Inference of publication type

ReCiter infers a single publication type for each publication record. This inference is stored in the publicationTypeCanonical field and is inferred based on the publication type values in PubMed. PubMed may have up to 6-7 publication types assigned to a given article.

Here's the logic tree:

1. If the PubMed publication type includes "Editorial", the article is classified as an "Editorial Article".
2. Otherwise, if it contains "Letter", it's classified simply as a "Letter".
3. Otherwise, if it contains "Comment", it's classified as a "Comment".
4. Otherwise, if it contains "Preprint", it's classified as a "Preprint".
5. Otherwise, if it contains "Published Erratum", it's classified as an "Erratum".
6. Otherwise, for various types related to conferences (like "Consensus Development Conference" or "Congress"), it's classified as a "Conference Paper".
7. Otherwise, for "Guideline" or "Practice Guideline", it's classified as a "Guideline".
8. Otherwise, for publications that are "Meta-Analysis", "Review", or similar, it's classified as a "Review".
9. Otherwise, if it's a "Case Reports", it's classified as a "Case Report".
10. Otherwise, for a wide range of types that imply a more formal research article, including "Journal Article", various phases of "Clinical Trial", and others, it's classified as an "Academic Article".
11. Otherwise, if none of the specific categories match, it defaults to classifying the publication as an "Article".