init commit

2025-12-08 15:20:22 +08:00
commit 1d0077510a
28 changed files with 9050234 additions and 0 deletions
--- a/README.md
+++ b/README.md
@@ -0,0 +1,194 @@
+# crawler_tiktok
+
+TikTok 数据抓取脚本，分为两阶段：
+- 按关键词搜索视频链接并生成快照（`links`）
+- 根据视频链接抓取评论与二级回复并生成快照与可选 CSV（`comments`）
+
+全项目基于 Python 标准库实现（`urllib`、`threading` 等），不依赖第三方包。
+
+## 目录结构
+
+```
+crawler_tiktok/
+├─ core/                # cURL 文本解析与请求发送
+│  └─ curl.py
+├─ tiktok/              # TikTok 业务逻辑
+│  ├─ search.py         # 关键词搜索视频链接
+│  └─ comments.py       # 抓取评论与二级回复
+├─ data/                # 示例数据与输出目录
+│  ├─ 1.text            # cURL 文本（包含多个 curl 命令块）
+│  ├─ keyword.txt       # 关键词文件（每行一个关键词）
+│  ├─ urls.json         # 链接搜索快照输出（示例已有）
+│  ├─ comments.csv      # 评论 CSV 输出（可选）
+│  └─ store.py          # 统一的快照写入工具
+├─ utils/               # 通用 IO 工具
+│  └─ io.py
+├─ main.py              # 命令行入口（子命令：links / comments / all）
+└─ __init__.py          # 包入口
+```
+
+## 准备工作
+
+- 安装 Python（建议 3.8+）
+- 准备 `data/1.text`：
+  - 打开浏览器访问 TikTok，登录后在开发者工具的 Network 面板选中相关请求，使用 “Copy as cURL” 复制。
+  - 将“评论接口”的 `curl ...` 放在第一段，“搜索接口”的 `curl ...` 放在第二段；两段之间可直接换行即可。
+  - 保留请求头（尤其是 `cookie`）以便接口正常返回。
+- 准备关键词文件 `data/keyword.txt`（每行一个关键词），或使用命令行传参。
+
+## 快速开始
+
+在仓库根目录（必须为 `D:\work\crawler_tiktok`）直接运行脚本：
+
+```
+python main.py -h
+```
+
+
+### 1) 搜索视频链接（links）
+
+将关键词并发搜索，统一去重并保存到 `urls.json`。
+
+```
+python main.py links \
+  --keywords-file data\keyword.txt \
+  --file-path data\1.text \
+  --out data\urls.json \
+  --max-pages 50 \
+  --count 12 \
+  --timeout 30 \
+  --workers 5
+```
+
+可选：
+- 通过 `--keyword` 重复传入多个关键词（可与 `--keywords-file` 混用）
+- `--keywords` 逗号分隔的关键词字符串
+
+输出 `urls.json` 结构示例：
+
+```json
+{
+  "keywords": ["xxx", "yyy"],
+  "items": [
+    {"keyword": "xxx", "count": 10, "links": ["https://www.tiktok.com/@user/video/123" ...]},
+    {"keyword": "yyy", "count": 8,  "links": [ ... ]}
+  ],
+  "total_count": 17855,
+  "links": ["https://www.tiktok.com/@user/video/123", ...]
+}
+```
+
+### 2) 抓取评论与回复（comments）
+
+从链接快照读取链接，抓取主评论与二级回复，并保存 JSON 与可选 CSV。
+
+```
+python main.py comments \
+  --links-json data\urls.json \
+  --out data\tik_comments.json \
+  --file-path data\1.text \
+  --count 100 \
+  --pages 100 \
+  --timeout 30 \
+  --reply-count 100 \
+  --reply-pages 100 \
+  --csv data\comments.csv \
+  --workers 8
+```
+
+输出 `tik_comments.json` 结构示例：
+
+```json
+{
+  "items": [
+    {
+      "link": "https://www.tiktok.com/@user/video/123",
+      "count": 42,
+      "comments": [
+        {
+          "cid": "xxx",
+          "text": "...",
+          "user": {"unique_id": "..."},
+          "replies": [{"text": "..."}, ...],
+          "reply_count": 3
+        }
+      ]
+    }
+  ]
+}
+```
+
+若提供 `--csv`，会将主评论与回复分别以 `username,text` 形式追加到该文件。
+
+### 3) 全流程一体化（all）
+
+一次性串联链接搜索与评论抓取，适合流水线执行：
+
+```
+python main.py all \
+  --keywords-file data\keyword.txt \
+  --file-path data\1.text \
+  --links-out data\urls.json \
+  --search-max-pages 50 \
+  --search-count 12 \
+  --search-timeout 30 \
+  --search-workers 5 \
+  --comments-out data\tik_comments.json \
+  --comments-count 100 \
+  --comments-pages 100 \
+  --comments-timeout 30 \
+  --comments-limit 1000 \
+  --reply-count 100 \
+  --reply-pages 100 \
+  --reply-limit 2000 \
+  --csv data\comments.csv \
+  --comments-workers 8
+```
+
+### 4) 写入 MySQL（从 CSV 导入）
+
+在 `D:\work\crawler_tiktok` 下执行：
+
+```
+pip install pymysql
+python main.py mysql \
+  --csv data\comments.csv \
+  --host localhost \
+  --port 3306 \
+  --user root \
+  --password <你的密码> \
+  --database crawler_tiktok \
+  --table comments
+```
+
+若数据库不存在，请先在 MySQL 中创建：
+
+```
+CREATE DATABASE IF NOT EXISTS `crawler_tiktok` DEFAULT CHARACTER SET utf8mb4 COLLATE utf8mb4_unicode_ci;
+```
+
+导入会在指定库中自动建表（如不存在），并批量插入 `username,text` 两列数据。
+
+## 重要参数说明
+
+- `--keyword / --keywords / --keywords-file`：三种方式提供关键词，最终会合并并去重。
+- `--file-path`：cURL 文本文件路径（包含多个 `curl ...` 命令块）。
+  - 第 1 块作为评论接口基准。
+  - 第 2 块作为搜索接口基准。
+- 搜索阶段：`--max-pages` 分页轮次上限；`--count` 每页条数（默认从 URL 中推断，通常为 12）；`--workers` 并发线程数。
+- 评论阶段：`--pages` 评论分页上限；`--count` 每页评论数；`--reply-count` / `--reply-pages` 回复分页与每页数；`--workers` 并发抓取线程数。
+- `--timeout`：请求超时秒数。
+- `--csv`：若提供则会将主评论与回复按 `username,text` 追加到该 CSV。
+
+## 输出文件约定
+
+- `data/urls.json`：链接搜索快照，包含 `keywords/items/total_count/links`。
+- `data/tik_comments.json`：评论抓取快照，包含 `items`（每项含 `link/count/comments`）。
+- `data/comments.csv`：CSV 格式的评论与回复（用户名、文本）。
+
+## 常见问题
+
+- 返回为空或报错：优先检查 `data/1.text` 的 cURL 是否有效，`cookie` 是否过期。
+- 速率限制：适当降低 `--workers`、提高 `--timeout`，或分批执行。
+- Windows 路径：示例中使用反斜杠；若在类 Unix 环境，改为 `/`。
+- 进度打印：抓取过程会打印 START/DONE/ERROR 以及评论统计，便于观察执行状态。
--- a/init.py
+++ b/init.py
@@ -0,0 +1,10 @@
+"""crawler_tiktok 包
+
+该包用于从 TikTok 搜集视频链接并抓取评论与回复。
+模块划分：
+- core：基础能力（如从 curl 文本解析 URL 与请求头）
+- tiktok：与 TikTok 相关的抓取逻辑（搜索、评论）
+- utils：通用 IO 工具（JSON/CSV 读写、关键词文件加载）
+- data：数据快照的写入工具（store），以及示例数据文件
+入口：main.py 提供命令行子命令 links/comments/all。
+"""
--- a/pycache/init.cpython-312.pyc
+++ b/pycache/init.cpython-312.pyc
--- a/pycache/main.cpython-312.pyc
+++ b/pycache/main.cpython-312.pyc
--- a/core/pycache/curl.cpython-312.pyc
+++ b/core/pycache/curl.cpython-312.pyc
--- a/core/pycache/har.cpython-312.pyc
+++ b/core/pycache/har.cpython-312.pyc
--- a/core/pycache/store.cpython-312.pyc
+++ b/core/pycache/store.cpython-312.pyc
--- a/core/curl.py
+++ b/core/curl.py
@@ -0,0 +1,74 @@
+import re
+import json
+from urllib.request import Request, urlopen
+
+def _split_curl_blocks(text):
+    """按出现的 `curl ` 关键字切分文本为多个命令块"""
+    blocks = []
+    indices = [m.start() for m in re.finditer(r"\bcurl\s", text)]
+    if not indices:
+        return blocks
+    for i, start in enumerate(indices):
+        end = indices[i + 1] if i + 1 < len(indices) else len(text)
+        blocks.append(text[start:end])
+    return blocks
+
+def _parse_block(block):
+    """从单个 curl 命令块中解析 URL 与头部
+
+    返回：`{'url': str, 'headers': dict}`，若无法解析 URL 返回 None
+    """
+    url_m = re.search(r"curl\s+['\"](.*?)['\"]", block, re.S)
+    if not url_m:
+        return None
+    url = url_m.group(1)
+    headers = {}
+    for hm in re.finditer(r"-H\s+['\"]([^:]+):\s*(.*?)['\"]", block):
+        k = hm.group(1).strip()
+        v = hm.group(2).strip()
+        headers[k.lower()] = v
+    cm = re.search(r"-b\s+['\"](.*?)['\"]", block, re.S)
+    if cm:
+        headers['cookie'] = cm.group(1)
+    return {'url': url, 'headers': headers}
+
+def parse_curl_file(file_path):
+    """读取 curl 文本文件并解析为请求描述列表
+
+    参数：`file_path` 文件路径
+    返回：列表，每项包含 `url` 与 `headers`
+    """
+    with open(file_path, 'r', encoding='utf-8') as f:
+        text = f.read()
+    blocks = _split_curl_blocks(text)
+    result = []
+    for b in blocks:
+        parsed = _parse_block(b)
+        if parsed:
+            result.append(parsed)
+    return result
+
+def fetch_from_curl(file_path, index=0, timeout=30):
+    """按索引选取解析出的请求并发起 GET
+
+    参数：`index` 为第几个 curl 块；`timeout` 请求超时秒数
+    返回：尝试解析为 JSON，失败则返回原始 bytes
+    """
+    reqs = parse_curl_file(file_path)
+    if not reqs or index < 0 or index >= len(reqs):
+        return None
+    item = reqs[index]
+    req = Request(item['url'], headers=item['headers'], method='GET')
+    with urlopen(req, timeout=timeout) as resp:
+        data = resp.read()
+        try:
+            return json.loads(data.decode('utf-8', errors='ignore'))
+        except Exception:
+            return data
+"""curl 文本解析与请求发送工具
+
+职责：
+- 将包含多个 curl 命令的文本切分为块
+- 从每个块解析 URL 与请求头（含 Cookie）
+- 基于解析结果发起 GET 请求并尝试返回 JSON
+"""
--- a/data/1.text
+++ b/data/1.text
@@ -0,0 +1,28 @@
+curl 'https://www.tiktok.com/api/comment/list/?WebIdLastTime=1762843898&aid=1988&app_language=en&app_name=tiktok_web&aweme_id=7554313767425985806&browser_language=zh-CN&browser_name=Mozilla&browser_online=true&browser_platform=Win32&browser_version=5.0%20%28Windows%20NT%2010.0%3B%20Win64%3B%20x64%29%20AppleWebKit%2F537.36%20%28KHTML%2C%20like%20Gecko%29%20Chrome%2F142.0.0.0%20Safari%2F537.36&channel=tiktok_web&cookie_enabled=true&count=20&cursor=0&data_collection_enabled=true&device_id=7571356851431851534&device_platform=web_pc&focus_state=true&from_page=video&history_len=4&is_fullscreen=false&is_page_visible=true&odinId=7571357724144124941&os=windows&priority_region=US&referer=https%3A%2F%2Fwww.google.com%2F&region=US&root_referer=https%3A%2F%2Fwww.google.com%2F&screen_height=1080&screen_width=1920&tz_name=Asia%2FShanghai&user_is_login=true&verifyFp=verify_mi5rex4u_QQ3WUuF2_qkrc_4zFV_9lPB_6ZiBsQO9Yg1Y&webcast_language=en&msToken=HGxJ50IHbcJxidsAy4biksW4jCfUpjG5IOfoNZd9m24WyL0muFz7f02qUT2A4HCKPQPheRtCr66460XMCJQ9mCplXR1zk1fKK81mU65TLKdczaVqauDay_1qAol348Cg_iQaiK74qPZ0EoJBKu_iZbCATA==&X-Bogus=DFSzsIVuMqxANCCACObDM-ZLJrPe&X-Gnarly=MxLz6L1B1jRWAkdMAhXfoRcZW7o7E89jQUXQepysu7jSC47hCDAgLaFj6ATg13br-ct2WppjvVuo3DrB5foJoo3XOjJH6TVzfVkPLs8Sw47ja/0uC5DB6DPtfPWekO9g9-ZviZeREnpG/N2SRXbqDr0-Go5o0OzoRp9wdRUoLSAM5nbo0niphLjDxyOzdsW/RqxqQNFbBnJJkNqIH4TiXbQmNafqX1Yk5cGaIFH7FHcjZsYRbtA2gc8cXePp2guxR5cXDepaBF2Wgsdmu8VM2q8ed8o7ohQXi56GiUuUXjLJJ130mlhYHJHgYHtxYynbeRw=' \
