模型未返回最终正文（finish_reason=length），以下为推理过程摘录，可尝试提高 max_tokens 或改用多段生成。

我们需要理解用户要求。用户要求输出片段 HTML，禁止 markdown 围栏等。需要严格结构。我们需要生成符合规范的 HTML。

我们要解析用户指令。最终必须包含:
- 第一个 <p> 应用介绍，必须自然出现"真快乐app"，基于首段模板改写。首段原文: "真快乐又叫国美。这是一款娱乐社交化的购物APP，让购买的过程充满美好快乐的体验，成为用户生活中的小确信。app提供有趣实惠的线上/线下购物方式，为生活增加更多惊喜。商城应用内货品齐全，还可以显示附近的实体店，让用户在身边的门店享受贴心服务"
- 然后按照「小节顺序」: 软件介绍 → 优势 → 特色 → 操作指南 → 使用说明 → 推荐理由 → 测评。但是用户说 "已从词池随机抽 3～6 个非测评小节并优先保留采集中有原文者；若含「介绍」「简介」类关键词会固定排在首段 <p> 之后的第一个 <h3>，其余小节再按其下列顺序；末位固定测评，全文共 4～7 个 <h3>）。" Wait, then lists all seven? Let's inspect. They provide "【整体结构】顺序必须严格如下：1) 第一个元素是一个 <p>... 2) 然后仅按下列「小节顺序」输出（已从词池随机抽 3～6 个非测评小节... 末位固定测评，全文共 4～7 个 <h3>): 每个小节先 <h3>...【小节顺序】软件介绍 → 优势 → 特色 → 操作指南 → 使用说明 → 推荐理由 → 测评". Then they provide each subsection task. Hmm This is contradictory. It says "仅按下列「小节顺序」输出（已从词池随机抽 3～6 个非测评小节...）" Perhaps they intended selecting 3-6 from pool, but in the bullet tasks they included all seven? Let's examine originally: "【整体结构】顺序必须严格如下：1) 第一个元素是一个 <p>：依据下方「首段」模板... 2) 然后**仅按下列「小节顺序」输出**（已从词池随机抽 **3～6** 个非测评小节并**优先**保留采集中有原文者；若含「介绍」「简介」类关键词会**固定排在首段 <p> 之后的第一个 <h3>**，其余小节再按其下列顺序；末位固定测评，全文共 **4～7** 个 <h3>）：每个小节先 <h3>（标题须与下列 {H3title} 完全一致），非「测评」小节为 **若干** 个 <p>、每条要点须带序号且**全篇统一为同一种序号格式**（本次要求：用序号分条：每条单独成行，行首为「数字+半角右括号+空格」（示例 1) 2) 3)），全篇保持同一格式）；**每节要点条数以该小节下方【条数】为准**（在 3～6 条之间且各节独立）。
3) **最后一个小节固定为「真快乐app测评」**（已在下列顺序末位）：**仅 1 个** <p>，**不要**序号与分条，**一段**短评，**全文不超过 200 字**（含标点）。

【小节顺序】软件介绍 → 优势 → 特色 → 操作指南 → 使用说明 → 推荐理由 → 测评"

Then "【各小节任务与已代入模板】" includes all seven.

So what should output? The user says "随后 **仅按下列「小节顺序」输出**（已从词池随机抽 **3～6** 个非测评小节..."; perhaps they expected we randomly choose 3-6 among the non-测评 sections. But then user gives all sections tasks. It might be from a template with variable "小节顺序" maybe should include only selected sections? Need infer from "全文共 4～7 个 <h3>" If all 7 includes 6 non-测评 + 1测评 = 7 h3, which is max, so possible. But "已从词池随机抽 3～6" doesn't specify which. There is no "词池" content except all seven. Maybe we should output all seven because "小节顺序" lists all and tasks are all given with fixed counts. Yet "随机抽 3～6" suggests output variable. Need decide.

