Bahamut Gaming Forum & VOC Scraper

English | 繁體中文

English

Collect structured, public discussion data from the Bahamut Gaming Forum (forum.gamer.com.tw) for game research, community monitoring, and auditable VOC analysis.

This Actor is purpose-built for Bahamut's bsn, snA, floor, comment, GP/BP, and board-search model. Generic phpBB or Discourse scrapers do not understand those relationships; broad web/forum search Actors usually return snippets instead of complete Bahamut threads.

Choose this Actor when

• You need to discover a Bahamut game board from a game/board name or start from known bsn values.
• You need keyword search inside selected boards with latest or GP-based ordering.
• You need the original post, every available floor, public comments, GP/BP, and available listing view counts in one discussion record.
• You need repeatable post-release monitoring that emits new or updated discussions and preserves a crawl audit.

Do not choose it for logged-in, age-restricted, deleted, private, or permission-gated content. The Actor records those URLs as inaccessible and does not bypass platform controls.

What it collects

Each successful discussion is one Dataset item containing:

• board bsn, discussion snA, board name, title, canonical URL, tags, and timestamps;
• original post and all available floors with content text/HTML, author identity handling, GP, BP, and timestamps;
• all public comments per floor with comment IDs, content, timestamps, GP, and BP;
• available view count from board listing discovery;
• public image URLs and external links referenced in content;
• pagination and comment completeness fields;
• source keywords/listings, monitoring status, and optional game version label.

The Actor never downloads images, videos, animations, or linked files. It excludes links hosted by Bahamut Anime (ani.gamer.com.tw) and Bahamut Store (buy.gamer.com.tw) and never visits external links.

Inputs

Provide at least one source:

• boardQueries: game/board names, such as 原神, for board discovery;
• bsns: known board IDs supplied as numeric strings, such as "36730";
• startUrls: direct public C.php discussion URLs.

The Console prefill uses one known public discussion and one-item limits so Apify's automated quality test can validate the Actor quickly. These prefill values do not change API, schedule, or Saved Task defaults; those runs must still provide at least one source.

Useful controls:

• keywords: searches every selected board. Empty keywords use board listings.
• sortBy: LATEST or HOT. Keyword hot ordering uses Bahamut GP sorting. Without keywords, hot ordering ranks the page-bounded candidates by GP.
• publishedAfter / publishedBefore: filter using the first floor's Asia/Taipei publication time.
• maxListPages, maxThreads, maxThreadPages: hard collection limits. maxThreadPages: 0 means all available thread pages.
• includeComments, includeImages, includeExternalLinks: output depth controls.
• authorMode: HASHED (default), RAW, or OMIT.
• authorHashSalt: optional secret salt for hashes stable across unrelated runs.
• proxyConfiguration: defaults to the verified Taiwan residential proxy route. The tested direct and generic data-center routes returned Bahamut HTTP 403 challenge pages; changing the proxy route can affect access and cost.

Example: keyword VOC collection

{
  "operation": "SCRAPE_DISCUSSIONS",
  "boardQueries": ["原神"],
  "keywords": ["更新", "卡頓", "閃退"],
  "sortBy": "LATEST",
  "publishedAfter": "2026-08-01",
  "maxListPages": 3,
  "maxThreads": 50,
  "maxThreadPages": 0,
  "includeComments": true,
  "authorMode": "HASHED",
  "proxyConfiguration": {
    "useApifyProxy": true,
    "apifyProxyGroups": ["RESIDENTIAL"],
    "apifyProxyCountry": "TW"
  }
}

Example: known board, popular discussions

{
  "bsns": ["36730"],
  "keywords": ["聖遺物"],
  "sortBy": "HOT",
  "maxListPages": 2,
  "maxThreads": 20,
  "includeComments": true
}

Incremental version monitoring

Run the same input repeatedly with the same monitor key. The Actor stores only fingerprints and monitoring metadata in a named Key-Value Store; successful discussion results remain in the current run's Dataset.