+  -H 'accept: */*' \
+  -H 'accept-language: zh-CN,zh;q=0.9' \
+  -b 'passport_csrf_token=fc48fa188f4d67baad733476f64baccf; passport_csrf_token_default=fc48fa188f4d67baad733476f64baccf; last_login_method=email; delay_guest_mode_vid=5; living_user_id=480687482446; tiktok_webapp_theme_source=auto; tt_chain_token=tTxp2ztaIaxTaXsGrJb+8Q==; tiktok_webapp_lang=zh-Hans; d_ticket=64bffcf490d4c9c7839c94bfa06f09cf36a43; ttwid=1%7C28VHivouaKIrGON9d-ZudJgYZUKezdC-9xuqts9saLQ%7C1763362651%7C2cd9c33c1a5733387a276a04d1af0af2026a80003217b55a841cb13520f1347f; myCookie=rap; fblo_1862952583919182=y; tiktok_webapp_theme=light; multi_sids=7571357724144124941%3Ab0685b23eb2eeb5f5e0a5604801f365b; cmpl_token=AgQQAPNSF-RO0rksMxu3N50083NFwqLP_4zZYKPv8w; sid_guard=b0685b23eb2eeb5f5e0a5604801f365b%7C1763435980%7C15552000%7CSun%2C+17-May-2026+03%3A19%3A40+GMT; uid_tt=b8d8d5d8b062bd2f3badb4e51420dc711f3305fec2a664f03306a64342912e7e; uid_tt_ss=b8d8d5d8b062bd2f3badb4e51420dc711f3305fec2a664f03306a64342912e7e; sid_tt=b0685b23eb2eeb5f5e0a5604801f365b; sessionid=b0685b23eb2eeb5f5e0a5604801f365b; sessionid_ss=b0685b23eb2eeb5f5e0a5604801f365b; tt_session_tlb_tag=sttt%7C1%7CsGhbI-su619eClYEgB82W__________pPerzW7vBXrU5jjfVhTeyC35OugALhjpAGeyHj7inp30%3D; sid_ucp_v1=1.0.1-KDVjMGU0Y2Q1OTg2ZjI0ODI2MjI2NWFmZDY0YWYyMmE0ZTkzY2M5M2QKIgiNiJX4w7e3iWkQzMvvyAYYswsgDDDgu8vIBjgBQOoHSAQQBBoHdXNlYXN0NSIgYjA2ODViMjNlYjJlZWI1ZjVlMGE1NjA0ODAxZjM2NWIyTgogV1B2-bFMjyKUv3NQtyvtMGmk7ui-Sl9aVTQStE9c4gkSIMJacGh0QuV1QyQEcXmFWpSTFvdWSbfDoe3vFgrRtb2YGAEiBnRpa3Rvaw; ssid_ucp_v1=1.0.1-KDVjMGU0Y2Q1OTg2ZjI0ODI2MjI2NWFmZDY0YWYyMmE0ZTkzY2M5M2QKIgiNiJX4w7e3iWkQzMvvyAYYswsgDDDgu8vIBjgBQOoHSAQQBBoHdXNlYXN0NSIgYjA2ODViMjNlYjJlZWI1ZjVlMGE1NjA0ODAxZjM2NWIyTgogV1B2-bFMjyKUv3NQtyvtMGmk7ui-Sl9aVTQStE9c4gkSIMJacGh0QuV1QyQEcXmFWpSTFvdWSbfDoe3vFgrRtb2YGAEiBnRpa3Rvaw; store-idc=useast5; store-country-code=us; store-country-code-src=uid; tt-target-idc=useast8; tt-target-idc-sign=jh5EgVgJzs2Zhfd2qcTVn1799rd4vK_dkjNT5E1hkY8ey6Iuuuo2qPfonmOJN-73-SEUPvtKH8L04xVHFdfuxGD4jEaT9iJGdE634_1n9RDR1aV8X7xR36LWBtgYwCfK96M28ozQElXgVDFsHS5jNIH0Jfq5gaisBGFWCAz7zEHd3YwWFjSVW96udWsQjGHM_y0UYLKmGEwmNh3nCmKOGntgfvFHrzuxfYL2T6upJ8x8WMb3GG-tGXKw4N05kaWH3LJCY3hGiCOIBX3s8_n0jvn1PLu9yiOUiF2f-K63HqcnVxPsuYfg8iYE47R08TALOuvQE9CFArejv2TMFIjiGKi-MB3BlWwwCUmUSbTyoTLBzSuKT5Elynh0l1JVMxrXlqZn39OMcs8_AB0n_RyyAF3FH9pV0sQ2lza6iJtZim0hmnAuTn8C26lyuss0BP3vlSPtp26rNMyqs1uZlpIclUzI7hmlBV6HNph2l1oBp7QMbvCQLFIGYMJXTL5kP_kX; tt_csrf_token=2alNSmcj-CbgYZfsy7TgGoLanfvaF_S5Ya4M; passport_fe_beating_status=true; ttwid=1%7C28VHivouaKIrGON9d-ZudJgYZUKezdC-9xuqts9saLQ%7C1763542066%7C310b419382cc569f618a86d012ca09f9c470719d9734b7ef5e637af0851cc2fc; odin_tt=081974ccd8c05aa99c9def51a2ae7bff2f2eff4b7187eaefc9035f6f29b3cf23b9e7e874eddd540c813b1aeacd524755eeea3b487b1604ab69034829839420a35eb841c2c1660309fb2e8733f8cec76e; store-country-sign=MEIEDEmOSATz1HVkE69LLQQgQ-xv4GcuilURIjdWweEXq86fX22G46h6wFIvL1YxVVUEEOiSOHWEhgwxBRqtF3ZvKh4; s_v_web_id=verify_mi5rex4u_QQ3WUuF2_qkrc_4zFV_9lPB_6ZiBsQO9Yg1Y; msToken=HGxJ50IHbcJxidsAy4biksW4jCfUpjG5IOfoNZd9m24WyL0muFz7f02qUT2A4HCKPQPheRtCr66460XMCJQ9mCplXR1zk1fKK81mU65TLKdczaVqauDay_1qAol348Cg_iQaiK74qPZ0EoJBKu_iZbCATA==; msToken=jWvit9euZoPFKgYZamkCxcNRJFbxE2efJZTwjJRsUOFUCqfJXFXMDMp6zFbk7VT7czHFPTMr0hdTmSdJybWT0mwnHYYrra7EQWKq9cwTr0NAdoyzkp-xrBuSlxD6a4Fcai6a72-qqTi1FxkAWjmrliiqwg==' \
+  -H 'priority: u=1, i' \
+  -H 'referer: https://www.tiktok.com/@user73001399001191/video/7561810082636287246' \
+  -H 'sec-ch-ua: "Chromium";v="142", "Google Chrome";v="142", "Not_A Brand";v="99"' \
+  -H 'sec-ch-ua-mobile: ?0' \
+  -H 'sec-ch-ua-platform: "Windows"' \
+  -H 'sec-fetch-dest: empty' \
+  -H 'sec-fetch-mode: cors' \
+  -H 'sec-fetch-site: same-origin' \
+  -H 'user-agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/142.0.0.0 Safari/537.36'
+
+
+curl 'https://www.tiktok.com/api/search/general/full/?WebIdLastTime=1762843898&aid=1988&app_language=en&app_name=tiktok_web&browser_language=zh-CN&browser_name=Mozilla&browser_online=true&browser_platform=Win32&browser_version=5.0%20%28Windows%20NT%2010.0%3B%20Win64%3B%20x64%29%20AppleWebKit%2F537.36%20%28KHTML%2C%20like%20Gecko%29%20Chrome%2F142.0.0.0%20Safari%2F537.36&channel=tiktok_web&client_ab_versions=70508271%2C72437276%2C73547759%2C73720540%2C74444736%2C74446915%2C74465410%2C74627577%2C74679798%2C74703728%2C74744616%2C74746519%2C74757744%2C74780477%2C74782564%2C74793838%2C74798355%2C74803471%2C74808328%2C74824020%2C74843467%2C74852654%2C74860161%2C74879745%2C74879783%2C74882809%2C74891662%2C74902367%2C74926160%2C74928117%2C74935708%2C74936938%2C74970253%2C74972148%2C74973673%2C74976255%2C74980175%2C74983940%2C74994853%2C75001423%2C75005876%2C70405643%2C70772958%2C71057832%2C71200802%2C71381811%2C71516509%2C71803300%2C71962127%2C72360691%2C72361743%2C72408100%2C72854054%2C72892778%2C73171280%2C73208420%2C73989921%2C74276218%2C74611443%2C74844724&cookie_enabled=true&count=16&data_collection_enabled=true&device_id=7571356851431851534&device_platform=web_pc&device_type=web_h265&focus_state=true&from_page=search&history_len=5&is_fullscreen=false&is_page_visible=true&keyword=%E6%B6%82%E9%B8%A6%E7%BB%98%E7%94%BB&odinId=7571357724144124941&offset=0&os=windows&priority_region=US&referer=https%3A%2F%2Fwww.google.com%2F&region=US&root_referer=https%3A%2F%2Fwww.google.com%2F&screen_height=1080&screen_width=1920&search_source=search_history&tz_name=Asia%2FShanghai&user_is_login=true&verifyFp=verify_mi5rex4u_QQ3WUuF2_qkrc_4zFV_9lPB_6ZiBsQO9Yg1Y&web_search_code=%7B%22tiktok%22%3A%7B%22client_params_x%22%3A%7B%22search_engine%22%3A%7B%22ies_mt_user_live_video_card_use_libra%22%3A1%2C%22mt_search_general_user_live_card%22%3A1%7D%7D%2C%22search_server%22%3A%7B%7D%7D%7D&webcast_language=en&msToken=Evo6YZn35dd6dAsaNUBm7WHaOCixR84Hwjo6DVlNBGE0L56xiDF_dmDWfyJIJGq8LDEsjNm5G9H3uMP9LlVsCunVwx0lMEnriQWWWuzpN7Xp4j0Fj5wXbgMEqU9KMd5YfkZ1iqFubWhu99nvT06p5qpUeg==&X-Bogus=DFSzsIVu7-iANCCACObDh-ZLJrOb&X-Gnarly=M5J8rVZ10jjW3H5JvTrrLC6MGn7Qq4X0NFfuLZ1UYP2F5Tyem4CUCigEriTnj4Ui3kZdlhNogxKstfzoLHWeSKWiubEdsYZpiegkx-Ot2OUSwbyC9mcwB8T80j7nJpzf6tMOisjjbGiGzYQDJuNrqgxehrCDKUfdA6CbLeoguoGy7XQjTDxmg3/VsSqdhziaenBlm72xVj0GyLUEgrboEwzp11Xphma3Qo8b-/uiMZWQDNyJaC7rcb11dW-ffpSTMrXvf6EU6QXJav2NYvS2gNjMJBhMf15s0-NNQQIC-USLgeAWEo5Wj-gXgn/YmaUd7hZ=' \
+  -H 'accept: */*' \
+  -H 'accept-language: zh-CN,zh;q=0.9' \
+  -b 'passport_csrf_token=fc48fa188f4d67baad733476f64baccf; passport_csrf_token_default=fc48fa188f4d67baad733476f64baccf; last_login_method=email; delay_guest_mode_vid=5; living_user_id=480687482446; tiktok_webapp_theme_source=auto; tt_chain_token=tTxp2ztaIaxTaXsGrJb+8Q==; tiktok_webapp_lang=zh-Hans; d_ticket=64bffcf490d4c9c7839c94bfa06f09cf36a43; ttwid=1%7C28VHivouaKIrGON9d-ZudJgYZUKezdC-9xuqts9saLQ%7C1763362651%7C2cd9c33c1a5733387a276a04d1af0af2026a80003217b55a841cb13520f1347f; myCookie=rap; fblo_1862952583919182=y; tiktok_webapp_theme=light; multi_sids=7571357724144124941%3Ab0685b23eb2eeb5f5e0a5604801f365b; cmpl_token=AgQQAPNSF-RO0rksMxu3N50083NFwqLP_4zZYKPv8w; sid_guard=b0685b23eb2eeb5f5e0a5604801f365b%7C1763435980%7C15552000%7CSun%2C+17-May-2026+03%3A19%3A40+GMT; uid_tt=b8d8d5d8b062bd2f3badb4e51420dc711f3305fec2a664f03306a64342912e7e; uid_tt_ss=b8d8d5d8b062bd2f3badb4e51420dc711f3305fec2a664f03306a64342912e7e; sid_tt=b0685b23eb2eeb5f5e0a5604801f365b; sessionid=b0685b23eb2eeb5f5e0a5604801f365b; sessionid_ss=b0685b23eb2eeb5f5e0a5604801f365b; tt_session_tlb_tag=sttt%7C1%7CsGhbI-su619eClYEgB82W__________pPerzW7vBXrU5jjfVhTeyC35OugALhjpAGeyHj7inp30%3D; sid_ucp_v1=1.0.1-KDVjMGU0Y2Q1OTg2ZjI0ODI2MjI2NWFmZDY0YWYyMmE0ZTkzY2M5M2QKIgiNiJX4w7e3iWkQzMvvyAYYswsgDDDgu8vIBjgBQOoHSAQQBBoHdXNlYXN0NSIgYjA2ODViMjNlYjJlZWI1ZjVlMGE1NjA0ODAxZjM2NWIyTgogV1B2-bFMjyKUv3NQtyvtMGmk7ui-Sl9aVTQStE9c4gkSIMJacGh0QuV1QyQEcXmFWpSTFvdWSbfDoe3vFgrRtb2YGAEiBnRpa3Rvaw; ssid_ucp_v1=1.0.1-KDVjMGU0Y2Q1OTg2ZjI0ODI2MjI2NWFmZDY0YWYyMmE0ZTkzY2M5M2QKIgiNiJX4w7e3iWkQzMvvyAYYswsgDDDgu8vIBjgBQOoHSAQQBBoHdXNlYXN0NSIgYjA2ODViMjNlYjJlZWI1ZjVlMGE1NjA0ODAxZjM2NWIyTgogV1B2-bFMjyKUv3NQtyvtMGmk7ui-Sl9aVTQStE9c4gkSIMJacGh0QuV1QyQEcXmFWpSTFvdWSbfDoe3vFgrRtb2YGAEiBnRpa3Rvaw; store-idc=useast5; store-country-code=us; store-country-code-src=uid; tt-target-idc=useast8; tt-target-idc-sign=jh5EgVgJzs2Zhfd2qcTVn1799rd4vK_dkjNT5E1hkY8ey6Iuuuo2qPfonmOJN-73-SEUPvtKH8L04xVHFdfuxGD4jEaT9iJGdE634_1n9RDR1aV8X7xR36LWBtgYwCfK96M28ozQElXgVDFsHS5jNIH0Jfq5gaisBGFWCAz7zEHd3YwWFjSVW96udWsQjGHM_y0UYLKmGEwmNh3nCmKOGntgfvFHrzuxfYL2T6upJ8x8WMb3GG-tGXKw4N05kaWH3LJCY3hGiCOIBX3s8_n0jvn1PLu9yiOUiF2f-K63HqcnVxPsuYfg8iYE47R08TALOuvQE9CFArejv2TMFIjiGKi-MB3BlWwwCUmUSbTyoTLBzSuKT5Elynh0l1JVMxrXlqZn39OMcs8_AB0n_RyyAF3FH9pV0sQ2lza6iJtZim0hmnAuTn8C26lyuss0BP3vlSPtp26rNMyqs1uZlpIclUzI7hmlBV6HNph2l1oBp7QMbvCQLFIGYMJXTL5kP_kX; tt_csrf_token=2alNSmcj-CbgYZfsy7TgGoLanfvaF_S5Ya4M; passport_fe_beating_status=true; ttwid=1%7C28VHivouaKIrGON9d-ZudJgYZUKezdC-9xuqts9saLQ%7C1763542066%7C310b419382cc569f618a86d012ca09f9c470719d9734b7ef5e637af0851cc2fc; odin_tt=081974ccd8c05aa99c9def51a2ae7bff2f2eff4b7187eaefc9035f6f29b3cf23b9e7e874eddd540c813b1aeacd524755eeea3b487b1604ab69034829839420a35eb841c2c1660309fb2e8733f8cec76e; store-country-sign=MEIEDEmOSATz1HVkE69LLQQgQ-xv4GcuilURIjdWweEXq86fX22G46h6wFIvL1YxVVUEEOiSOHWEhgwxBRqtF3ZvKh4; s_v_web_id=verify_mi5rex4u_QQ3WUuF2_qkrc_4zFV_9lPB_6ZiBsQO9Yg1Y; msToken=imJTbrm-SxNJWwT4U4KOTLwsN5UnjQQzm-bHzXVVdiKnymtLbyQXbM_dziPgdrG3rcvbHJf_WWoBIuJ8AegguxYLz_gA0dQNa8suc7aNA3RA_Z7FqCUJd7iO4TJX5lU3dG0Ahchd28z0ip0HADyOygq2sQ==; msToken=Na9bB8_PXmSHmEprxHatJ3iAi6DYMS4DGaKymzxCSv5ho-vkxkGi2Oh4LHpL3LntQloywKO0p5gTBoGjA3BKW7uguLGHfS6FiTPzo5JkbwAMPkYyGdaoQh3yikWSJFGPNMpdrwaN8-ta5IcfWZlfv_QQcQ==' \
+  -H 'priority: u=1, i' \
+  -H 'referer: https://www.tiktok.com/search?q=%E6%B6%82%E9%B8%A6%E7%BB%98%E7%94%BB&t=1763542137936' \
+  -H 'sec-ch-ua: "Chromium";v="142", "Google Chrome";v="142", "Not_A Brand";v="99"' \
+  -H 'sec-ch-ua-mobile: ?0' \
+  -H 'sec-ch-ua-platform: "Windows"' \
+  -H 'sec-fetch-dest: empty' \
+  -H 'sec-fetch-mode: cors' \
+  -H 'sec-fetch-site: same-origin' \
+  -H 'user-agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/142.0.0.0 Safari/537.36'
--- a/data/pycache/store.cpython-312.pyc
+++ b/data/pycache/store.cpython-312.pyc
--- a/data/comments.csv
+++ b/data/comments.csv
--- a/data/key_comment.csv
+++ b/data/key_comment.csv
--- a/data/keyword.txt
+++ b/data/keyword.txt
@@ -0,0 +1,596 @@
+direct liquid soft head acrylic marker pen
+guangna direct liquid soft head acrylic marker pen
+guangna direct liquid soft head acrylic marker pen 24 colors
+@huangaa_3 guangna direct liquid soft head acrylic marker pen
+how to draw with water based markers
+acrylic marker holder
+acrylic pen markers
+liquid acrylic pens
+direct liquid pen
+direct liquid acrylic marker pen japan
+direct liquid acrylic marker pen japanese
+direct liquid acrylic marker pen jumbo
+direct liquid acrylic marker pen
+acrylic marker pens
+liquitex acrylic markers review
+acrylic marker art
+liquitex acrylic marker
+direct liquid acrylic marker pen instructions
+direct liquid acrylic marker pen ink
+direct liquid acrylic marker pen ii
+direct liquid acrylic marker pen instructions pdf
+direct liquid acrylic marker pen in usa
+direct liquid acrylic marker pen msds
+direct liquid acrylic marker pen markers
+direct liquid acrylic marker pen michaels
+direct liquid acrylic marker pen msds sheet
+direct liquid acrylic marker pen large
+direct liquid acrylic marker pen lowes
+direct liquid acrylic marker pen label
+direct liquid acrylic marker pen liquid
+direct liquid acrylic marker pen light
+direct liquid acrylic marker pen liner
+direct liquid acrylic marker pen liner review
+what pen to use to write on acrylic
+liquid marker pen
+direct liquid acrylic marker pen kit
+direct liquid acrylic marker pen kit instructions
+direct liquid acrylic marker pen orange
+direct liquid acrylic marker pen only
+direct liquid acrylic marker pen on amazon
+direct liquid acrylic marker pen oil
+direct liquid acrylic marker pen on sale
+direct liquid acrylic marker pen off clothes
+direct liquid acrylic marker pen off golf balls
+direct liquid acrylic marker pen on walls
+direct liquid acrylic marker pen pack
+direct liquid acrylic marker pen pink
+direct liquid acrylic marker pen pen
+direct liquid acrylic marker pen purple
+direct liquid acrylic marker pen price
+direct liquid acrylic marker pen paper mate
+direct liquid acrylic marker pen quality
+direct liquid acrylic marker pen quick dry
+direct liquid acrylic marker pen qvc
+direct liquid acrylic marker pen quizlet
+direct liquid acrylic marker pen quiz
+direct liquid acrylic marker pen refill
+direct liquid acrylic marker pen review
+direct liquid acrylic marker pen reddit
+direct liquid acrylic marker pen red
+direct liquid acrylic marker pen target
+direct liquid acrylic marker pen tip
+direct liquid acrylic marker pen type
+direct liquid acrylic marker pen tint
+direct liquid acrylic marker pen use
+direct liquid acrylic marker pen usa
+direct liquid acrylic marker pen us
+direct liquid acrylic marker pen walmart
+direct liquid acrylic marker pen white
+direct liquid acrylic marker pen wholesale
+direct liquid acrylic marker pen waterproof
+direct liquid acrylic marker pen walgreens
+direct liquid acrylic marker pen video
+direct liquid acrylic marker pen vintage
+direct liquid acrylic marker pen vs
+direct liquid acrylic marker pen volume
+direct liquid acrylic marker pen vs regular
+direct liquid acrylic marker pen vape
+direct liquid acrylic marker pen xl
+direct liquid acrylic marker pen x2
+acrylic marker diy
+diy acrylic paint markers
+direct liquid acrylic marker pen zoom
+direct liquid acrylic marker pen zero
+direct liquid acrylic marker pen zipper
+direct liquid acrylic marker pen zip
+direct liquid acrylic marker pen zoominfo
+acrylic marker paper
+acrylic marker painting
+acrylic liquid pens
+acrylic markers blick
+what markers can you use on acrylic
+are liquid chalk markers permanent
+acrylic marker edding
+direct liquid acrylic marker pen amazon
+direct liquid acrylic marker pen amazon prime
+direct liquid acrylic marker pen acrylic
+direct liquid acrylic marker pen app
+direct liquid acrylic marker pen art
+acrylic marker fine tip
+acrylic paint markers fine tip
+are sharpie gel pens waterproof
+liqui-mark gel pens
+acrylic marker graffiti
+acrylic paint markers how to use
+acrylic ink marker
+krink acrylic markers
+permanent marker on acrylic plastic
+liqui-mark permanent markers
+liquitex acrylic paint markers
+marker acrylic
+acrylic paint marker waterproof
+acrylic pen marker
+direct liquid acrylic marker pen black
+direct liquid acrylic marker pen bulk
+direct liquid acrylic marker pen blue
+direct liquid acrylic marker pen brand
+direct liquid acrylic marker pen brand name
+direct liquid acrylic marker pen bleeding through paper
+direct liquid acrylic marker pen bleeding through paint
+direct liquid acrylic marker pen brush photoshop
+oil based marker on acrylic paint
+acrylic markers on plastic
+acrylic paint marker refill
+are sharpies water based
+are sharpies oil or water based
+acrylic marker liquitex
+liquitex acrylic pen
+direct liquid acrylic marker pen directions
+direct liquid acrylic marker pen dollar tree
+direct liquid acrylic marker pen dispenser
+direct liquid acrylic marker pen directions for use
+direct liquid acrylic marker pen dry
+direct liquid acrylic marker pen disguises
+direct liquid acrylic marker pen depot
+direct liquid acrylic marker pens fine line
+direct liquid acrylic marker pens fine point
+direct liquid acrylic marker pens fine tip
+direct liquid acrylic marker pens for painting
+direct liquid acrylic marker pen ebay
+direct liquid acrylic marker pen ewg
+direct liquid acrylic marker pen elite
+direct liquid acrylic marker pen elite review
+direct liquid acrylic marker pen elite vaporizer
+uni acrylic markers
+using acrylic markers
+what marker to use on acrylic
+what type of marker to use on acrylic
+what marker works on plastic
+where to buy acrylic markers
+where to buy acrylic paint pens
+are acrylic markers permanent
+acrylic marker uses
+acrylic marker tutorial
+can you use permanent marker on acrylic paint
+who acrylic markers
+who sells liquitex acrylic paint
+are acrylic markers water based
+is acrylic liquid monomer
+are acrylic pens oil based
+are acrylic paint markers waterproof
+can you use permanent marker on acrylic
+can dry erase markers be used on acrylic
+can acrylic markers be used on fabric
+acrylic vs oil marker
+acrylic markers vs acrylic paint