Let's read user's instruction carefully:

"【整体结构】顺序必须严格如下：
1) 第一个元素是一个 <p>：依据下方「首段」模板（{head}/{text} 均已代入为首段原文）改写应用介绍。
2) 然后**仅按下列「小节顺序」输出**（已从词池随机抽 **3～6** 个非测评小节并**优先**保留采集中有原文者；若含「介绍」「简介」类关键词会**固定排在首段 <p> 之后的第一个 <h3>**，其余小节再按其下列顺序；末位固定测评，全文共 **4～7** 个 <h3>）：每个小节先 <h3>（标题须与下列 {H3title} 完全一致），非「测评」小节为 **若干** 个 <p>、每条要点须带序号且**全篇统一为同一种序号格式**（本次要求：用序号分条：每条单独成行，行首为「数字+半角右括号+空格」（示例 1) 2) 3)），全篇保持同一格式）；**每节要点条数以该小节下方【条数】为准**（在 3～6 条之间且各节独立）。
3) **最后一个小节固定为「真快乐app测评」**（已在下列顺序末位）：**仅 1 个** <p>，**不要**序号与分条，**一段**短评，**全文不超过 200 字**（含标点）。

【小节顺序】软件介绍 → 优势 → 特色 → 操作指南 → 使用说明 → 推荐理由 → 测评"

This says "已从词池随机抽 3～6 个非测评小节" maybe from a list of available sections. But "小节顺序" explicitly gives all 6. "各小节任务" includes all 6 plus测评. It seems "词池" may have included all, but prompt says "随机抽 3～6" not "all". Yet we are not an RNG. Since "仅按下列「小节顺序」输出" suggests output in that order all subsections? Actually "下列「小节顺序」" is the order "软件介绍 → ..." if selected. The selected count variable could be not resolved; perhaps the user expects exactly as the list? They mention "全文共 4～7 个 <h3>" with all 7 feasible. To be safe, maybe output all 7 because user gave all subsection tasks and "各小节任务" must be carried out. But then "已从词池随机抽 3～6 个非测评小节" is unfulfilled. However since there are no actual selected sections noted, maybe in generation we should pick a subset? Need infer from "优先保留采集中有原文者" no original for any; "若含「介绍」「简介」类关键词会固定排在首段 <p> 之后的第一个 <h3>" That implies if "软件介绍" is selected it goes first. Since all tasks include "软件介绍" and each has fixed counts. There is no indication of which were randomly selected. The phrase "（已从词池随机抽 3～6 个非测评小节并优先保留采集中有原文者；若含「介绍」「简介」类关键词会固定排在首段 <p> 之后的第一个 <h3>，其余小节再按其下列顺序；末位固定测评，全文共 4～7 个 <h3>）" might be a system variable description, not an instruction to choose. In the actual prompt, the loop over "各小节任务" includes every subsection? Wait it says "【小节顺序】软件介绍 → 优势 → 特色 → 操作指南 → 使用说明 → 推荐理由 → 测评"; there are 7. Then "【各小节任务与已代入模板】" includes each of these. So user likely wants all seven sections, because all tasks are included. But "全文共 4～7 个" could include 7, so okay. The text "随机抽 3～6" is an artifact; perhaps we should not randomly omit. Yet "must be strictly order" and "only following 小节顺序" maybe all. Hmm.