{
  "bsns": ["36730"],
  "keywords": ["7.0", "更新", "BUG"],
  "publishedAfter": "2026-08-12",
  "sortBy": "LATEST",
  "monitoring": {
    "enabled": true,
    "key": "genshin-7-0",
    "onlyNew": true,
    "includeUpdated": true,
    "versionLabel": "7.0"
  }
}

• First run: records are NEW.
• Later runs: unseen discussions are NEW; changed floors/comments/metrics are UPDATED; unchanged discussions can be skipped.
• Set reset: true only when intentionally starting a new baseline for that monitor key.

Audit and completeness

Audit records are stored in the default Key-Value Store so they are not mixed with business results:

• RUN_AUDIT: sanitized input summary, counts, deduplication, monitoring, and media policy;
• PAGE_AUDITS: requested/final URLs, response status, attempts, bytes, and timing;
• FAILED_URLS: failed, deleted, sensitive, or login-required URLs and reasons;
• INCOMPLETE_DISCUSSIONS: successfully parsed but capped/incomplete discussions. These are deliberately not pushed to the default Dataset.

completeness.is_complete requires all detected thread pages, expected floors when known, and requested comments to be present. View count can be null for direct URLs and keyword results because Bahamut exposes exact views on board listings, not consistently on every discussion/search response.

Responsible use

Collect only public data and follow applicable laws, Bahamut terms, and research ethics. Use conservative limits and delays. Hashed author IDs reduce direct exposure but are pseudonymous, not anonymous; use a private salt and restrict access when identity linkage is unnecessary.

繁體中文

從巴哈姆特電玩資訊站哈啦區（forum.gamer.com.tw）擷取結構化的公開討論資料，用於遊戲研究、社群監測及可稽核的 VOC（使用者聲音）分析。

此 Actor 專為巴哈姆特的 bsn、snA、樓層、留言、GP/BP 及看板搜尋架構而設計。通用 phpBB 或 Discourse 爬蟲無法理解這些關聯；一般網站或論壇搜尋 Actor 通常只會回傳摘要，而不是完整的巴哈姆特討論串。

適合使用此 Actor 的情境

• 您需要透過遊戲或看板名稱尋找巴哈姆特看板，或從已知的 bsn 開始擷取。
• 您需要在指定看板內搜尋關鍵字，並依最新或 GP 熱度排序。
• 您需要在同一筆討論資料中取得主樓、所有可取得的樓層、公開留言、GP/BP，以及看板列表中可取得的瀏覽數。
• 您需要可重複執行的版本發布後監測，只輸出新增或更新的討論，並保留爬取稽核紀錄。

此 Actor 不適用於需要登入、限制年齡、已刪除、非公開或受權限限制的內容。Actor 會將這些網址記錄為無法存取，且不會繞過平台的存取控制。

蒐集內容

每個成功擷取的討論會成為一筆 Dataset 項目，包含：

• 看板 bsn、討論 snA、看板名稱、標題、標準網址、標籤及時間戳記；
• 主樓及所有可取得的樓層，包括內容文字/HTML、作者身分處理結果、GP、BP 及時間戳記；
• 每個樓層下的所有公開留言，包括留言 ID、內容、時間戳記、GP 及 BP；
• 從看板列表找到的可取得瀏覽數；
• 內容中引用的公開圖片網址及外部連結；
• 分頁與留言完整性欄位；
• 來源關鍵字/列表、監測狀態及可選的遊戲版本標籤。

此 Actor 絕不下載圖片、影片、動畫或連結檔案。它會排除巴哈姆特動畫瘋（ani.gamer.com.tw）與巴哈商城（buy.gamer.com.tw）上的連結，也不會造訪外部連結。

輸入

請至少提供一種來源：

• boardQueries：用於尋找看板的遊戲或看板名稱，例如 原神；
• bsns：以數字字串提供的已知看板 ID，例如 "36730"；
• startUrls：公開 C.php 討論頁面的直接網址。

Console 的預填值使用一個已知的公開討論及單筆限制，讓 Apify 自動品質測試能快速驗證 Actor。這些預填值不會變更 API、排程或 Saved Task 的預設值；這些執行仍須自行提供至少一種來源。

