bellingcat / alias-generator Goto Github PK

Take a optional argument for locale and use this to influence the types of variants returned, such as culturally specific name-order conventions, and alternative transliterations.

getAliases("Hayao Miyazaki", "jp") => [ ... , "Miyazaki Hayao",  ...]
getAliases("Екатерина Александровна Гордеева", "ru") => [... "Ekaterina Alexandrovna Gordeeva",  ...]

Probably as a separate feature/function:

getUsernames("John Doe") => [ j.doe, jdoe, doejohn ]

There's just a basic smoke test at time of writing, but we should have a proper test suite with lots of different examples.

Tailor name proposals according to the submissions in singular form:

• Add spaces to names only when more than one name is present
• Add commas to last-name, first-name only where applicable
• Consider omitting the single-letter candidates where a single name is given

I realize this hits an edge-case that goes against best practice (fully qualify the name where possible). However, there are likely valid use-cases where only one name is known.

Currently, alias generator does not predict common spellings. For example:
Josef / Joseph
Philipp / Phillip / Philip

Especially in multilingual contexts, it is common for one person to use multiple such variants.

I don't know much about this topic, but Bohemian people historically used both German and Czech variants of their names. These are currently not predicted by the name variant search tool.

One example is:
Dora Rudolfine Richter, also known as Dora Rudolfa Richterová

van der Laan -> v.d. Laan

From social media feedback: "I have three first names, the third one is my primary day to day one (not uncommon for swedes). But combinations generated only care about my first first name. "

For example, a search for "Qasim Suleymany" should return:

• Qasem Soleimani
• Qassem Soleimani
• Qasim Sulemani
• etc.

Propose expanded-form candidate names where short versions are submitted:

Both mike and mikey should receive (optional) proposal michael in the generated candidate list.

alias generator fails to predict common German nicknames, such as:
Anton → Toni
Josef → Sepp
Johannes → Hans