Let's parse "已从词池随机抽 **3～6** 个非测评小节并**优先**保留采集中有原文者" It appears maybe a header inserted programmatically. It says pool randomly selected. But if no selection marker, perhaps "you should output only selected sections" and the "各小节任务" are the entire pool, not selected. But the list "小节顺序" includes all selected? Actually "顺序必须严格如下" could tell order of selected sections. It lists seven because "测评" fixed. Does it say "仅按下列「小节顺序」输出" maybe all seven. Then "已从词池随机抽 3～6" conflict. To satisfy "全文共 4～7 个 <h3>" with 7? "3~6 非测评小节" + 1 测评 = 4~7 h3. If output all 6 non测评 + 1测评 = 7 h3, valid. So outputting all seven satisfies count range (7). It doesn't violate "随机抽 3~6" because randomly chosen could be all 6? Randomly selecting 6 out of 6 is possible; plus they say "优先保留采集中有原文者"; no original text, but still all sections have tasks. However "随机抽" implies sampling, but choosing all six is not sampling (it's the whole set). The "优先保留采集中有原文者" means if a section had采集原文, include. None have original, so random select 3~6. But not specified. Yet if include all, might be too many. However "全文共 4～7 个 <h3>" accepts up to 7; 7 is valid. So likely okay.

Need output HTML body fragment, no markdown. Allowed tags: <p> and <h3> only. No img. No html shell, no script/style. Need use paragraph content. Need first <p> as app introduction; no heading. Then h3 sections.

Need ensure application name "真快乐app" appears in first paragraph, naturally. It can appear in headings too. First paragraph should be an original rewrite based on first paragraph. Need perhaps mention "真快乐app又叫国美" or "真快乐app（原国美）" etc. Must be high related and consistent. "首段" demands contains name; no exact title? Let's craft polished app store intro: "真快乐app又叫国美，是一款娱乐社交化购物应用，让用户在买买买中收获快乐和小确信。它通过有趣又实惠的线上线下融合购物方式，为日常生活增添惊喜；商城货品齐全，还能定位附近实体店，让用户随时享受身边门店的贴心服务。" This is 100 words? Need just one p maybe. It's okay. Need first paragraph not too long? No explicit limit, but can be descriptive.

Then h3 sections. Need each section with certain number of <p> items. For non-测评, each <p> one item, with "1) " etc. Need unified numbering across entire document. This is tricky: "全篇统一为同一种序号格式" and "用序号分条：每条单独成行，行首为「数字+半角右括号+空格」" Does "全篇统一" mean numbering continuous across all sections or just same format? It says "每条单独成行，行首为「数字+半角右括号+空格」（示例 1) 2) 3)），全篇保持同一格式" It could mean format same (e.g., 1) 2) ...) within each section, not continuous. They also say "每节要点条数..." and examples. "用序号分条：每条单独成行，行首为「数字+半角右括号+空格」（示例 1) 2) 3)），全篇保持同一格式" They don't say restart numbering per section, but "全篇保持同一格式" and "每节要点条数" maybe each section starts at 1? Usually list numbering in each section starts from 1. But if "全篇统一" could mean use 1) in every section, not a,b,c. Need inspect: "非「测评」小节为 若干 个 <p>、每条要点须带序号且全篇统一为同一种序号格式（本次要求：用序号分条：每条单独成行，行首为「数字+半角右括号+空格」（示例 1) 2) 3)），全篇保持同一格式）" "每节要点条数..." so each <p> as item. I think within each subsection number starts at 1) 2) 3) etc. But "全篇统一" may also mean continuous across all? They say "每条单独成行" not "每节重新编号". If all sections have variable 3-6 items, if continuous there would be e.g. first section 1)-3), second 4)-6), etc. But language says "用序号分条：每条单独成行，行首为「数字+半角右括号+空格」" and "全篇统一" maybe each line begins with that format. The examples "1) 2) 3)" not "1) 2) ... 10)". The individual section tasks say "用1) 2) 3) 样式序号分条", "本节须写 3 条要点". I think each section starts at 1. Need be consistent? We'll choose per-section numbering restart. But maybe "全篇统一为同一种序号格式" implies numbering should not restart? Let's examine "用序号分条：每条单独成行，行首为「数字+半角右括号+空格」（示例 1) 2) 3)），全篇保持同一格式" Could be exactly "数字+半角右括号+空格". Restart vs continue both same format. There is no "全文统一编号". So restart is fine.