常用控制項：

• keywords：在每個指定看板中搜尋。若留空，則使用看板列表。
• sortBy：LATEST 或 HOT。關鍵字的熱門排序使用巴哈姆特 GP 排序；未提供關鍵字時，熱門排序會依 GP 排列受頁數限制的候選討論。
• publishedAfter / publishedBefore：依主樓的 Asia/Taipei 發布時間篩選。
• maxListPages、maxThreads、maxThreadPages：明確的蒐集上限。maxThreadPages: 0 代表擷取所有可取得的討論分頁。
• includeComments、includeImages、includeExternalLinks：控制輸出內容的深度。
• authorMode：HASHED（預設）、RAW 或 OMIT。
• authorHashSalt：可選的私密雜湊鹽值，讓不相關的多次執行仍可產生穩定雜湊。
• proxyConfiguration：預設使用已驗證的台灣住宅代理路由。測試中的直接連線與一般資料中心路由會收到巴哈姆特 HTTP 403 驗證頁；變更代理路由可能影響存取能力與費用。

範例：關鍵字 VOC 蒐集

{
  "operation": "SCRAPE_DISCUSSIONS",
  "boardQueries": ["原神"],
  "keywords": ["更新", "卡頓", "閃退"],
  "sortBy": "LATEST",
  "publishedAfter": "2026-08-01",
  "maxListPages": 3,
  "maxThreads": 50,
  "maxThreadPages": 0,
  "includeComments": true,
  "authorMode": "HASHED",
  "proxyConfiguration": {
    "useApifyProxy": true,
    "apifyProxyGroups": ["RESIDENTIAL"],
    "apifyProxyCountry": "TW"
  }
}

範例：已知看板的熱門討論

{
  "bsns": ["36730"],
  "keywords": ["聖遺物"],
  "sortBy": "HOT",
  "maxListPages": 2,
  "maxThreads": 20,
  "includeComments": true
}

遊戲版本增量監測

使用相同的監測鍵重複執行同一份輸入。Actor 只會將指紋與監測中繼資料儲存在具名 Key-Value Store 中；成功擷取的討論結果仍會保存在當次執行的 Dataset 中。

{
  "bsns": ["36730"],
  "keywords": ["7.0", "更新", "BUG"],
  "publishedAfter": "2026-08-12",
  "sortBy": "LATEST",
  "monitoring": {
    "enabled": true,
    "key": "genshin-7-0",
    "onlyNew": true,
    "includeUpdated": true,
    "versionLabel": "7.0"
  }
}

• 第一次執行：紀錄的狀態為 NEW。
• 後續執行：未曾出現的討論為 NEW；樓層、留言或指標有變更的討論為 UPDATED；未變更的討論可以略過。
• 只有在確定要為該監測鍵建立全新基準時，才設定 reset: true。

稽核與完整性

稽核紀錄儲存在預設 Key-Value Store 中，因此不會與業務結果混在一起：

• RUN_AUDIT：經過清理的輸入摘要、數量、去重、監測及媒體政策；
• PAGE_AUDITS：請求網址/最終網址、回應狀態、嘗試次數、位元組數及耗時；
• FAILED_URLS：擷取失敗、已刪除、敏感或需要登入的網址及原因；
• INCOMPLETE_DISCUSSIONS：已成功解析但因上限而不完整的討論。這些紀錄不會被推送到預設 Dataset。

只有在所有偵測到的討論分頁、可確認的預期樓層及使用者要求的留言都已取得時，completeness.is_complete 才會成立。直接網址與關鍵字搜尋結果的瀏覽數可能為 null，因為巴哈姆特會在看板列表顯示精確瀏覽數，但不一定會在每個討論頁或搜尋結果中提供。

負責任地使用

只蒐集公開資料，並遵守適用法律、巴哈姆特條款及研究倫理。請採用保守的蒐集上限與請求延遲。雜湊後的作者 ID 可降低直接暴露風險，但屬於假名化而非匿名化；不需要連結作者身分時，請使用私密鹽值並限制資料存取權限。