4.5
+
+[中文阅读](./README_zh.md)
+
+<div align="center">
+<img src=resources/logo.svg width="15%"/>
+</div>
+<p align="center">
+    👋 Join our <a href="resources/WECHAT.md" target="_blank">WeChat</a> or <a href="https://discord.gg/QR7SARHRxK" target="_blank">Discord</a> community.
+    <br>
+    📖 Check out the GLM-4.5 <a href="https://z.ai/blog/glm-4.5" target="_blank">technical blog</a>.
+    <br>
+    📍 Use GLM-4.5 API services on <a href="https://docs.z.ai/guides/llm/glm-4.5">Z.ai API Platform (Global)</a> or <br> <a href="https://docs.bigmodel.cn/cn/guide/models/text/glm-4.5">Zhipu AI Open Platform (Mainland China)</a>.
+    <br>
+    👉 One click to <a href="https://chat.z.ai">GLM-4.5</a>.
+</p>
+
+## Model Introduction
+
+The **GLM-4.5** series models are foundation models designed for intelligent agents. GLM-4.5 has **355** billion total
+parameters with **32** billion active parameters, while GLM-4.5-Air adopts a more compact design with **106** billion
+total parameters and **12** billion active parameters. GLM-4.5 models unify reasoning, coding, and intelligent agent
+capabilities to meet the complex demands of intelligent agent applications.
+
+Both GLM-4.5 and GLM-4.5-Air are hybrid reasoning models that provide two modes: thinking mode for complex reasoning and
+tool usage, and non-thinking mode for immediate responses.
+
+We have open-sourced the base models, hybrid reasoning models, and FP8 versions of the hybrid reasoning models for both
+GLM-4.5 and GLM-4.5-Air. They are released under the MIT open-source license and can be used commercially and for
+secondary development.
+
+As demonstrated in our comprehensive evaluation across 12 industry-standard benchmarks, GLM-4.5 achieves exceptional
+performance with a score of **63.2**, in the **3rd** place among all the proprietary and open-source models. Notably,
+GLM-4.5-Air delivers competitive results at **59.8** while maintaining superior efficiency.
+
+![bench](resources/bench.png)
+
+For more eval results, show cases, and technical details, please visit
+our [technical blog](https://z.ai/blog/glm-4.5). The technical report will be released soon.
+
+The model code, tool parser and reasoning parser can be found in the implementation
+of [transformers](https://github.com/huggingface/transformers/tree/main/src/transformers/models/glm4_moe), [vLLM](https://github.com/vllm-project/vllm/blob/main/vllm/model_executor/models/glm4_moe_mtp.py)
+and [SGLang](https://github.com/sgl-project/sglang/blob/main/python/sglang/srt/models/glm4_moe.py).
+
+## Model Downloads
+
+You can directly experience the model on [Hugging Face](https://huggingface.co/spaces/zai-org/GLM-4.5-Space)
+or [ModelScope](https://modelscope.cn/studios/ZhipuAI/GLM-4.5-Demo) or download the model by following the links below.
+
+| Model            | Download Links                                                                                                                                | Model Size | Precision |
+|------------------|-----------------------------------------------------------------------------------------------------------------------------------------------|------------|-----------|
+| GLM-4.5          | [🤗 Hugging Face](https://huggingface.co/zai-org/GLM-4.5)<br> [🤖 ModelScope](https://modelscope.cn/models/ZhipuAI/GLM-4.5)                   | 355B-A32B  | BF16      |
+| GLM-4.5-Air      | [🤗 Hugging Face](https://huggingface.co/zai-org/GLM-4.5-Air)<br> [🤖 ModelScope](https://modelscope.cn/models/ZhipuAI/GLM-4.5-Air)           | 106B-A12B  | BF16      |
+| GLM-4.5-FP8      | [🤗 Hugging Face](https://huggingface.co/zai-org/GLM-4.5-FP8)<br> [🤖 ModelScope](https://modelscope.cn/models/ZhipuAI/GLM-4.5-FP8)           | 355B-A32B  | FP8       |
+| GLM-4.5-Air-FP8  | [🤗 Hugging Face](https://huggingface.co/zai-org/GLM-4.5-Air-FP8)<br> [🤖 ModelScope](https://modelscope.cn/models/ZhipuAI/GLM-4.5-Air-FP8)   | 106B-A12B  | FP8       |
+| GLM-4.5-Base     | [🤗 Hugging Face](https://huggingface.co/zai-org/GLM-4.5-Base)<br> [🤖 ModelScope](https://modelscope.cn/models/ZhipuAI/GLM-4.5-Base)         | 355B-A32B  | BF16      |
+| GLM-4.5-Air-Base | [🤗 Hugging Face](https://huggingface.co/zai-org/GLM-4.5-Air-Base)<br> [🤖 ModelScope](https://modelscope.cn/models/ZhipuAI/GLM-4.5-Air-Base) | 106B-A12B  | BF16      |
+
+## System Requirements
+
+### Inference
+
+We provide minimum and recommended configurations for "full-featured" model inference. The data in the table below is
+based on the following conditions:
+
+1. All models use MTP layers and specify
+   `--speculative-num-steps 3 --speculative-eagle-topk 1 --speculative-num-draft-tokens 4` to ensure competitive
+   inference speed.
+2. The `cpu-offload` parameter is not used.
+3. Inference batch size does not exceed `8`.
+4. All are executed on devices that natively support FP8 inference, ensuring both weights and cache are in FP8 format.
+5. Server memory must exceed `1T` to ensure normal model loading and operation.
+
+The models can run under the configurations in the table below:
+
+| Model       | Precision | GPU Type and Count   | Test Framework |
+|-------------|-----------|----------------------|----------------|
+| GLM-4.5     | BF16      | H100 x 16 / H200 x 8 | sglang         |
+| GLM-4.5     | FP8       | H100 x 8 / H200 x 4  | sglang         |
+| GLM-4.5-Air | BF16      | H100 x 4 / H200 x 2  | sglang         |
+| GLM-4.5-Air | FP8       | H100 x 2 / H200 x 1  | sglang         |
+
+Under the configurations in the table below, the models can utilize their full 128K context length:
+
+| Model       | Precision | GPU Type and Count    | Test Framework |
+|-------------|-----------|-----------------------|----------------|
+| GLM-4.5     | BF16      | H100 x 32 / H200 x 16 | sglang         |
+| GLM-4.5     | FP8       | H100 x 16 / H200 x 8  | sglang         |
+| GLM-4.5-Air | BF16      | H100 x 8 / H200 x 4   | sglang         |
+| GLM-4.5-Air | FP8       | H100 x 4 / H200 x 2   | sglang         |
+
+### Fine-tuning
+
+The code can run under the configurations in the table below
+using [Llama Factory](https://github.com/hiyouga/LLaMA-Factory):
+
+| Model       | GPU Type and Count | Strategy | Batch Size (per GPU) |
+|-------------|--------------------|----------|----------------------|
+| GLM-4.5     | H100 x 16          | Lora     | 1                    |
+| GLM-4.5-Air | H100 x 4           | Lora     | 1                    |
+
+The code can run under the configurations in the table below using [Swift](https://github.com/modelscope/ms-swift):
+
+| Model       | GPU Type and Count | Strategy | Batch Size (per GPU) |
+|-------------|--------------------|----------|----------------------|
+| GLM-4.5     | H20 (96GiB) x 16   | Lora     | 1                    |
+| GLM-4.5-Air | H20 (96GiB) x 4    | Lora     | 1                    |
+| GLM-4.5     | H20 (96GiB) x 128  | SFT      | 1                    |
+| GLM-4.5-Air | H20 (96GiB) x 32   | SFT      | 1                    |
+| GLM-4.5     | H20 (96GiB) x 128  | RL       | 1                    |
+| GLM-4.5-Air | H20 (96GiB) x 32   | RL       | 1                    |
+
+## Quick Start
+
+Please install the required packages according to `requirements.txt`.
+
+```shell
+pip install -r requirements.txt
+```
+
+### transformers
+
+Please refer to the `trans_infer_cli.py` code in the `inference` folder.
+
+### vLLM
+
++ Both BF16 and FP8 can be started with the following code:
+
+```shell
+vllm serve zai-org/GLM-4.5-Air \
+    --tensor-parallel-size 8 \
+    --tool-call-parser glm45 \
+    --reasoning-parser glm45 \
+    --enable-auto-tool-choice \
+    --served-model-name glm-4.5-air
+```
+
+If you're using 8x H100 GPUs and encounter insufficient memory when running the GLM-4.5 model, you'll need
+`--cpu-offload-gb 16` (only applicable to vLLM).
+
+If you encounter `flash infer` issues, use `VLLM_ATTENTION_BACKEND=XFORMERS` as a temporary replacement. You can also
+specify `TORCH_CUDA_ARCH_LIST='9.0+PTX'` to use `flash infer` (different GPUs have different TORCH_CUDA_ARCH_LIST
+values, please check accordingly).
+
+### SGLang
+
++ BF16
+
+```shell
+python3 -m sglang.launch_server \
+  --model-path zai-org/GLM-4.5-Air \
+  --tp-size 8 \
+  --tool-call-parser glm45  \
+  --reasoning-parser glm45 \
+  --speculative-algorithm EAGLE \
+  --speculative-num-steps 3 \
+  --speculative-eagle-topk 1 \
+  --speculative-num-draft-tokens 4 \
+  --mem-fraction-static 0.7 \
+  --served-model-name glm-4.5-air \
+  --host 0.0.0.0 \
+  --port 8000
+```
+
++ FP8
+
+```shell
+python3 -m sglang.launch_server \
+  --model-path zai-org/GLM-4.5-Air-FP8 \
+  --tp-size 4 \
+  --tool-call-parser glm45  \
+  --reasoning-parser glm45  \
+  --speculative-algorithm EAGLE \
+  --speculative-num-steps 3  \
+  --speculative-eagle-topk 1  \
+  --speculative-num-draft-tokens 4 \
+  --mem-fraction-static 0.7 \
+  --disable-shared-experts-fusion \
+  --served-model-name glm-4.5-air-fp8 \
+  --host 0.0.0.0 \
+  --port 8000
+```
+
+### Request Parameter Instructions
+
++ When using `vLLM` and `SGLang`, thinking mode is enabled by default when sending requests. If you want to disable the
+  thinking switch, you need to add the `extra_body={"chat_template_kwargs": {"enable_thinking": False}}` parameter.
++ Both support tool calling. Please use OpenAI-style tool description format for calls.
++ For specific code, please refer to `api_request.py` in the `inference` folder.