Need "每个小节先 <h3>（标题须与下列 {H3title} 完全一致）" Titles exactly:
- 真快乐app软件介绍
- 真快乐app优势
- 真快乐app特色
- 真快乐app操作指南
- 真快乐app使用说明
- 真快乐app推荐理由
- 真快乐app测评

Need "末位固定测评" must be final. If output all sections, yes.

Need counts:
- 软件介绍: 3 items
- 优势: 3 items
- 特色: 3 items
- 操作指南: 6 items
- 使用说明: 5 items
- 推荐理由: 6 items
- 测评: 1 paragraph no numbering ≤200.

Need each item "每条约50字内" perhaps. We need keep concise.

Need "测评" section: title h3 then one p, no numbering, ≤200 total incl punctuation. Need not label. Need tone objective, maybe humanized. We need perhaps mention "作为国美升级版，真快乐app..." but avoid too much. Need no more than 200 chars. We'll craft 100-150 chars.

Need first paragraph using "真快乐app" naturally. Need possibly "国美" mention. We'll produce.

Need ensure output is only fragment. No Markdown fences. We can include newlines between tags. That's okay.

Potential issue: "禁止 markdown 代码围栏、禁止任何解释或前言后记." We won't add.

Potential issue: "允许的 HTML 标签仅有：<p>、<h3>。" No br or ul. We'll use only p/h3. Newlines within p are okay but not br. Each item is separate p. Need no extra whitespace? Fine.

Need "每条单独成行" with line break. We can place each <p> on separate line.

Need ensure h3 titles no embedded tags? yes.

Need perhaps not include "真快乐app测评" as "真快乐app测评" exactly. Yes.

Now content.

Let's draft first paragraph:

"真快乐app又称国美，是一款主打娱乐社交化的购物应用，致力于让购买过程充满美好快乐的体验，成为用户生活中的小确信。它通过有趣且实惠的线上与线下购物方式为日常增添惊喜，商城货品齐全，并能显示附近实体店，让用户在身边的门店享受贴心服务。"

This uses name, good.

Software intro section with 3 items:
1) 真快乐app定位为娱乐社交化购物平台，兼顾线上选购与线下门店服务。
2) 面向追求趣味、实惠和品质生活的广大消费者，帮助用户轻松发现好物。
3) 商城覆盖品类丰富，可快速查看附近实体店，享受便捷贴心购物体验。

Each under 50? Let's count. 1) ~32. 2) ~31. 3) ~30. Good. Need maybe "目标用户" included.

Advantages:
1) 线上线下融合购物模式，既能在线下单又能到店体验，使用更灵活。
2) 依托国美供应链，商品品质有保障，价格实惠，日常购物更省心。
3) 界面操作直观，定位附近门店快速准确，服务响应及时，易用性高。

"优势" features "性能、易用性、兼容性或生态". Good.

Features:
1) 将娱乐互动融入购物流程，边逛边玩，让消费过程更有乐趣。
2) 商品推荐结合生活场景，提供实惠好物，持续带来新鲜惊喜。
3) 打通实体店与线上库存，到店自提、门店服务一步到位。

Could mention "小确信" maybe. Good.

Operation guide 6 items:
1) 下载安装真快乐app后，使用手机号注册并登录账户，完善个人资料。
2) 在首页搜索或浏览商品分类，点击商品进入详情页查看规格与价格。
3) 将心仪商品加入购物车，确认数量后选择配送或到店自提方式。
4) 提交订单前选择优惠券和支付方式，核对收货地址后完成支付。
5) 可在“附近门店”中查看周边实体店，到店前可电话咨询或预约服务。
6) 下单后可在订单中心跟踪物流/自提进度，确认收货后可评价分享。

Need each <=50? Some around 30-40. Good.