+acrylic paint marker vs oil paint marker
+are acrylic paint markers permanent
+will dry erase markers work on plexiglass
+will permanent marker stick to plastic
+will permanent marker stay on silicone
+does dry erase markers work on plexiglass
+worst acrylic paint
+worst acrylic paint brands
+worst acrylic powder
+art supplies acrylic marker testing comparison
+acrylic paint markers permanent
+do dry erase markers work on acrylic
+do acrylic markers work on fabric
+do acrylic paint pens work on plastic
+do dry erase markers work on plexiglass
+best acrylic markers
+best acrylic markers for artists
+best acrylic paint pens for plastic
+top rated acrylic paint markers
+top rated acrylic paint pens
+can acrylic paint markers be used on canvas
+acrylic marker pens
+acrylic marker pen set
+acrylic marker pen price
+acrylic marker pen uses
+acrylic marker pen drawing
+acrylic marker pens uk
+acrylic marker pen black
+acrylic marker pen 24 shades
+acrylic marker pen art
+acrylic marker pen painting
+acrylic marker pens hobbycraft
+acrylic marker pen amazon
+acrylic marker pen mr diy
+acrylic marker pen for fabric
+acrylic marker pen how to use
+acrylic pen and marker holder
+acrylic marker pen himic
+acrylic marker pens home bargains
+acrylic marker pens hobby lobby
+acrylic paint pen hobby lobby
+acrylic paint pen holder
+acrylic paint pen hobbycraft
+acrylic paint pen home depot
+acrylic paint pen how to use
+acrylic paint pens home bargains
+acrylic paint pens hs code
+best acrylic paint pens hobbycraft
+acrylic marker pen gold
+acrylic marker pens guangna
+acrylic paint pen gold
+acrylic paint pen green
+acrylic paint pen grey
+acrylic paint pen glass
+acrylic paint pen grey set
+acrylic paint pen golf ball
+acrylic paint pen graffiti
+acrylic paint pens grabie
+acrylic paint pens glitter
+acrylic paint pens guangna
+acrylic paint pens gundam
+acrylic paint pens gunpla
+acrylic paint pen ideas
+acrylic paint pen ideas for beginners
+acrylic paint pen icon
+acrylic paint pens ireland
+acrylic paint pens in national bookstore
+acrylic paint pens in store
+acrylic paint pens india
+acrylic paint in pen
+acrylic paint pen art ideas
+acrylic paint pen drawing ideas
+easy acrylic paint pen ideas
+which acrylic paint pen is best
+acrylic paint pen craft ideas
+acrylic paint pens ak interactive
+acrylic marker pen flair
+acrylic marker pen faber castell
+acrylic marker pen fine tip
+acrylic marker pen for kids
+acrylic marker pen flipkart
+acrylic marker pens for glass
+acrylic paint pen fine tip
+acrylic paint pen for fabric
+acrylic paint pen for wood
+acrylic paint pen flowers
+acrylic paint pen for art and crafts
+acrylic paint pen for canvas
+acrylic paint pen for glass
+acrylic marker pen shopee
+acrylic paint pen set
+acrylic paint pen storage
+acrylic marker sketch pen
+acrylic paint pen set uk
+acrylic paint pen silver
+acrylic paint pen sealer
+acrylic paint pen spotlight
+acrylic paint pen set nearby
+acrylic paint sketch pen
+acrylic paint marker pen set
+acrylic marker 12 pen set flair brand
+acrylic marker pen posca
+acrylic marker pen popular
+acrylic marker pen peak
+acrylic paint pen projects
+acrylic paint pen posca
+acrylic paint pen price
+acrylic paint pen painting
+acrylic paint pen pink
+acrylic paint pens pna
+acrylic paint pens permanent
+acrylic paint pens pastel
+acrylic paint pens professional
+acrylic paint pen joann
+acrylic paint pens jumbo
+acrylic pen ideas
+what pens write on acrylic
+do acrylic paint pens work on plastic
+what are acrylic paint pens used for
+acrylic marker pen ohuhu
+acrylic paint pen on fabric
+acrylic paint pen on glass
+acrylic paint pen officeworks
+acrylic paint pen on canvas
+acrylic paint pen on wood
+acrylic paint pen on plastic
+acrylic paint pen on mirror
+acrylic paint pen on metal
+acrylic paint pen on leather
+acrylic paint pen organizer
+acrylic paint pen on skin
+acrylic paint pen on shirt
+acrylic paint pen on ceramic
+acrylic marker pen
+acrylic marker pen white
+acrylic marker pen near me
+acrylic paint pens video
+acrylic paint pens vs markers
+acrylic paint pen vs sharpie
+sharpie acrylic paint pens vs posca
+acrylic paint pens vs alcohol markers
+acrylic pen vs marker
+acrylic pen vs permanent marker
+acrylic marker vs brush pen
+acrylic marker vs color pen
+acrylic marker vs gel pen
+acrylic paint pens velles
+acrylic marker pen under ₹ 100
+acrylic marker pen under 200
+acrylic marker pen under ₹ 200
+acrylic marker pen under ₹ 300
+acrylic marker pen under 100
+acrylic marker pen under ₹ 400
+acrylic paint pen uses
+acrylic paint pen ultra fine
+acrylic paint pen uk
+acrylic paint pens ultra fine tip
+acrylic paint pens uk amazon
+best acrylic marker pens uk
+acrylic marker pens kmart
+acrylic paint pen kmart
+acrylic paint pen kits
+acrylic paint pens kids
+acrylic paint pens kuwait
+acrylic paint pens nz kmart
+acrylic paint dot pens kmart
+kokuyo camlin acrylic marker pen
+what are acrylic markers used for
+does permanent marker stay on acrylic
+acrylic paint pens not working
+acrylic pens how to use
+acrylic marker pen meesho
+acrylic marker pen malaysia
+acrylic marker pens michaels
+acrylic paint pen michaels
+acrylic paint pen mont marte
+acrylic paint pen metallic
+acrylic paint pen molotow
+acrylic paint pen mug
+acrylic paint pens miniatures
+acrylic paint pens mr price
+acrylic paint pens medium tip
+acrylic paint pens mitre 10
+acrylic paint pens michaels nearby
+acrylic marker pen quality
+acrylic marker pen quiz
+acrylic marker pen quick dry
+acrylic marker pen qvc
+acrylic marker pen quick release
+artecho acrylic marker pen
+artecho dual tip acrylic marker pen
+arrtx acrylic marker pen
+acrylic marker vs acrylic paint pen
+acrylic paint marker pen amazon
+difference between acrylic marker and brush pen
+akarued white paint pen acrylic marker
+acrylic marker brush pen amazon
+marker acrylic pen allegro
+acrylic marker and brush pen
+is an acrylic marker a paint pen
+best acrylic marker pen
+black acrylic marker pen
+brustro acrylic marker pen
+best white acrylic marker pen
+baoke acrylic marker pen
+acrylic marker brush pen
+brush pen vs acrylic marker
+acrylic brush marker pen set
+acrylic paint marker brush pen
+acrylic paint marker pen black
+acrylic marker pen box
+acrylic paint marker calligraphy brush pen
+best acrylic paint marker pen
+acrylic marker pens the works
+acrylic marker pens the range
+acrylic marker pens tesco
+acrylic paint pen tooli art
+acrylic paint pen tutorial
+acrylic paint pen target
+acrylic paint pen techniques
+acrylic paint pen tips
+acrylic paint pen thin
+acrylic paint pen thick
+acrylic paint pen the range
+acrylic paint pen tool art
+acrylic paint pen temu
+acrylic paint pens the works
+acrylic paint pen white
+acrylic paint pen walmart
+acrylic paint pen water based
+acrylic paint pen waterproof
+acrylic paint pen warhammer
+acrylic paint pen white michaels
+acrylic marker with pen
+acrylic paint pen washable
+acrylic paint pen wood
+acrylic paint pens warehouse
+acrylic paint pens with brush tip
+whsmith acrylic paint pens
+acrylic paint pens wholesale
+acrylic marker pen refill
+acrylic marker pen review
+acrylic paint pen refillable
+acrylic paint pen reviews
+acrylic paint pen removal
+acrylic paint pen red
+acrylic paint pen reddit
+acrylic paint pens range
+acrylic paint pens rocks
+acrylic paint pens rymans
+acrylic paint pens reject shop
+acrylic paint pens red dot
+acrylic paint pen for resin
+acrylic marker pen xl
+acrylic marker pen xray
+acrylic marker pen xtool
+guangna direct liquid soft head acrylic marker pen
+gold acrylic marker pen
+grasp acrylic marker pen
+guangna acrylic marker pen
+grabie acrylic marker pen
+languo acrylic marker gel pen
+acrylic marker brush pen guangna
+marker acrylic pen m&g
+goffi acrylic paint marker pen
+doloha acrylic.marker pen
+deli acrylic marker pen
+doms acrylic marker pen
+dual tip acrylic marker pen
+direct liquid soft head acrylic marker pen
+direct liquid acrylic marker pen
+dual tip acrylic paint pen marker
+double.sided acrylic pen marker
+double sided acrylic pen marker set of 24
+acrylic marker pen diy
+dual tip acrylic paint pen marker - 24/48/72 colours
+beyond draw-dual-tip-acrylic-paint-pen-marker
+fine tip acrylic marker pen
+flair acrylic marker pen
+marker pen for acrylic painting
+marker pen for acrylic board
+acrylic marker pen used for
+marker pen for acrylic
+white marker pen for acrylic
+flair acrylic paint marker pen
+ohuhu acrylic marker pen for diy
