Jump to content

Wikidata:Property proposal/Baidu Baike article

From Wikidata

Baidu Baike identifier

[edit]

Return to Wikidata:Property proposal/Authority control

   Under discussion
Descriptionidentifier for an article (lemma) in Baidu Baike
Data typeExternal identifier
Example 1Liu Huan (Q461415): Chinese singer-songwriter (1963–2026) → https://baike.baidu.com/item/刘欢/6703
Example 2Wikipedia (Q52): free multilingual online encyclopedia → https://baike.baidu.com/item/维基百科/106382
Example 3Pinglu Canal (Q119804497): canal in Guangxi, China → https://baike.baidu.com/item/平陆运河/5069924
Example 4Japanese role-playing video game (Q5923834): video game genre → https://baike.baidu.com/item/日式角色扮演游戏/58641007

Motivation

[edit]

Baidu Baike (Q803722) (百度百科) is a Chinese-language online encyclopedia operated by Baidu with 31,694,089 entries so far. (This is 4 times more than English Wikipedia.) Its entries are called lemmas (词条).

Baidu identifies its lemmas using a numeric lemma ID (lemmaId). Baidu's current official APIs expose the lemmaId and lemmaTitle as separate fields, and use the lemma ID to distinguish different lemmas that may have the same title (disambiguation page (Q138717185)).

A normal Baidu Baike article URL:

https://baike.baidu.com/item/{lemmaTitle}/{lemmaId}

For Liu Huan (Q461415): Chinese singer-songwriter (1963–2026), for example:

https://baike.baidu.com/item/刘欢/6703

Thus, there are two possible designs for this property which should be considered:

Option 1: store only the lemma ID

[edit]

Example value:

6703

This corresponds directly to Baidu's lemmaId / lemma_id field and treats the numeric ID as the external identifier.

The lemma ID can also be used without the title in some Baidu Baike URLs. For example, the history interface accepts:

https://baike.baidu.com/p/history?lemmaId=6703

The disadvantage is that the normal human-facing article URL cannot be constructed from the numeric ID alone. The canonical article URL requires both the lemma title and lemma ID.

Option 2: store lemmaTitle/lemmaId

[edit]

Example value:

刘欢/6703

Under this design, the property would store the two components required by the normal Baidu Baike article URL. A formatter URL could then simply be:

https://baike.baidu.com/item/$1

This has the practical advantage that every value can link directly to the corresponding Baidu Baike main article without an API request or other lookup.

The disadvantage is that this is a composite value rather than the numeric lemmaId itself. Baidu's own current API represents lemma_title and lemma_id as separate fields rather than representing lemmaTitle/lemmaId as one identifier. Also, if there is only one page with a given lemmaTitle, the lemmaId can be omitted; if multiple, there seems to be a default one, which is the most popular entity.

The English Wikipedia already has Template:Baidu Baike (Template:Baidu Baike (Q104804206): Wikimedia template), which is used to link to Baidu Baike articles. Its documentation supports a title together with numeric identifiers.

内存溢出的猫 (talk) 17:47, 26 September 2026 (UTC)reply

Discussion

[edit]
  • Previous proposals: 2015, 2018, 2021, 2021, 2023.--GZWDer (talk) 20:18, 26 September 2026 (UTC)reply
    Ah, thank you. I searched before creating it, but I missed these previous discussions. —— 内存溢出的猫 (talk) 01:34, 28 September 2026 (UTC)reply
  • Regarding the copyvio issue discussed above, this is a common problem with UGC websites, and it seems that the objections raised in the discussion mainly came from the Chinese Wikipedia community. I think most reasons have been "addressed" by User:Habst. Copyright infringement requires a complaint from the rights holder, which reflects Baidu Baike’s lack of proactive measures to remove copyright-violating (copyvio) content. However, such measures are not legally required. In addition, Baidu Baike has been cited extensively in Chinese academic papers. A search for “baike.baidu.com” on CNKI (Q12857515): China's academic research database returns 185,928 hits. ——内存溢出的猫 (talk) 01:58, 28 September 2026 (UTC)reply
  • Support with the lemmaTitle/lemmaId format for linkable URLs. Should the title change, a bot should be able to update the property value based on the lemmaId. The only way I would support a bare-lemmaId property is if someone was willing to make a Toolforge redirector site that transforms links like https://baike.toolforge.org/lemmaId to their full Baidu Baike article URL. In that case, we could just link to the Toolforge redirector from Wikidata.
If this property is approved, I would be open to making that site. --Habst (talk) 13:14, 28 September 2026 (UTC)reply
  •  Support, with no current preference between the options as I have not looked into them. As I stated at a prior proposal, adding a property ID for an entity does [not] carry any sort of implied endorsement, and like it or not, Baidu Baike is probably the best source of encyclopedic information for non-controversial Chinese topics that might be too obscure to have much or any coverage on Wikipedia. Sdkb talk 04:20, 5 October 2026 (UTC)reply
Retrieved from "https://www.wikidata.org/w/index.php?title=Wikidata:Property_proposal/Baidu_Baike_article&oldid=2553293049"