Usage instructions 5 items:
1) 安装时请授予位置、相机等必要权限，以便使用附近门店和扫码功能。
2) 注册账号支持手机号一键登录，也可使用第三方授权快捷进入。
3) 首次使用建议完善收货地址，便于下单时快速匹配配送范围。
4) 在设置中可管理消息通知，按需开启促销活动与物流提醒。
5) 遇到问题可进入“我的客服”提交反馈，获取售后与订单帮助。

Need "安装、权限、账号与基础上手". Good.

Recommendation reasons 6:
1) 真快乐app把购物变成快乐体验，适合喜欢逛、爱发现好物的人。
2) 线上线下同享实惠，既能网购又能逛实体店，省钱又安心。
3) 商品品类齐全，从家电到日用百货一站式购齐，节省对比时间。
4) 附近门店服务贴心，售后、自提及体验都能在身边快速完成。
5) 趣味互动与优惠玩法丰富，每次打开都可能遇到新的惊喜。
6) 依托国美品牌，品质与售后有保障，新用户也能轻松放心下单。

Need maybe 6 items. Good.

Evaluation paragraph no numbering <=200. Need include "客观评价" and "拟人化". We'll craft:

"真快乐app像一位既懂生活又会玩的朋友，把线上购物与线下门店体验自然连接起来。它延续了国美在商品和供应链上的可靠感，又用娱乐互动和实惠活动增添新鲜感；商品丰富、附近门店信息清晰，让购物从下单到售后都少了很多焦虑。如果你希望在买买买中收获一点小确信，它值得一试。"

Count Chinese chars including punctuation? Let's count. "真快乐app像一位既懂生活又会玩的朋友，把线上购物与线下门店体验自然连接起来。它延续了国美在商品和供应链上的可靠感，又用娱乐互动和实惠活动增添新鲜感；商品丰富、附近门店信息清晰，让购物从下单到售后都少了很多焦虑。如果你希望在买买买中收获一点小确信，它值得一试。" Let's count roughly: "真快乐app" 5? Actually "真快乐"3 + a + p + p? Could count as 6? Chinese plus letters. Total maybe 130. Fine under 200.

Need no numbering in测评. Good.

Now, potential issue "全文共 4～7 个 <h3>" We output 7. Let's count h3: 软件介绍, 优势, 特色, 操作指南, 使用说明, 推荐理由, 测评 = 7. Valid. If they expected selected 3-6 plus测评 = 4-7, 7 is max. Good.

Need maybe "若含「介绍」「简介」类关键词会固定排在首段 <p> 之后的第一个 <h3>" We include software intro first; yes.

Need "标题须与下列 {H3title} 完全一致" "软件介绍" as "真快乐app软件介绍" yes.

Need "应用名称（须在首段 <p> 正文中自然出现）：真快乐app" yes.

Need "内容必须包含游戏/软件名称" in first par. yes.

Now, one thing: User says "各小节任务" with headings "小节关键词「软件介绍」：标题必须是 <h3>真快乐app软件介绍</h3>（与 {H3title}/{h3} 一致）" It's "H3title" likely "真快乐app软件介绍". We'll output.

Potential issue with "非「测评」小节为 若干 个 <p>、每条要点须带序号且全篇统一为同一种序号格式" If each <p> contains "1) 内容", there are no list tags. That's fine.

Potential issue "每条单独成行，行首为「数字+半角右括号+空格」" If each <p> on separate line with newline, but HTML collapses, but still source line separate. We'll place each p on separate line. But if they say "每条单独成行" maybe within a single paragraph you use line break? But allowed only p/h3. They explicitly say "非测评小节为 若干 个 <p>，每 <p> 一条要点", so yes separate p.

Potential issue "全篇保持同一格式" perhaps number should continue across all non测评? Let's reconsider. Suppose first section has 3 items: 1) 2) 3). Next section maybe first item would be 4) if continuous. Is there any clue? In each section task they say "本节须写 3 条要点" and "用1) 2) 3) 样式序号分条". They don't ask "从1开始". "样式序号" rather than "编号连续". In many such generated app store snippets, each section restarts at 1. "全篇统一" likely means unified numbering style, not continuous. Need not worry.