+camel acrylic marker pen
+carissa acrylic marker pen
+camlin acrylic marker pen
+acrylic colour marker pen
+acrylic marker pen china
+acrylic marker pen in chinese
+acrylic marker brush pen 80 cores
+sharpie creative marker acrylic paint pen
+acrylic marker brush pen 60 crore
+caneta acrylic marker brush pen
+miya acrylic marker pen
+metallic acrylic marker pen
+led acrylic writing message board night lamp with marker pen
+acrylic marker brush pen mercado livre
+b&m acrylic marker pens
+sketching pens & markers acrylic marker pen
+marcadores acrylic marker pen
+set acrylic marker pen
+soft head acrylic marker pen
+silver acrylic marker pen
+acrylic marker 12 pen set
+acrylic marker pen 48 shades
+acrylic marker pen 36 shades
+acrylic marker pen 12 shades
+pen peak acrylic marker pen
+price acrylic marker pen
+posca acrylic marker pen
+paint acrylic marker pen
+acrylic permanent marker pen
+wotek acrylic paint marker pen
+acrylic paint marker pen white
+acrylic marker vs paint pen
+thick acrylic marker pen
+the acrylic paint marker pen set
+acrylic tip marker pen
+is an acrylic marker the same as a paint pen
+how to use acrylic marker pen
+enmy acrylic marker pen
+emmy acrylic marker pen
+marker acrylic pen empik
+what are acrylic markers
+are there acrylic paint pens
+how to use acrylic markers
+nicety acrylic marker pen
+acrylic marker pen nearby
+acrylic pen marker national bookstore
+what markers can you use on acrylic
+ohuhu acrylic marker pen
+ohuhu acrylic marker pen price
+unicorn acrylic marker pen
+what is acrylic marker pen
+what are acrylic paint markers used for
+white acrylic marker pen
+@huangaa_3 guangna direct liquid soft head acrylic marker pen
+hightune acrylic marker brush pen
+liquid soft head acrylic marker pen
+how to use acrylic marker
+how do acrylic markers work
+languo acrylic marker pen
+liquid acrylic marker pen
+best acrylic marker pens
+best acrylic paint marker pens
+what are the best acrylic markers
+what is the best acrylic paint pens
+what are the best acrylic pens
+acrylic marker pen ideas
+acrylic marker pen price in bangladesh
+is sharpie acrylic
+acrylic paint pen zeyar
+acrylic paint pens new zealand
+acrylic marker brush pen zjw
+acrylic marker pen zuixua
+restly acrylic marker pen
+how long do acrylic paint pens last
+are acrylic paint pens permanent
+are sharpies acrylic
+acrylic paint pens lyuvie
+acrylic paint pens like posca
+acrylic paint pens liquitex
+acrylic paint pens life of colour
+acrylic paint pens lowes
+acrylic paint pens large
+acrylic paint pens large set
+acrylic paint pens languo
+acrylic paint pens for leather shoes
+acrylic marker pen best
+acrylic marker pen blinkit
+acrylic marker pens b&m
+acrylic paint pen black
+acrylic paint pen brush tip
+acrylic paint pen brands
+acrylic paint pen brush
+acrylic paint pen bunnings
+acrylic paint pen brown
+acrylic paint pen big w
+acrylic paint pen blue
+acrylic paint pen by numbers
+acrylic marker pen doms
+acrylic marker pen deli
+acrylic marker pen dual tip
+acrylic paint pen drawing
+acrylic paint pen drying time
+acrylic paint pen designs
+acrylic paint pen doodles
+acrylic paint pen dry
+dual tip acrylic paint.pen
+acrylic paint pens dried out
+acrylic paint pens desire deluxe
+acrylic marker pen camlin
+acrylic marker pen camel
+acrylic paint pen coloring book
+acrylic paint pen crafts
+acrylic paint pen case
+acrylic paint pen canvas
+acrylic marker colour pen
+acrylic paint pen car
+acrylic paint pen colouring book
+acrylic paint pen canada
+acrylic paint pens cheap
+acrylic paint pens crockd
+acrylic marker pen enmy
+acrylic marker pens ebay
+acrylic paint pen extra fine
+acrylic paint pen extra fine tip
+acrylic paint pen empty
+acrylic paint pens ebay
+acrylic paint pens earth tones
+acrylic paint pens eckersley
+acrylic paint pens enmy
+bia acrylic paint pen extra fine tip
+sharpie acrylic paint pens earth tones
+artistro acrylic paint pens extra fine tip
+acrylic paint pens double ended
+acrylic marker pens arrtx
+acrylic marker pens arrtx simptap
+acrylic paint pen art
+acrylic paint pen amazon
+acrylic paint pen artwork
+acrylic paint pens australia
+acrylic paint pens argos
+acrylic paint pens asda
+acrylic paint pens artistro
+acrylic paint pens arrtx
+acrylic paint pens at michaels
--- a/data/links.json
+++ b/data/links.json
@@ -0,0 +1,18 @@
+{
+  "keyword": "马克笔绘画",
+  "count": 12,
+  "links": [
+    "https://www.tiktok.com/@drawing_board8/video/7569235583214587150",
+    "https://www.tiktok.com/@seekingartsupplier_my/video/7259632291306032402",
+    "https://www.tiktok.com/@huangaa_3/video/7522044666745818382",
+    "https://www.tiktok.com/@acrylicmarkerasmr/video/7470837517571345695",
+    "https://www.tiktok.com/@fungraffiti/video/7550422681103895863",
+    "https://www.tiktok.com/@yzd20328cuq/video/7553666411760012574",
+    "https://www.tiktok.com/@acrylicmarkerasmr/video/7495580197811440926",
+    "https://www.tiktok.com/@miss.uk3/video/7561345033010416916",
+    "https://www.tiktok.com/@nashvibes/video/7472932371990416670",
+    "https://www.tiktok.com/@muse1378/video/7322534704643722539",
+    "https://www.tiktok.com/@miss.uk3/video/7567258402661960981",
+    "https://www.tiktok.com/@miss.uk3/video/7569627806678846740"
+  ]
+}
--- a/data/store.py
+++ b/data/store.py
@@ -0,0 +1,25 @@
+from utils.io import write_json
+
+def save_links_snapshot(path, keywords, items, links):
+    """写入链接快照
+
+    结构：`{'keywords': list, 'items': list, 'total_count': int, 'links': list}`
+    """
+    write_json(path, {'keywords': list(keywords), 'items': items, 'total_count': len(links), 'links': all_links(links)})
+    return path
+
+def save_comments_snapshot(path, items):
+    """写入评论快照：`{'items': items}`"""
+    write_json(path, {'items': items})
+    return path
+
+def all_links(links):
+    """将任意可迭代链接转为列表（用于 JSON 序列化）"""
+    return list(links)
+"""数据快照写入工具
+
+职责：
+- 将链接搜索的结果按统一结构写入 JSON
+- 将评论抓取的结果写入 JSON
+仅负责序列化，不包含业务逻辑。
+"""
--- a/data/tik_comments.json
+++ b/data/tik_comments.json
--- a/data/urls.json
+++ b/data/urls.json
--- a/db/pycache/mysql_import.cpython-312.pyc
+++ b/db/pycache/mysql_import.cpython-312.pyc
--- a/db/mysql_import.py
+++ b/db/mysql_import.py
@@ -0,0 +1,49 @@
+import csv
+import os
+
+def import_csv_to_mysql(csv_path, host='localhost', port=3306, user='root', password='', database='crawler_tiktok', table='comments'):
+    try:
+        import pymysql
+    except Exception:
+        print('missing dependency: pip install pymysql', flush=True)
+        raise SystemExit(1)
+    if not os.path.exists(csv_path):
+        print('csv not found: ' + csv_path, flush=True)
+        raise SystemExit(1)
+    conn = pymysql.connect(host=host, port=int(port), user=user, password=password, database=database, charset='utf8mb4')
+    cur = conn.cursor()
+    cur.execute(f"CREATE TABLE IF NOT EXISTS `{table}` (\n        `id` BIGINT AUTO_INCREMENT PRIMARY KEY,\n        `username` VARCHAR(255),\n        `text` TEXT\n      ) CHARACTER SET utf8mb4 COLLATE utf8mb4_unicode_ci")
+    rows = []
+    with open(csv_path, 'r', encoding='utf-8', newline='') as f:
+        r = csv.reader(f)
+        first = True
+        for row in r:
+            if first and row and row[0].lower() == 'username':
+                first = False
+                continue
+            first = False
+            if not row:
+                continue
+            username = row[0] if len(row) > 0 else ''
+            text = row[1] if len(row) > 1 else ''
+            rows.append((username, text))
+    if rows:
+        cur.executemany(f"INSERT INTO `{table}` (`username`,`text`) VALUES (%s,%s)", rows)
+        conn.commit()
+    cur.close()
+    conn.close()
+    print(f"inserted={len(rows)}", flush=True)
+
+def create_database_if_not_exists(host='localhost', port=3306, user='root', password='', database='yunque'):
+    try:
+        import pymysql
+    except Exception:
+        print('missing dependency: pip install pymysql', flush=True)
+        raise SystemExit(1)
+    conn = pymysql.connect(host=host, port=int(port), user=user, password=password, charset='utf8mb4')
+    cur = conn.cursor()
+    cur.execute(f"CREATE DATABASE IF NOT EXISTS `{database}` CHARACTER SET utf8mb4 COLLATE utf8mb4_unicode_ci")
+    conn.commit()
+    cur.close()
+    conn.close()
+    print(f"database_ready={database}", flush=True)
--- a/main.py
+++ b/main.py
@@ -0,0 +1,185 @@
+import argparse
+import os
+from utils.io import load_keywords_from_file, read_json
