Files
HeurAMS/examples/repo.ipynb
2026-01-01 20:18:18 +08:00

367 lines
9.6 KiB
Plaintext

{
"cells": [
{
"cell_type": "markdown",
"id": "51b89355",
"metadata": {},
"source": [
"# 演练场\n",
"此笔记本将带你了解 repomgr 与 particles 对象相关操作"
]
},
{
"cell_type": "markdown",
"id": "f5c49014",
"metadata": {},
"source": [
"# 从一个例子开始\n",
"## 了解文件结构\n",
"了解一下文件结构"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "a5ed9864",
"metadata": {},
"outputs": [],
"source": [
"!tree # 了解文件结构"
]
},
{
"cell_type": "markdown",
"id": "4e10922b",
"metadata": {},
"source": [
"如果你先前运行了单元格, 请运行下面一格清理."
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "9777730e",
"metadata": {},
"outputs": [],
"source": [
"!rm -rf test_new_repo\n",
"!rm -rf heurams.log*"
]
},
{
"cell_type": "markdown",
"id": "058c098f",
"metadata": {},
"source": [
"## 导入模块\n",
"导入所需模块, 你会看到欢迎信息, 标示了库所使用的配置. \n",
"HeurAMS 在基础设施也使用配置文件实现隐式的依赖注入. "
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "bf1b00c8",
"metadata": {},
"outputs": [],
"source": [
"import heurams.kernel.repolib as repolib # 这是 RepoLib 子模块, 用于管理和结构化 repo(中文含义: 仓库) 数据结构与本地文件间的联系\n",
"import heurams.kernel.particles as pt # 这是 Particles(中文含义: 粒子) 子模块, 用于运行时的记忆管理操作\n",
"from pathlib import Path # 这是 Python 的 Pathlib 模块, 用于表示文件路径, 在整个项目中, 都使用此模块表示路径"
]
},
{
"cell_type": "markdown",
"id": "ea1f68bb",
"metadata": {},
"source": [
"## 运行时检查\n",
"如你所见, repo 在文件系统内存储为一个文件夹. \n",
"因此在载入之前, 首先要检查这是否是一个合乎标准的 repo 文件夹. "
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "897b62d7",
"metadata": {},
"outputs": [],
"source": [
"is_vaild = repolib.Repo.check_repodir(Path(\"./test_repo\"))\n",
"print(f\"这是一个 {'合规' if is_vaild else '不合规'} 的 repo!\")"
]
},
{
"cell_type": "markdown",
"id": "24a19991",
"metadata": {},
"source": [
"## 加载仓库\n",
"接下来, 正式加载 repo."
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "708ae7e4",
"metadata": {},
"outputs": [],
"source": [
"test_repo = repolib.Repo.create_from_repodir(Path(\"./test_repo\"))"
]
},
{
"cell_type": "markdown",
"id": "474f8eb7",
"metadata": {},
"source": [
"## 导出为字典\n",
"作为一个数据容器, repo 相应地建立了导入和导出的功能. \n",
"我们刚刚从本地文件夹导入了一个 repo. \n",
"现在试试导出为一个字典."
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "a11115fb",
"metadata": {},
"outputs": [],
"source": [
"test_repo_dic = test_repo.export_to_single_dict()\n",
"from pprint import pprint\n",
"pprint(test_repo_dic)"
]
},
{
"cell_type": "markdown",
"id": "35a2e06f",
"metadata": {},
"source": [
"## 持久化与部分保存\n",
"如你所见, 所有内容被结构化地输出了! \n",
"\n",
"现在写回到文件夹! \n",
"\n",
"我们注意到, 并非所有的内容都要被修改. \n",
"我们可以只保存接受修改的一部分, 默认情况下, 是迭代的记忆数据(algodata). \n",
"这就是为什么我们一般不使用单个 json 或 toml 来存储 repo.\n",
"\n",
"persist_to_repodir 接受两个可选参数: \n",
"- save_list: 默认为 [\"algodata\"], 是要持久化的数据.\n",
"- source: 默认为原目录, 你也可以手动指定为其他文件夹(通过 Path)\n",
"\n",
"现在做一些演练, 我们将创建一个位于 test_new_repo 的\"克隆\", 此时我们!\n",
"除非文件夹已经存在, Repo 对象将会为你自动创建新文件夹."
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "05eeaacc",
"metadata": {},
"outputs": [],
"source": [
"test_repo.persist_to_repodir(save_list=[\"schedule\", \"payload\", \"manifest\", \"typedef\", \"algodata\"], source=Path(\"test_new_repo\"))\n",
"!tree"
]
},
{
"cell_type": "markdown",
"id": "059d7bdf",
"metadata": {},
"source": [
"如你所见, test_new_repo 已被生成!"