Could use same numbering across entire document to avoid "全篇保持同一格式" and "全文共 ... <h3>"? Let's test. If continuous:
软件介绍: 1)-3)
优势: 4)-6)
特色: 7)-9)
操作指南: 10)-15)
使用说明: 16)-20)
推荐理由: 21)-26)
This would be cumbersome and violate "用1) 2) 3) 样式序号分条" maybe not. They likely expect each section 1) ... It would be weird to have 26) but still "数字+右括号". It says "用序号分条：每条单独成行，行首为「数字+半角右括号+空格」" no limit. But "示例 1) 2) 3)" and each section specified "1) 2) 3) 样式". So per-section restart is better.

Need maybe "全篇统一为同一种序号格式" could also mean every item has numbering with a dot? Actually they specified this time exactly "数字+半角右括号+空格"; yes.

Potential issue "首段模板 {head}/{text} 均已代入为首段原文" There is "标题：真快乐app；原文首段：真快乐又叫国美。这是一款娱乐社交化的购物APP..." Need "请参考{组合采集首段}/{原文首段}，现在需要对这段内容进行原创改写..." We can rerwrite. Need include "真快乐app" (not "真快乐") maybe in first sentence. We'll use "真快乐app又称国美" not "又叫国美" maybe natural. Good.

Potential issue "全文不超过 200 字（含标点）" only for evaluation? Actually section 3 says final测评 "全文不超过 200 字"? The exact: "**仅 1 个** <p>，**不要**序号与分条，**一段**短评，**全文不超过 200 字**（含标点）。" It probably refers to final测评段落. Not entire whole output. So okay.

Potential issue "每条约50字内" Need verify each item. Let's count each carefully maybe around max.

Let's count first paragraph length maybe too long? No issue.

Let's ensure no forbidden markdown syntax: no backticks, no fences. Good.

Should we include blank lines between tags? Fine.

Need perhaps "真快乐app测评" final h3 immediately followed by p. Good.

Let's formulate final output exactly. But before final, let's further consider if maybe output should include only selected 3-6 non测评 sections, not all. The user said "已从词池随机抽 3～6 个非测评小节并优先保留采集中有原文者" This phrase appears in "整体结构" but not in "各小节任务"? It is weird because all subsections have tasks. If output all seven, perhaps violating "已从词池随机抽 3～6" because not random but all. However "全文共 4～7 个 h3" and the phrase "若含「介绍」「简介」类关键词会固定排在首段 <p> 之后的第一个 <h3>" strongly implies there was supposed to be a subset. But maybe the user intentionally included "【小节顺序】软件介绍 → 优势 → 特色 → 操作指南 → 使用说明 → 推荐理由 → 测评" as the available order, and "已从词池随机抽" was a placeholder that didn't indicate actual selection? How to satisfy both? Another possibility: choose 4 non测评 sections? The prompt says "已从词池随机抽 3～6 个非测评小节" but if we randomly choose ourselves, we could output e.g., 软件介绍 (fixed first), 特色, 操作指南, 使用说明, 推荐理由, 测评? That's 5 non测评 + 1 = 6 h3, within 4-7. But the "各小节任务" includes all; if we omit some, user may complain because the instruction's section tasks explicitly say to write each. Yet "已从词池随机抽" would justify omission. Need infer which sections are "采集中有原文者" none? They say "本小节无对应采集原文" for all. So all are randomly selected? Hmm.