+from tiktok.search import save_links_multi
+from tiktok.comments import save_comments_from_links
+from db.mysql_import import import_csv_to_mysql, create_database_if_not_exists
+
+def run_links(args):
+    """运行链接收集阶段
+
+    参数来源：命令行（关键词、请求文件、分页、并发等）
+    流程：
+    1. 汇总关键词（--keyword/--keywords/--keywords-file）
+    2. 校验非空
+    3. 调用 `save_links_multi` 并发搜索与去重，保存到 `args.out`
+    """
+    kws = []
+    if args.keyword:
+        kws.extend([k for k in args.keyword if k])
+    if args.keywords:
+        for k in args.keywords.split(','):
+            k = k.strip()
+            if k:
+                kws.append(k)
+    if args.keywords_file:
+        kws.extend(load_keywords_from_file(args.keywords_file))
+    kws = [k for k in kws if k]
+    if not kws:
+        raise SystemExit('no keywords')
+    save_links_multi(kws, out_path=args.out, file_path=args.file_path, max_pages=args.max_pages, timeout=args.timeout, count=args.count, workers=args.workers)
+
+def run_comments(args):
+    """运行评论与回复抓取阶段
+
+    输入：`args.links_json`（可为统一快照或简单结构）
+    读取逻辑：优先 `links` 字段；若无，则聚合 `items[*].links`
+    调用：`save_comments_from_links` 执行并发抓取，输出 JSON 与可选 CSV
+    """
+    obj = read_json(args.links_json)
+    links = obj.get('links') or []
+    if not links and os.path.exists(args.links_json):
+        try:
+            for name in ['links', 'items']:
+                if name == 'items':
+                    tmp = []
+                    for it in obj.get('items', []):
+                        tmp.extend(it.get('links', []))
+                    links = tmp
+                    break
+        except Exception:
+            pass
+    if not links:
+        raise SystemExit('no links')
+    save_comments_from_links(links, out_path=args.out, file_path=args.file_path, count=args.count, pages=args.pages, timeout=args.timeout, reply_count=args.reply_count, reply_pages=args.reply_pages, total_limit=args.limit, reply_total_limit=args.reply_limit, csv_path=args.csv, workers=args.workers)
+
+def run_all(args):
+    """串联执行链接收集与评论抓取
+
+    1. 解析关键词并调用搜索阶段输出到 `args.links_out`
+    2. 读取链接快照，兼容两种结构
+    3. 调用评论抓取阶段输出到 `args.comments_out` 并可写入 CSV
+    适用于一体化流水线执行。
+    """
+    kws = []
+    if getattr(args, 'keyword', None):
+        kws.extend([k for k in args.keyword if k])
+    if getattr(args, 'keywords', None):
+        for k in args.keywords.split(','):
+            k = k.strip()
+            if k:
+                kws.append(k)
+    if getattr(args, 'keywords_file', None):
+        kws.extend(load_keywords_from_file(args.keywords_file))
+    kws = [k for k in kws if k]
+    if not kws:
+        raise SystemExit('no keywords')
+    save_links_multi(kws, out_path=args.links_out, file_path=args.file_path, max_pages=args.search_max_pages, timeout=args.search_timeout, count=args.search_count, workers=args.search_workers)
+    obj = read_json(args.links_out)
+    links = obj.get('links') or []
+    if not links and os.path.exists(args.links_out):
+        try:
+            for name in ['links', 'items']:
+                if name == 'items':
+                    tmp = []
+                    for it in obj.get('items', []):
+                        tmp.extend(it.get('links', []))
+                    links = tmp
+                    break
+        except Exception:
+            pass
+    if not links:
+        raise SystemExit('no links')
+    save_comments_from_links(links, out_path=args.comments_out, file_path=args.file_path, count=args.comments_count, pages=args.comments_pages, timeout=args.comments_timeout, reply_count=args.reply_count, reply_pages=args.reply_pages, total_limit=args.comments_limit, reply_total_limit=args.reply_limit, csv_path=args.csv, workers=args.comments_workers)
+
+def main():
+    """命令行解析并分发到对应子命令函数"""
+    p = argparse.ArgumentParser()
+    sub = p.add_subparsers(dest='cmd')
+    p_links = sub.add_parser('links')
+    p_links.add_argument('--keyword', action='append')
+    p_links.add_argument('--keywords', default=None)
+    p_links.add_argument('--keywords-file', default=None)
+    p_links.add_argument('--file-path', default=r'data\1.text')
+    p_links.add_argument('--out', default='data\\urls.json')
+    p_links.add_argument('--max-pages', type=int, default=50)
+    p_links.add_argument('--count', type=int, default=None)
+    p_links.add_argument('--timeout', type=int, default=30)
+    p_links.add_argument('--workers', type=int, default=5)
+    p_links.set_defaults(func=run_links)
+
+    p_comments = sub.add_parser('comments')
+    p_comments.add_argument('--links-json', default='data\\urls.json')
+    p_comments.add_argument('--out', default='data\\tik_comments.json')
+    p_comments.add_argument('--file-path', default=r'data\\1.text')
+    p_comments.add_argument('--count', type=int, default=100)
+    p_comments.add_argument('--pages', type=int, default=100)
+    p_comments.add_argument('--timeout', type=int, default=30)
+    p_comments.add_argument('--limit', type=int, default=None)
+    p_comments.add_argument('--reply-count', type=int, default=100)
+    p_comments.add_argument('--reply-pages', type=int, default=100)
+    p_comments.add_argument('--reply-limit', type=int, default=None)
+    p_comments.add_argument('--csv', default='data\\comments.csv')
+    p_comments.add_argument('--workers', type=int, default=None)
+    p_comments.set_defaults(func=run_comments)
+
+    p_all = sub.add_parser('all')
+    p_all.add_argument('--keyword', action='append')
+    p_all.add_argument('--keywords', default=None)
+    p_all.add_argument('--keywords-file', default=None)
+    p_all.add_argument('--file-path', default=r'data\\1.text')
+    p_all.add_argument('--links-out', default='data\\urls.json')
+    p_all.add_argument('--search-max-pages', type=int, default=50)
+    p_all.add_argument('--search-count', type=int, default=None)
+    p_all.add_argument('--search-timeout', type=int, default=30)
+    p_all.add_argument('--search-workers', type=int, default=5)
+    p_all.add_argument('--comments-out', default='data\\tik_comments.json')
+    p_all.add_argument('--comments-count', type=int, default=100)
+    p_all.add_argument('--comments-pages', type=int, default=100)
+    p_all.add_argument('--comments-timeout', type=int, default=30)
+    p_all.add_argument('--comments-limit', type=int, default=None)
+    p_all.add_argument('--reply-count', type=int, default=100)
+    p_all.add_argument('--reply-pages', type=int, default=100)
+    p_all.add_argument('--reply-limit', type=int, default=None)
+    p_all.add_argument('--csv', default='data\\comments.csv')
+    p_all.add_argument('--comments-workers', type=int, default=None)
+    p_all.set_defaults(func=run_all)
+
+    p_mysql = sub.add_parser('mysql')
+    p_mysql.add_argument('--csv', default='data\\comments.csv')
+    p_mysql.add_argument('--host', default='localhost')
+    p_mysql.add_argument('--port', type=int, default=3306)
+    p_mysql.add_argument('--user', default='root')
+    p_mysql.add_argument('--password', default='')
+    p_mysql.add_argument('--database', default='crawler_tiktok')
+    p_mysql.add_argument('--table', default='comments')
+    def run_mysql(args):
+        import_csv_to_mysql(args.csv, host=args.host, port=args.port, user=args.user, password=args.password, database=args.database, table=args.table)
+    p_mysql.set_defaults(func=run_mysql)
+
+    p_mysql_db = sub.add_parser('mysql-db')
+    p_mysql_db.add_argument('--host', default='localhost')
+    p_mysql_db.add_argument('--port', type=int, default=3306)
+    p_mysql_db.add_argument('--user', default='root')
+    p_mysql_db.add_argument('--password', default='')
+    p_mysql_db.add_argument('--database', default='yunque')
+    def run_mysql_db(args):
+        create_database_if_not_exists(host=args.host, port=args.port, user=args.user, password=args.password, database=args.database)
+    p_mysql_db.set_defaults(func=run_mysql_db)
+
+    args = p.parse_args()
+    if not args.cmd:
+        p.print_help()
+        raise SystemExit(1)
+    args.func(args)
+
+if __name__ == '__main__':
+    main()
+"""命令行入口模块
+
+提供三类子命令：
+- links：根据关键词并发搜索视频链接并保存快照
+- comments：根据链接列表抓取评论与回复并保存快照与 CSV
+- all：串联 links 与 comments，一次性完成全流程
+运行方式建议使用 `python -m crawler_tiktok.main ...` 以避免导入路径问题。
+"""
--- a/1
+++ b/1
@@ -0,0 +1 @@
+MySQL80
--- a/tiktok/pycache/comments.cpython-312.pyc
+++ b/tiktok/pycache/comments.cpython-312.pyc
--- a/tiktok/pycache/search.cpython-312.pyc
+++ b/tiktok/pycache/search.cpython-312.pyc
--- a/tiktok/comments.py
+++ b/tiktok/comments.py
