Add "«", "»" to blacklist characters list

[SiarheiFedartsou]

👋

We faced an issue that our Pelias instance couldn't find ТРЦ «Respublika Park» (https://www.openstreetmap.org/way/293888317) via queries which don't contain "«", "»" symbols(e.g. I expect Respublika Park to be almost perfect match for this place). After some experimentations I narrowed it down to the fact that Pelias's ElasticSearch schema doesn't consider these symbols as blacklist ones, i.e. it becomes part of token (now we get «Respublika and Park» as separate tokens). So I propose to add it to list of blacklisted characters to fix the issue. Such symbols are quite popular in some regions, see this wiki article https://en.wikipedia.org/wiki/Guillemet

---

Here's the reason for this change 🚀

"«", "»" symbols are quite popular delimiters in some languages.

---

Here's what actually got changed 👏

Added "«", "»" symbols to blacklist characters list

---

Here's how others can test the changes 👀

I updated tests + it can be tested manually via bootstrapping ES instance.

[missinglink]

Seems reasonable, we actually have those symbols listed as quote characters in https://github.com/pelias/parser/blob/7d28b15f3cebee2d5739266f82cb3f0daa0f17f4/tokenization/split_funcs.js#L10

[missinglink]

Is this good to merge or would you like to add additional symbols?

[missinglink]

I took the liberty to clean up your commits, add some additional symbols and clean up some code, hope you don»t mind 🙏

Thanks!

[SiarheiFedartsou]

@missinglink I definitely don't, thanks!

[missinglink]

I screwed up the commit message so the docker images only just got published now after I fixed it 🤦‍♂️.

• https://github.com/pelias/schema/releases/tag/v7.1.0
• https://hub.docker.com/layers/pelias/schema/v7.1.0/images/sha256-72db96494aa1b56d29a2db02b1984db0b2bb47d855e390644bfbc1aadf4493d1