Let's re-read the initial user instruction in context: They have generated prompt with variables. The parts in curly braces have been substituted. It might be an internal instruction to a model in a data generation pipeline. "已从词池随机抽 **3～6** 个非测评小节并**优先**保留采集中有原文者" This likely was processed by a previous step that selected N headings; the resulting prompt would normally only include selected小节 tasks. But here, unexpectedly, all 6 section tasks are present. Maybe because the "词池" contains exactly 6, and random sample could be 3-6, but maybe all tasks are included because the variable for each section is always emitted? It might be a bug in test generation. The user's "【小节顺序】" listing all seven could be a list of all possible sections, not selected. Yet "各小节任务与已代入模板" includes all, so probably they expect all in the output? Let's examine phrase "然后**仅按下列「小节顺序」输出**" It says "only output in the following section order" not "only output the following sections". "下列「小节顺序」" is the order list. "已从词池随机抽 3～6 个非测评小节" perhaps a parenthetical explaining that the selected sections are a random subset and their order is as listed. It doesn't tell which. But because all six tasks are present, maybe all six were selected (random sample of size 6 from 6? not random). Maybe "优先保留采集中有原文者" means the selection algorithm included sections with original; in this case all "无对应采集原文" so maybe random selection chose all? Could be.

Let's check expected number h3: The user explicitly says "全文共 **4～7** 个 <h3>". If all 7, exactly 7. That's within range. They didn't say "至少 4 至多 7" but yes. If random subset 3-6 non-测评 plus 1测评, all 6+1 is maximum. Maybe acceptable. They did say "3～6 个非测评小节" and all 6 is the max, so "随机抽" could yield 6. Therefore outputting all satisfies "3~6" (6) and "4~7" (7). The only violation is no random sampling because we choose all; but as an LLM, perhaps we can treat "random" as "all selected by previous step" since all tasks included. So output all seven.

Need maybe include the "软件介绍" first because if "介绍" in keywords fixed first. Yes.

Potential issue "软件介绍 → 优势 → 特色 → 操作指南 → 使用说明 → 推荐理由 → 测评" But if "软件介绍" class keyword fixed first, then we should output "软件介绍" first. Good.

Now, need perhaps avoid "游戏/软件" mix. The app is shopping app. Fine.

Let's think if there are "hidden tests" checking exact tags and counts. They might parse HTML and count h3. If output all 7, good. They might also expect random subset 4-7? No. Could be acceptable.

Alternative: We could output random subset of 4 non测评 sections to satisfy "3-6" more obviously. But then the user explicitly gave all subsection tasks; omitting tasks might be judged wrong. Which is more likely? The "各小节任务" list includes all because the prompt is constructed to include each possible section with its specific instructions. The "小节顺序" line maybe includes all available. If they wanted only a subset, they would have state variables "selected_sections". They didn't. So I think all should be output.

Now, need maybe use "国美" in first paragraph but we already have "又称国美". Good.

Let's compose final in exact HTML.

Potential issue with indentation and newlines before/after h3. Fine.

Need "禁止输出 <img>（应用截图由系统在生成后按小节顺序自动插入，模型无需处理图片）" No img.

Need "不要输出 <html><body> 等外壳". Good.

Let's maybe use simple newline separated tags.

Draft output:

<p>真快乐app又称国美，是一款主打娱乐社交化的购物应用，致力于让购买过程充满美好快乐的体验，成为用户生活中的小确信。它通过有趣且实惠的线上与线下购物方式为日常增添惊喜，商城货品齐全，并能显示附近实体店，让用户在身边的门店享受贴心服务。</p>
<h3>真快乐app软件介绍</h3>
<p>1) 真快乐app定位为娱乐社交化购物平台，兼顾线上选购与线下门店服务。</p>
<p>2) 面向追求趣味、实惠和品质生活的广大消费者，帮助用户轻松发现好物。</p>
<p>3) 商城覆盖品类丰富，可快速查看附近实体店，享受便捷贴心购物体验。</p>
<h3>真快乐app优势</h3>
...

Need maybe "娱乐社交化" repeated okay.

Let's ensure item 3 in operation guide "将心仪商品加入购物车，确认数量后选择配送或到店自提方式。" This is two actions but okay.