@@ -0,0 +1,208 @@
+import json
+import re
+import threading
+import time
+from urllib.parse import urlparse, parse_qs, urlencode
+from urllib.request import Request, urlopen
+from core.curl import parse_curl_file
+from utils.io import ensure_csv_header, append_csv_rows
+from data.store import save_comments_snapshot
+
+def _extract_aweme_id(link):
+    """从视频链接中提取 aweme_id（/video/<id>）"""
+    m = re.search(r"/video/(\d+)", link)
+    return m.group(1) if m else None
+
+def fetch_comments_aweme(aweme_id, file_path, count=20, max_pages=50, timeout=30, total_limit=None, referer=None):
+    """分页抓取某个视频的评论
+
+    参数：
+    - `aweme_id` 视频 ID
+    - `file_path` curl 文本文件（第 1 块为评论接口基准）
+    - `count/max_pages/timeout` 分页与超时控制
+    - `total_limit` 总条数上限（可选）
+    - `referer` 用于设置请求头的来源页（可选）
+    行为：失败重试、必要时切换到兜底评论接口；处理 `has_more/next_cursor`。
+    返回：评论对象列表。
+    """
+    reqs = parse_curl_file(file_path)
+    if not reqs:
+        return []
+    base = reqs[0]
+    headers = dict(base['headers'])
+    if referer:
+        headers['referer'] = referer
+    cursor = 0
+    all_comments = []
+    for _ in range(max_pages):
+        u_parsed = urlparse(base['url'])
+        q = parse_qs(u_parsed.query)
+        q['aweme_id'] = [str(aweme_id)]
+        q['count'] = [str(count)]
+        q['cursor'] = [str(cursor)]
+        u = u_parsed._replace(query=urlencode(q, doseq=True)).geturl()
+        data = None
+        for i in range(3):
+            try:
+                req = Request(u, headers=headers, method='GET')
+                with urlopen(req, timeout=timeout) as resp:
+                    data = resp.read()
+                break
+            except Exception:
+                time.sleep(0.5 * (i + 1))
+                data = None
+        try:
+            obj = json.loads(data.decode('utf-8', errors='ignore'))
+        except Exception:
+            obj = {}
+        if not obj.get('comments'):
+            alt_params = {'aid': 1988, 'aweme_id': aweme_id, 'count': count, 'cursor': cursor}
+            alt_url = 'https://www.tiktok.com/api/comment/list/?' + urlencode(alt_params)
+            for i in range(2):
+                try:
+                    req = Request(alt_url, headers=headers, method='GET')
+                    with urlopen(req, timeout=timeout) as resp:
+                        data2 = resp.read()
+                    obj2 = json.loads(data2.decode('utf-8', errors='ignore'))
+                    if obj2.get('comments'):
+                        obj = obj2
+                        break
+                except Exception:
+                    time.sleep(0.5 * (i + 1))
+        comments = obj.get('comments') or []
+        for c in comments:
+            all_comments.append(c)
+        if isinstance(total_limit, int) and total_limit > 0 and len(all_comments) >= total_limit:
+            break
+        has_more = obj.get('has_more')
+        next_cursor = obj.get('cursor') or obj.get('next_cursor')
+        if has_more in (True, 1) and isinstance(next_cursor, int):
+            cursor = next_cursor
+            continue
+        if comments and isinstance(next_cursor, int):
+            cursor = next_cursor
+            continue
+        break
+    return all_comments
+
+def fetch_replies(comment_id, aweme_id, file_path, count=20, max_pages=50, timeout=30, total_limit=None):
+    """分页抓取某条评论的二级回复
+
+    参数：`comment_id/aweme_id` 标识；其他参数同评论抓取。
+    返回：回复对象列表。
+    """
+    reqs = parse_curl_file(file_path)
+    if not reqs:
+        return []
+    headers = reqs[0]['headers']
+    base = 'https://www.tiktok.com/api/comment/list/reply/'
+    cursor = 0
+    replies = []
+    for _ in range(max_pages):
+        params = {'aid': 1988, 'aweme_id': aweme_id, 'comment_id': comment_id, 'count': count, 'cursor': cursor}
+        url = base + '?' + urlencode(params)
+        data = None
+        for i in range(3):
+            try:
+                req = Request(url, headers=headers, method='GET')
+                with urlopen(req, timeout=timeout) as resp:
+                    data = resp.read()
+                break
+            except Exception:
+                time.sleep(0.5 * (i + 1))
+                data = None
+        try:
+            obj = json.loads(data.decode('utf-8', errors='ignore'))
+        except Exception:
+            obj = {}
+        arr = obj.get('comments') or []
+        for r in arr:
+            replies.append(r)
+        if isinstance(total_limit, int) and total_limit > 0 and len(replies) >= total_limit:
+            break
+        has_more = obj.get('has_more')
+        next_cursor = obj.get('cursor')
+        if has_more in (True, 1) and isinstance(next_cursor, int):
+            cursor = next_cursor
+            continue
+        break
+    return replies
+
+_csv_lock = threading.Lock()
+_print_lock = threading.Lock()
+_results_lock = threading.Lock()
+
+def save_comments_from_links(links, out_path, file_path, count=20, pages=50, timeout=30, reply_count=20, reply_pages=50, total_limit=None, reply_total_limit=None, csv_path=None, workers=None):
+    """并发从视频链接抓取评论与回复并保存快照
+
+    并发：可选信号量限制；每个链接独立线程抓取；
+    CSV：若提供 `csv_path`，按 `username,text` 追加主评论与回复；
+    输出：写入 `out_path`，结构为 `{'items': [{link,count,comments: [...]}, ...]}`。
+    """
+    ensure_csv_header(csv_path, ['username', 'text'])
+    results = []
+    sem = None
+    if isinstance(workers, int) and workers > 0:
+        sem = threading.Semaphore(workers)
+
+    def _process(link):
+        if sem:
+            sem.acquire()
+        with _print_lock:
+            print(f"[START] {link}", flush=True)
+        try:
+            cs = fetch_comments_aweme(_extract_aweme_id(link), file_path=file_path, count=count, max_pages=pages, timeout=timeout, total_limit=total_limit, referer=link)
+            enriched = []
+            for c in cs:
+                cid = c.get('cid')
+                if cid:
+                    rs = fetch_replies(cid, _extract_aweme_id(link), file_path=file_path, count=reply_count, max_pages=reply_pages, timeout=timeout, total_limit=reply_total_limit)
+                    c = dict(c)
+                    c['replies'] = rs
+                    c['reply_count'] = len(rs)
+                enriched.append(c)
+                try:
+                    with _print_lock:
+                        print(f"{link} | cid={c.get('cid')} | create_time={c.get('create_time')} | reply_count={c.get('reply_count', 0)} | text={c.get('text')}", flush=True)
+                except Exception:
+                    pass
+                if csv_path:
+                    u = c.get('user') or {}
+                    uname = u.get('unique_id') or u.get('nickname') or u.get('uid') or ''
+                    rows = [[uname, c.get('text')]]
+                    for r in c.get('replies', []) or []:
+                        ru = r.get('user') or {}
+                        runame = ru.get('unique_id') or ru.get('nickname') or ru.get('uid') or ''
+                        rows.append([runame, r.get('text')])
+                    with _csv_lock:
+                        append_csv_rows(csv_path, rows)
+            with _results_lock:
+                results.append({'link': link, 'count': len(cs), 'comments': enriched})
+            reply_total = sum(len(c.get('replies') or []) for c in enriched)
+            with _print_lock:
+                print(f"[DONE] {link} comments={len(cs)} replies={reply_total}", flush=True)
+        except Exception as e:
+            with _print_lock:
+                print(f"[ERROR] {link} {e}", flush=True)
+        finally:
+            if sem:
+                sem.release()
+
+    threads = []
+    for link in links:
+        t = threading.Thread(target=_process, args=(link,))
+        t.daemon = True
+        t.start()
+        threads.append(t)
+    for t in threads:
+        t.join()
+    save_comments_snapshot(out_path, results)
+    return out_path
+"""TikTok 评论与回复抓取模块
+
+能力：
+- 根据视频链接提取 aweme_id
+- 通过评论接口分页拉取评论（支持兜底接口）
+- 针对每条评论抓取二级回复并汇总
+- 可选写入 CSV 与打印进度日志
+"""
--- a/tiktok/search.py
+++ b/tiktok/search.py
@@ -0,0 +1,171 @@
+import json
+import re
+import threading
+import time
+from urllib.parse import urlparse, parse_qs, urlencode, urlunparse
+from urllib.request import Request, urlopen
+from core.curl import parse_curl_file
+from data.store import save_links_snapshot
+
+def _update_query(url, updates):
+    """在原始 URL 上用 `updates` 更新查询参数并返回新 URL"""
+    p = urlparse(url)
+    q = parse_qs(p.query)
+    for k, v in updates.items():
