manubot version 0.2.2 release notes
===================================

Manubot version 0.2.2 contains citation and web request enhancements.

New features
------------

This release adds citation support for two additional types of identifiers (`isbn` and `wikidata`).

ISBNs are the primary persistent identifier for many books, so many books no longer need to be cited by URL (see [GH79](https://github.com/manubot/manubot/pull/79) and [GH14](https://github.com/manubot/manubot/issues/14)).
However, ISBN metadata is sometimes missing or erroneous.
Users may need to still need to set manual CSL JSON, but Manubot can at least produce a reasonable starting template.
Try for example `manubot cite isbn:9780062316097`.

[Wikidata](https://www.wikidata.org) is a free and open knowledge base that contains many records of scholarly works.
Wikidata can store metadata on records that do not have their own persistent identifiers, and thus can help Manubot users assign a stable identifier to works that otherwise would not have one (see [GH67](https://github.com/manubot/manubot/issues/67) and [GH86](https://github.com/manubot/manubot/pull/86)).
Try for example `manubot cite wikidata:Q50051684`.

Manubot now uses Zotero's [translation-server](https://github.com/zotero/translation-server) infrastructure to provide metadata for wikidata, ISBN, and URL citations (see [GH70](https://github.com/manubot/manubot/issues/70) and [GH84](https://github.com/manubot/manubot/pull/84)).
Manubot now hosts its own instance of translation-server at https://translate.manubot.org (see [GH82](https://github.com/manubot/manubot/issues/82)).
As such, Manubot users can benefit from Zotero's impressive collection of [translators](https://github.com/zotero/translators) for retrieving metadata from different webpages.
Manubot's ISBN and URL citation metadata retrievers now first attempt to generate metadata using translation-server, and fallback to other methods if that fails.

Bug fixes
---------

NCBI E-Utility requests are now rate limited to 2 per second (see [GH83](https://github.com/manubot/manubot/pull/83)).
Previously, certain situations that caused rapid E-Utility requests would return status code 429 for "too many requests".

Internal changes
----------------

Outgoing web requests made by Manubot now set the User-Agent header (see [GH83](https://github.com/manubot/manubot/pull/83)).
These headers provide high-level information of a user's system, as shown in the following examples:

```
manubot/0.2.2 (Linux; Python/3.6) <contact@manubot.org>
manubot/0.2.2 (Windows; Python/3.7) <contact@manubot.org>
```

Setting the header will help upstream resources contact the Manubot developers should our requests be problematic or should downtime be anticipated.
Furthermore, it will allow Manubot's translation-server to monitor Manubot usage, including which operating system, Python version, and package version are being used.

Manubot's test suite has been reorganized such that testing modules correspond one-to-one with package modules (see [GH87](https://github.com/manubot/manubot/pull/87)).