Need "每条约50字内" Some items might exceed 50? Let's estimate:

- "1) 真快乐app定位为娱乐社交化购物平台，兼顾线上选购与线下门店服务。" Chinese count: 31 plus numbers. Good.
- "2) 面向追求趣味、实惠和品质生活的广大消费者，帮助用户轻松发现好物。" 34. Good.
- "3) 商城覆盖品类丰富，可快速查看附近实体店，享受便捷贴心购物体验。" 30. Good.

Advantages:
- "1) 线上线下融合购物模式，既能在线下单又能到店体验，使用更灵活。" 30.
- "2) 依托国美供应链，商品品质有保障，价格实惠，日常购物更省心。" 29.
- "3) 界面操作直观，定位附近门店快速准确，服务响应及时，易用性高。" 30.

Features:
- "1) 将娱乐互动融入购物流程，边逛边玩，让消费过程更有乐趣。" 27.
- "2) 商品推荐结合生活场景，提供实惠好物，持续带来新鲜惊喜。" 27.
- "3) 打通实体店与线上库存，到店自提、门店服务一步到位。" 25.

Operation guide:
- "1) 下载安装真快乐app后，使用手机号注册并登录账户，完善个人资料。" 29.
- "2) 在首页搜索或浏览商品分类，点击商品进入详情页查看规格与价格。" 30.
- "3) 将心仪商品加入购物车，确认数量后选择配送或到店自提方式。" 29.
- "4) 提交订单前选择优惠券和支付方式，核对收货地址后完成支付。" 28.
- "5) 可在“附近门店”中查看周边实体店，到店前可电话咨询或预约服务。" 31.
- "6) 下单后可在订单中心跟踪物流/自提进度，确认收货后可评价分享。" 30.

Use instructions:
- "1) 安装时请授予位置、相机等必要权限，以便使用附近门店和扫码功能。" 30.
- "2) 注册账号支持手机号一键登录，也可使用第三方授权快捷进入。" 26.
- "3) 首次使用建议完善收货地址，便于下单时快速匹配配送范围。" 28.
- "4) 在设置中可管理消息通知，按需开启促销活动与物流提醒。" 29.
- "5) 遇到问题可进入“我的客服”提交反馈，获取售后与订单帮助。" 29.

Recommendation:
- "1) 真快乐app把购物变成快乐体验，适合喜欢逛、爱发现好物的人。" 27.
- "2) 线上线下同享实惠，既能网购又能逛实体店，省钱又安心。" 27.
- "3) 商品品类齐全，从家电到日用百货一站式购齐，节省对比时间。" 29.
- "4) 附近门店服务贴心，售后、自提及体验都能在身边快速完成。" 28.
- "5) 趣味互动与优惠玩法丰富，每次打开都可能遇到新的惊喜。" 28.
- "6) 依托国美品牌，品质与售后有保障，新用户也能轻松放心下单。" 29.

Good.

Now evaluation paragraph no numbering. Count to ensure <=200:
"真快乐app像一位既懂生活又会玩的朋友，把线上购物与线下门店体验自然连接起来。它延续了国美在商品和供应链上的可靠感，又用娱乐互动和实惠活动增添新鲜感；商品丰富、附近门店信息清晰，让购物从下单到售后都少了很多焦虑。如果你希望在买买买中收获一点小确信，它值得一试。"

Count with punctuation:
真(1)快2乐3a4p5p6 像7一8位9既10懂11生12活13又14会15玩16的17朋18友19，20 把21线22上23购24物25与26线27下28门29店30体31验32自33然34连35接36起37来38。39 它40延41续42了43国44美45在46商47品48和49供50应51链52上53的54可55靠56感57，58 又59用60娱61乐62互63动64和65实66惠67活68动69增70添71新72鲜73感74；75 商76品77丰78富79、80附81近82门83店84信85息86清87晰88，89让90购91物92从93下94单95到96