organicmaps/name-suggestion-index
13
Fork 0
Code Issues Pull requests Projects Releases Wiki Activity Actions
20955 commits 1 branch 0 tags 536 MiB
Commit graph

113 commits

Author SHA1 Message Date
Bryan Housel
800ca866f5 Update dependencies, npm run build 2023-03-01 17:59:30 -05:00
Bryan Housel
8450326e78 Ignore markdown/readme files 
(closes #7292)
 2022-10-31 09:13:20 -04:00
Minh Nguyễn
eec89184cb
Add local Goodwill affiliates as operators (#5734) 
* Added local Goodwill affiliates as operators

* Remove Goodwill from brands tree

Also added some more Goodwill organizations

Put operator in front of name for operators tree in idgen.js

Co-authored-by: archpdx <archpdx@pm.me>
 2022-10-30 19:44:10 -07:00
Bryan Housel
85015fe18b Use import assertions syntax for importing JSON files 
(closes #6014)
 2022-02-16 11:40:10 -05:00
Bryan Housel
d6d8b5acef s/colors/chalk 
(re: #6039)
 2022-01-10 23:40:27 -05:00
Max Erickson
7e08ce1b6c Revert "Add json type assertions for node 17." 
This reverts commit f4a11709e3.

Undo last change.
 2022-01-05 18:14:19 -05:00
Max Erickson
f4a11709e3 Add json type assertions for node 17. 2022-01-05 17:20:27 -05:00
Bryan Housel
e23bfba28b Add the ability to stuff extra metadata into the file headers 2021-11-03 13:18:46 -04:00
Bryan Housel
983308090f Raise a warning if a same-named item exists in several trees 
(re: #5625)
 2021-11-02 16:36:34 -04:00
Bryan Housel
e2bfcf3913 Make the Matcher a class 
This makes using it a bit more idiomatic: `const matcher = new Matcher();`
 2021-06-22 16:56:21 -04:00
Bryan Housel
b21ac2eb66 Enable --experimental-json-modules so we can import json files 
https://nodejs.org/docs/latest-v14.x/api/esm.html#esm_experimental_json_modules
This was something that we did a lot of with require()
 2021-06-22 15:17:54 -04:00
Bryan Housel
1c4fc70e3b WIP Modernize 
- switch to type: module
- replace all CJS require/module.exports with ES6 import/expor
 2021-06-22 00:04:52 -04:00
Bryan Housel
f760ac5721 Matcher should not throw on unrecognized locationSets 
(closes #5122)
 2021-06-14 15:52:52 -04:00
Bryan Housel
43e50b725d Drop entire general and supplemental unicode punctuation blocks 
U+2000-U206F is the General Punctuation block (en/em dashes are here)
U+2E00-U2E7F is the Supplemental Punctuation block
(closes #5117)
 2021-06-07 19:43:19 -04:00
Bryan Housel
7654d3663d Fix "Bim" -> "BİM" 
(closes #5017)
 2021-04-11 13:15:53 -04:00
Bryan Housel
21f5b4e838 In trees schema, name tag patterns do not need to be Arrays 2021-03-15 11:36:34 -04:00
Bryan Housel
4777daf743 Don't index excluded words in the matchIndex, even if they appear as values 
(i.e. the "nextbike" issue)
https://github.com/osmlab/name-suggestion-index/issues/4543#issuecomment-791762633
 2021-03-05 17:57:31 -05:00
Bryan Housel
cccc32bc16 Have the matcher handle generic/common words too 
(closes #4924)
 2021-02-26 11:18:23 -05:00
Bryan Housel
8c2d61eaaf More es6 in the matcher 2021-02-25 15:35:07 -05:00
Bryan Housel
723e98d4ca Flip the matchIndex around to be kv -> bucket -> names 
This lets us store properties per kv in preparation for #4924
Also upgrade code to use es6 Maps instead of objects
 2021-02-25 14:03:31 -05:00
Bryan Housel
366b13b4ff preserveTags should have format regex in the json schema 2021-02-24 10:02:03 -05:00
Bryan Housel
deb87996c8 Swap out mergeStrategy idea with simpler preserveTags array 
(closes #4906)
 2021-02-24 09:48:52 -05:00
Bryan Housel
db2475de01 Experimenting with introducing a tag merge strategy 
(re: #4906)
 2021-02-23 11:12:49 -05:00
Bryan Housel
58d638f4fe Restore the filtering in build script, 
Add skipCollection for categories which should not do it
(re: #4906)
 2021-02-18 15:20:53 -05:00
Bryan Housel
ea9d441ffb Wrap category properties in a properties object 2021-02-18 12:30:42 -05:00
Bryan Housel
b939ae2af4 Remove temp code, use ES6 Map for the itemID lookup 2021-02-17 15:10:27 -05:00
Bryan Housel
953b5531e0 Relax schamas, start refactor of cache and tree structures 
This moves towards storing more information about categories in the data/* files
(re: #4906)
 2021-02-17 09:48:25 -05:00
Bryan Housel
58bfd675d1 Use locale-aware sorting everywhere, npm run build 
Before, sort depended on the locale of the computer running the script
(re: #4850, #4871)
 2021-02-05 10:20:21 -05:00
Bryan Housel
eeaf81aec1 Rename sort -> sort_object 
(preparing to introduce more locale-specific sorting)
 2021-02-05 09:43:50 -05:00
Bryan Housel
f94b455aac Make sure each k/v pair appears in only one tree, npm run build 2021-01-28 12:03:48 -05:00
Bryan Housel
93ae4a408c Add templateInclude/templateExclude for UPS mailboxes 
(closes #4866)
 2021-01-19 17:06:08 -05:00
Bryan Housel
2c24a7d2d2 When expanding templates, don't copy matchTags (but do copy matchNames) 2021-01-19 15:32:10 -05:00
Bryan Housel
2b2a60d47f Exclude name:pronunciation 
I don't think we'd include these in NSI anyway, but it is a thing..
 2021-01-19 12:12:21 -05:00
Bryan Housel
c05dd74a25 Matcher refactor, many improvements 
- Separate caches for primary/alternate names
- This lets us perform a single indexing pass and return more relevant name matches
- Add primary/alternate name tag patterns to trees.json
- match() returns a result object that sorts primary before alternate
- More speed improvements
- More tests
 2021-01-15 12:01:12 -05:00
Bryan Housel
a93b7534c4 Add several categories where we don't want to match generic building/yes 2021-01-12 17:08:49 -05:00
Bryan Housel
023401cea0 Force localeCompare to use en-US locale for sorting consistency 2021-01-04 13:25:30 -05:00
Bryan Housel
504a8a2a30 Force newline '\n' on the end of all the files 
There are good reasons for files to end in a newline, and text editors keep trying to add it.
However `stringify` doesn't add it by default so this has resulted in `npm run
build` introducing a lot of churn in the files.
Hoping that this change avoids future churn.
 2021-01-04 10:56:19 -05:00
Bryan Housel
81fe88c1a6 Resolve locationSets when reading the file tree 
Because location-conflation has gotten a lot faster in v0.7.0, we can now resolve
all the locationSets into GeoJSON as we read them in reasonable time.