+        q[k] = [str(v)]
+    new_q = urlencode(q, doseq=True)
+    return urlunparse((p.scheme, p.netloc, p.path, p.params, new_q, p.fragment))
+
+def _extract_links(obj):
+    """从返回对象中提取视频链接
+
+    优先从 `data -> item -> author.uniqueId + item.id` 组合；
+    同时遍历字符串字段，用正则匹配 tiktok 链接作为兜底。
+    返回：链接列表（可能包含重复，外层负责去重）。
+    """
+    links = []
+    data = obj.get('data') if isinstance(obj, dict) else None
+    if isinstance(data, list):
+        for e in data:
+            if isinstance(e, dict) and e.get('type') == 1 and isinstance(e.get('item'), dict):
+                it = e['item']
+                author = it.get('author') or {}
+                uid = author.get('uniqueId')
+                vid = it.get('id')
+                if uid and vid:
+                    links.append(f"https://www.tiktok.com/@{uid}/video/{vid}")
+    patterns = [
+        r"https?://www\.tiktok\.com/[\w@._-]+/video/\d+",
+        r"https?://www\.tiktok\.com/video/\d+",
+        r"https?://vm\.tiktok\.com/[\w-]+",
+        r"https?://vt\.tiktok\.com/[\w-]+",
+    ]
+    def rec(x):
+        if isinstance(x, dict):
+            for v in x.values():
+                rec(v)
+        elif isinstance(x, list):
+            for v in x:
+                rec(v)
+        elif isinstance(x, str):
+            s = x
+            for pat in patterns:
+                for m in re.finditer(pat, s):
+                    links.append(m.group(0))
+    rec(obj)
+    return links
+
+def search_video_links(keyword, file_path, max_pages=50, timeout=30, count=None, on_link=None):
+    """按关键词分页搜索视频链接
+
+    输入：从 `file_path` 的第 2 个 curl 请求获取基准 URL 与头部
+    行为：分页拉取、重试、解析链接；对新链接触发 `on_link` 回调
+    返回：所有发现的链接列表（不去重本地返回，外层统一去重）。
+    """
+    reqs = parse_curl_file(file_path)
+    if len(reqs) < 2:
+        return []
+    base = reqs[1]
+    headers = base['headers']
+    parsed = urlparse(base['url'])
+    q = parse_qs(parsed.query)
+    if count is None:
+        if 'count' in q:
+            try:
+                count = int(q['count'][0])
+            except Exception:
+                count = 12
+        else:
+            count = 12
+    all_links = []
+    seen = set()
+    offset = 0
+    cursor = None
+    for _ in range(max_pages):
+        params = {'keyword': keyword, 'count': count}
+        if cursor is not None:
+            params['offset'] = cursor
+        else:
+            params['offset'] = offset
+        u = _update_query(base['url'], params)
+        data = None
+        for i in range(3):
+            try:
+                req = Request(u, headers=headers, method='GET')
+                with urlopen(req, timeout=timeout) as resp:
+                    data = resp.read()
+                break
+            except Exception:
+                time.sleep(0.5 * (i + 1))
+                data = None
+        try:
+            obj = json.loads(data.decode('utf-8', errors='ignore'))
+        except Exception:
+            obj = {}
+        links = _extract_links(obj)
+        has_more = obj.get('has_more')
+        next_cursor = obj.get('cursor')
+        new = 0
+        for l in links:
+            if l not in seen:
+                seen.add(l)
+                all_links.append(l)
+                new += 1
+                if on_link:
+                    try:
+                        on_link(l)
+                    except Exception:
+                        pass
+        if has_more in (True, 1) and isinstance(next_cursor, int):
+            cursor = next_cursor
+            continue
+        if new == 0:
+            break
+        offset += count
+    return all_links
+
+_print_lock = threading.Lock()
+
+def save_links_multi(keywords, out_path, file_path, max_pages=50, timeout=30, count=None, workers=5):
+    """并发按多个关键词搜索并保存快照
+
+    并发：使用线程 + 信号量限制并发；跨关键词统一去重；
+    输出：写入 `out_path`，包含 `keywords/items/total_count/links`。
+    """
+    all_links = []
+    seen = set()
+    items = []
+    seen_lock = threading.Lock()
+    sem = threading.Semaphore(max(1, int(workers)))
+
+    def worker(kw):
+        with sem:
+            item_links = []
+            def on_new(l):
+                with seen_lock:
+                    if l not in seen:
+                        seen.add(l)
+                        all_links.append(l)
+                        item_links.append(l)
+                        with _print_lock:
+                            print(l, flush=True)
+            search_video_links(kw, file_path=file_path, max_pages=max_pages, timeout=timeout, count=count, on_link=on_new)
+            items.append({'keyword': kw, 'count': len(item_links), 'links': item_links})
+
+    threads = []
+    for kw in keywords:
+        t = threading.Thread(target=worker, args=(kw,))
+        t.daemon = True
+        t.start()
+        threads.append(t)
+    for t in threads:
+        t.join()
+    save_links_snapshot(out_path, keywords, items, all_links)
+    return out_path
+"""TikTok 视频链接搜索模块
+
+核心能力：
+- 构造查询 URL（更新 keyword/offset/count 等参数）
+- 发起请求并解析返回中的视频链接（结构化 + 正则兜底）
+- 对多个关键词并发搜索、统一去重与快照保存
+"""
--- a/utils/pycache/io.cpython-312.pyc
+++ b/utils/pycache/io.cpython-312.pyc
--- a/utils/filter_comments.py
+++ b/utils/filter_comments.py
@@ -0,0 +1,55 @@
+import argparse
+import csv
+import os
+
+def filter_comments(csv_in, csv_out, keywords):
+    ks = set(k.lower() for k in keywords if k)
+    rows_out = []
+    with open(csv_in, 'r', encoding='utf-8', newline='') as f:
+        r = csv.reader(f)
+        first = True
+        for row in r:
+            if first and row and row[0].lower() == 'username':
+                first = False
+                continue
+            first = False
+            if not row:
+                continue
+            text = row[1] if len(row) > 1 else ''
+            s = (text or '').lower()
+            if any(k in s for k in ks):
+                rows_out.append(row)
+    os.makedirs(os.path.dirname(csv_out), exist_ok=True)
+    with open(csv_out, 'w', encoding='utf-8', newline='') as wf:
+        w = csv.writer(wf)
+        w.writerow(['username', 'text'])
+        for r in rows_out:
+            w.writerow(r)
+    print(f"input={csv_in} keywords={len(ks)} matched_rows={len(rows_out)} out={csv_out}")
+
+def main():
+    p = argparse.ArgumentParser()
+    p.add_argument('--extern-keywords', default=r'd:\work\test\test\all_keywords.txt')
+    p.add_argument('--local-keywords', default=r'data\keyword.txt')
+    p.add_argument('--csv-in', default=r'data\comments.csv')
+    p.add_argument('--csv-out', default=r'data\key_comment.csv')
+    args = p.parse_args()
+    def _load(path):
+        arr = []
+        try:
+            with open(path, 'r', encoding='utf-8') as f:
+                for line in f:
+                    s = line.strip()
+                    if s:
+                        arr.append(s)
+        except Exception:
+            arr = []
+        return arr
+    kws = []
+    kws.extend(_load(args.extern_keywords))
+    kws.extend(_load(args.local_keywords))
+    kws.append('pen')
+    filter_comments(args.csv_in, args.csv_out, kws)
+
+if __name__ == '__main__':
+    main()
--- a/utils/io.py
+++ b/utils/io.py
@@ -0,0 +1,54 @@
+import json
+import os
+import csv
+
+def load_keywords_from_file(path):
+    """逐行读取关键词文件，忽略空行，返回列表"""
+    arr = []
+    try:
+        with open(path, 'r', encoding='utf-8') as f:
+            for line in f:
+                s = line.strip()
+                if s:
+                    arr.append(s)
+    except Exception:
+        arr = []
+    return arr
+
+def write_json(path, obj):
+    """以 UTF-8 写入 JSON，使用非 ASCII 保留与缩进"""
+    with open(path, 'w', encoding='utf-8') as f:
+        json.dump(obj, f, ensure_ascii=False, indent=2)
+
+def read_json(path):
+    """读取 JSON 文件，失败时返回空对象"""
+    try:
+        with open(path, 'r', encoding='utf-8') as f:
+            return json.load(f)
+    except Exception:
+        return {}
+
+def ensure_csv_header(path, headers):
+    """若 CSV 不存在则创建并写入表头；为空路径直接返回"""
+    if not path:
+        return
+    if not os.path.exists(path):
+        with open(path, 'w', newline='', encoding='utf-8') as wf:
+            w = csv.writer(wf)
+            w.writerow(headers)
+
+def append_csv_rows(path, rows):
+    """向 CSV 追加多行，行元素按列表给出；为空路径直接返回"""
+    if not path:
+        return
+    with open(path, 'a', newline='', encoding='utf-8') as af:
+        w = csv.writer(af)
+        for r in rows:
+            w.writerow(r)
+"""通用 IO 工具
+
+提供：
+- 关键词文件加载
+- JSON 读写
+- CSV 文件写入（确保表头、追加行）
+"""