Unrelated information parsed into the last entry in Authors field

[handecelikkanat]

Some unrelated information (journal, year) sometimes is parsed as part of the last entry in Authors field.

Example:

"authors":["WT Tsai, L Zhang - 2025 IEEEI International Conference on Artificial …, 2025"]

There are multiple instances, I found more and list as part of my review of PR 11.
Some older entries also have it, so I dont think PR 11 introduced it.

[sebastian-nagel]

Thanks, @handecelikkanat! A few explanations why the authors fields may include other information.

The Scholar alerts parser only parses the structure of the alerts: an email (EML), the embedded HTML which holds the list of citations. Every citation has four blocks:

• title
• author(s)
• snippet
• sharing links

The author(s) block starts with the author list, but may also include the publication year, the publication source, or rarely some additional information about the authors ("bronze sponsors"). The publication source can be a journal, a conference, etc. Because the source might be truncated, it often is not really useful, e.g., "Proceedings of the Fifth …"

The publication year is definitely redundant, the publication source, although noisy, might provide some value.

Sure, we can try to split the other information from the authors field and write it to a separate field(s).

But we should keep in mind: the Scholar alerts provide semi-structured, noisy data. There might be better ways to get clean citations. It's maybe not worth to invest a lot of work in data cleansing.

[handecelikkanat]

@sebastian-nagel This makes sense, thank you for reminding we might not continue with GScholar - Semantic Scholar was discussed to take this information from. Let me close this issue then!