Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
    rwrx
    @rwrx
    @albertors90 I still did not solved this problem with elasticsearch index being removed. Do you solved it?
    Julian Simioni
    @orangejulius
    @rwrx @albertors90 indices disappearing suddenly is very concerning. I can't think of anything that would cause it, but Elasticsearch can definitely do weird things. The Elasticsearch logs are confusing and verbose, but there might be something in there that gives hints as to what happened
    Rowan Winsemius
    @rowanwins
    Hey Pelias Team, I was wondering if there was any interest in kicking off the work on getting Pelias to support better units
    pelias/pelias#861 ?
    There appears to be a bunch of PRs that are reasonably close but have never been merged for various reasons. If there is anything I can contribute let me know :)
    akshay-skills19
    @akshay-skills19
    hello I have one doubt, our company already configured OSM with overpass API, but now I want to do reverse geocoding on that OSM database so can I use peliase with existing OSM DB?
    Brad Hards
    @bradh
    @akshay-skills19 You need the pelias ES approach. You won't (without a significant amount of code) be able use overpass.
    albertors90
    @albertors90
    @rwrx Still stuck with it, and can't find any useful information on ES logs
    Ondřej Machulda
    @OndraM
    Hi there! We have set up pelias.json imports to use data from OSM, OA, WOF for our country (Czech republic). We run prepare & import and start the containers, and everything seems OK. But when we run the same search request for an eg. street name using Pelias Compare Tool against official api.geocode.earth vs. our service, we are getting slightly different results (our service return less and also worse results - on the same request). It seems it returns less polylines points than official API. Any idea what could be causing this?
    robertoamoreno
    @robertoamoreno
    Anybody have an idea why when using the reverse api its only returning 1 result instead of multiple
    5 replies
    robertoamoreno
    @robertoamoreno
    overlap
    Enrico Hofmann
    @gigo1980
    hi together, does pelias also give me the "district" name inside a city ?
    Enrico Hofmann
    @gigo1980
    And how can I get the ZIP code
    Sergiusz Kierat
    @sergiuszkierat
    Hi. I've created PR pelias/docker#232 for Poland and it waits for your review
    Keyur
    @iamkeyur
    Or is there any way I can stop deferring to pelias parser?
    Keyur
    @iamkeyur
    And I also wonder why "QwERRR, Ahmedabad" would fail?
    Keyur
    @iamkeyur
    Same happens with 'Unjejkdkoojjaj, Jakarta'. So, anything not valid before city name fails. This has to fallback on city name match.
    bullbates
    @bullbates
    Hello everyone! Really struggling getting this working on a windows machine; Has anyone managed this that can run through it with me? I've read and tried the comments here but haven't succeeded yet - pelias/docker#124
    Benoît Bouré
    @bboure

    Hi all,
    What would be the minimum recommended setup for a full planet build for low/moderate traffic?
    I am currently running:

    • 1 m5.xlarge for services on Docker (API, placeholder, pip. I am not using libpostal)
    • 2 t3.large for ES

    I guess I could reduce the services instance to a m5.large. Memory might be a bit tight but it should work
    Also, what would you recommend regarding ES. Better have 2 medium instances OR 1 bigger instance (if we sacrifice redundancy) and what type? I was thinking to change for t3 for m5 or r5. What is best?

    Right now with 2 t3.medium ,performance is not always good (sometimes it takes several seconds for a response), and I have zero traffic (still testing).

    Thanks!