This lets us catch situations where a locationSet resolves to an empty geometry
 e.g. (`include: ['us'], exclude: ['us'])
to prevent mistakes like that from being added to the index in the first place.
 2021-01-04 10:24:06 -05:00
Bryan Housel
8a40582d47 Add a test for flag matching, fix bug in buildLocationIndex 2021-01-01 09:35:07 -05:00
Bryan Housel
1e8d2ddb49 Adjust matcher result sort order 
- by area descending, if no location specified
- by area ascending, if location specified
 2020-12-31 12:09:08 -05:00
Bryan Housel
be84f4ac9e Can't use require for path relative to project root 2020-12-30 17:24:36 -05:00
Bryan Housel
175ef9964d Create a convenience function for writing files with metadata 2020-12-30 17:01:40 -05:00
Bryan Housel
a2b53cbf9c Always build the location index, check for invalid resolved features 2020-12-30 00:56:24 -05:00
Bryan Housel
e27e2ae232 Add tweaks to the matcher code, add more tests 
(re: #4788)
 2020-12-28 16:42:48 -05:00
Bryan Housel
503adedf05 matcher: Use the locationSet id (+[Q30]), not the feature id (Q30) 
(this was causing a mixup when doing location-based matching)
 2020-12-24 11:04:27 -05:00
Bryan Housel
b13f500cc8 Ensure that simplify/stemmer can deal with non-string input 2020-12-23 11:17:05 -05:00
Bryan Housel
32b51d3964 Raise error if a file doesn't have the right file extension 2020-12-15 11:55:11 -05:00
Bryan Housel
01d4d3fa05 Rename index.json -> nsi.json, match_groups.json -> matchGroups.json 
Trying to be less confusing, more consistent
 2020-12-08 12:23:45 -05:00
Bryan Housel
392a8fbf14 Watch out for non-string locationset values 
(closes #4776)
 2020-12-05 16:26:43 -05:00
Bryan Housel
e20dc348fb Do a better job of displaying the duplicate 2020-12-02 13:01:15 -05:00