]
},
{
"cell_type": "markdown",
"id": "4ef8925c",
"metadata": {},
"source": [
"# 数据结构\n",
"现在讲解 repo 的数据结构"
]
},
{
"cell_type": "markdown",
"id": "c19fed95",
"metadata": {},
"source": [
"## Lict 对象\n",
"Lict 对象集成了部分列表和字典的功能, 数据在这两种风格的 API 间都可用, 且修改是同步的. \n",
"Lict 默认情况下不会保存序列顺序, 而是在列表形式下, 自动按索引字符序排布, 详情请参阅源代码. \n",
"现在导入并初始化一个 Lict 对象:"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "7e88bd7c",
"metadata": {},
"outputs": [],
"source": [
"from heurams.utils.lict import Lict\n",
"lct = Lict() # 空的\n",
"lct = Lict(initlist=[(\"name\", \"tom\"), (\"age\", 12), (\"enemy\", \"jerry\")]) # 基于列表\n",
"print(lct)\n",
"lct = Lict(initdict={\"name\": \"tom\", \"age\": 12, \"enemy\": \"jerry\"}) # 基于字典\n",
"print(lct)\n"
]
},
{
"cell_type": "markdown",
"id": "4d760bf9",
"metadata": {},
"source": [
"### 输出形式\n",
"lct 的\"官方\"输出形式是列表形式\n",
"你也可以选择输出字典形式"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "248f6cba",
"metadata": {},
"outputs": [],
"source": [
"print(lct.dicted_data)"
]
},
{
"cell_type": "markdown",
"id": "29dce184",
"metadata": {},
"source": [
"### dicted_data 属性与修改方式\n",
"dicted_data 属性是一个字典, 它自动同步来自 Lict 对象操作的修改.\n",
"一个注意事项: 不要直接修改 dicted_data, 这将不会触发同步 hook.\n",
"如果你一定要这样做, 请在完事后手动运行同步 hook.\n",
"推荐的修改方式是直接把 lct 当作一个字典"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "a0eb07a7",
"metadata": {},
"outputs": [],
"source": [
"# 由于 jupyter 的环境处理, 请不要重复运行此单元格, 如果想再看一遍, 请重启 jupyter 后再全部运行\n",
"\n",
"# 错误的方式\n",
"lct.dicted_data[\"type\"] = \"cat\"\n",
"print(lct) # 将不会同步修改\n",
"\n",
"# 不推荐, 但可用的方式\n",
"lct.dicted_data[\"type\"] = \"cat\"\n",
"lct._sync_based_on_dict()\n",
"print(lct)\n",
"\n",
"# 推荐方式\n",
"lct['is_human'] = False\n",
"print(lct)"
]
},
{
"cell_type": "markdown",
"id": "2337d113",
"metadata": {},
"source": [
"### data 属性与修改方式\n",
"data 属性是一个列表, 它自动同步来自 Lict 对象操作的修改.\n",
"一个注意事项: 不要直接修改 data, 这将不会触发同步 hook, 并且可能破坏排序.\n",
"如果你一定要这样做, 请在完事后手动运行同步 hook 和 sort, 此处不演示.\n",
"推荐的修改方式是直接把 lct 当作一个列表, 且避免使用索引修改"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "0ab442d4",
"metadata": {},
"outputs": [],
"source": [
"# 由于 jupyter 的环境处理, 请不要重复运行此单元格, 如果想再看一遍, 请重启 jupyter 后再全部运行\n",
"\n",
"# 唯一推荐方式\n",
"lct.append(('enemy_2', 'spike'))\n",
"print(lct.dicted_data)"
]
},
{
"cell_type": "markdown",
"id": "a3383f59",
"metadata": {},
"source": [
"### 多面手\n",
"Lict 有一些很酷的功能\n",
"详情请看源文件\n",
"此处是一些例子"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "f3ca752f",
"metadata": {},
"outputs": [],
"source": [
"lct = Lict(initdict={'age': 12, 'enemy': 'jerry', 'is_human': False, 'name': 'tom', 'type': 'cat', 'enemy_2': 'spike'})\n",
"print(lct)\n",
"print(lct.dicted_data)\n",
"print(\"------\")\n",
"for i in lct:\n",
" print(i)\n",
"print(len(lct))\n",
"while len(lct) > 0:\n",
" print(lct.pop())\n",
" print(lct)\n",
"lct = Lict(initdict={'age': 12, 'enemy': 'jerry', 'is_human': False, 'name': 'tom', 'type': 'cat', 'enemy_2': 'spike'})\n",
"..."
]
},
{
"cell_type": "markdown",
"id": "2d6d3483",
"metadata": {},
"source": [
"关爱环境 从你我做起"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "773bf99c",
"metadata": {},
"outputs": [],
"source": [
"!rm -rf test_new_repo\n",
"!rm -rf heurams.log*"
]
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.13.11"
}
},
"nbformat": 4,
"nbformat_minor": 5
}