internetarchive / internetarchive/openlibrary
When importing authors from Amazon, also import the Amazon ID when available
- Dominant language
- Python
- Stars
- 6.7k
- Forks
- 2k
- Avg merge
- 2d 19h
- Merged PRs (30d)
- 138
Description
At the moment, the importbot must be very careful when importing new authors, because otherwise different authors with similar names might get mixed up. That has the side effect that the most carefully curated author's on OpenLibrary hardly get any new automatic imports - they will usually have some minor difference from the Amazon entry.
As far as I understand it, most books which are automatically imported are imported from Amazon. Amazon offers an author ID that can be set in OpenLibrary too.
Would it be possible to import the Amazon ID alongside the author name and check it when deciding whether to create a new author or use an existing one?
### Describe the problem that you'd like solved
I imagine it something like that:
Bot looks for an ISBN from Amazon. Amazon reports "I got this ISBN".
ImportBot requests Amazon author ID and check whether it already exists on OpenLibrary.
If it exists, it request the author's name. This name is checked against the author's name and pseudonyms on OpenLibrary.
If both ID and name / pseudonym match, the book should be included with the author.
Otherwise, a new author would be created.
If the Amazon ID does not yet exist, Importbot only runs the checks it uses now, but also imports the Amazon ID for future use.
### Proposal & Constraints
I am not sure how reliable Amazon author IDs are. It might lead to an increase in faulty imports.
As far as I understood @seabelis, a wrongly added author is much better than a book wrongly added to an existing author, as merging is easier than picking apart. So, the new implementation should not increase the risk of false attributions.
### Additional context
### Stakeholders
@mekarpeles
Contributor guide
Research direction
Start by tracing the importbot path that handles Amazon ISBN imports and author matching. Review how existing authors and pseudonyms are checked, then determine how Amazon author IDs could be retrieved and stored without increasing false attributions. Done means the matching and new-author behavior are covered for both existing and previously unseen Amazon IDs.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- backend, data
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 30/100