    Benoît Bouré
    @bboure
    The goal is obviously to reduce cost. We are on a tight budget.
    Julian Simioni
    @orangejulius
    hi @bboure, in our experience you want as few ES instances as possible. one that's twice as big is better than two
    and the main bottleneck is CPU. so you want to prefer the c5, m5, and then r5 instances in that order
    Benoît Bouré
    @bboure
    Thanks @orangejulius !! Will do that.
    What would you recommend for the services instance?
    Julian Simioni
    @orangejulius
    that really doesn't matter as much. you can go cheap there with a t3 or r5
    Benoît Bouré
    @bboure
    ok, thanks. I am using PIP, so I guess minimum is 8GB of RAM
    Julian Simioni
    @orangejulius
    yeah. i believe the t3s and r5s are almost identical in price per-gigabyte-of-ram, and for a production environment dealing with the CPU burst balance "feature" of the t3s is not worth the marginally cheaper cost
    Benoît Bouré
    @bboure
    Agree. r5 starts at 16GB though
    Julian Simioni
    @orangejulius
    if you're running a full planet PIP service 8GB ram total would be pretty tight
    the PIP service itself is almost 8, and the API and Placeholder each take a few hundred MB. Doesn't leave much room
    Benoît Bouré
    @bboure
    ok thanks for the tips.
    Will try with r5.large + 1 c5.large to start with and see from there
    cyfugr
    @cyfugr
    is it possible searching with pelias general types of venues, like coffee shop in london?
    Benoît Bouré
    @bboure
    @cyfugr yes, you can use the (undocumented) /nearby endpoint. It has a "categories" parameter
    Benoît Bouré
    @bboure
    /v1/nearby?categories=food:coffee_shop&boundary.circle.radius=10&point.lat=51.4896314&point.lon=-0.1248264
    cyfugr
    @cyfugr
    Cheers @bboure , i guess you can not just type it in search, like i will need fist to get the points od london and then perform that search.
    Benoît Bouré
    @bboure
    You can use search/autocomplete, but the use-case is a bit different. With search/autocomplete you would search for a specific coffee-shop: eg: Starbucks
    cyfugr
    @cyfugr
    @bboure got it, thank you.
    mdoakes42
    @mdoakes42
    Hello, New to Pelias but have docker setup and have some custom data loaded. THings working well. Question I have and hoping someone can help. Does the standard pelias schema contain geo location "shapes" (Polygons/lines)... Or is there just a centroid lat/lon. If so, can we do a spatial query on top of the polygon. IE, Only return records where polygon falls inside of bbox. (verse just checking the centroid...which may or may not be in the bbox)
    Edward Mac Gillavry
    @emacgillavry
    Would like to contribute updates and improvements to the dictionaries for the Netherlands as used by the parser (https://github.com/pelias/parser/tree/master/resources/libpostal/dictionaries/nl). How would these changes be made available to libpostal too?
    16 replies
    cyfugr
    @cyfugr
    Is there a way to get the type of a point? ie 'hotel', 'bar' etc from osm, wof data ?
    Francisco Sauceda
    @saucedaf
    Hello, has this been implemented? https://pelias.io/milestones/categories/
    gdi-disy
    @DisyGdi_twitter

    Hello everyone,
    i am struggling with the configuration for authentication via token for elasticsearch.
    curl calls from pelias-api pod to elasticsearch-master work with api-key auth but using pelias-api calls it fails with the following error

    Does anyone have an idea?

    authenticate={ 0=\"Bearer realm=\\\"security\\\"\" & 1=\"ApiKey\" & 2=\"Basic realm=\\\"security\\\" charset=\\\"UTF-8\\\"\" } } } :: {\"path\":\"/pelias/_search\",\"query\":{\"search_type\":\"dfs_query_then_fetch\"},\"body\":\"{\\\"query\\\":{\\\"function_score\\\":{\\\"query\\\":{\\\"bool\\\":{\\\"minimum_should_match\\\":1,\\\"should\\\":[{\\\"bool\\\":{\\\"_name\\\":\\\"fallback.street\\\",\\\"must\\\":[{\\\"match_phrase\\\":{\\\"address_parts.street\\\":{\\\"query\\\":\\\"westring\\\",\\\"slop\\\":4,\\\"analyzer\\\":\\\"peliasQuery\\\"}}}],\\\"filter\\\":{\\\"term\\\":{\\\"layer\\\":\\\"street\\\"}},\\\"boost\\\":5}},{\\\"bool\\\":{\\\"_name\\\":\\\"fallback.address\\\",\\\"must\\\":[{\\\"match_phrase\\\":{\\\"address_parts.number\\\":{\\\"query\\\":\\\"501\\\"}}},{\\\"match_phrase\\\":{\\\"address_parts.street\\\":{\\\"query\\\":\\\"westring\\\",\\\"slop\\\":4,\\\"analyzer\\\":\\\"peliasQuery\\\"}}}],\\\"filter\\\":{\\\"term\\\":{\\\"layer\\\":\\\"address\\\"}},\\\"boost\\\":10}}],\\\"filter\\\":{\\\"bool\\\":{\\\"must\\\":[]}}}},\\\"functions\\\":[]}},\\\"size\\\":20,\\\"track_scores\\\":true}\",\"statusCode\":401,\"response\":\"{\\\"error\\\":{\\\"root_cause\\\":[{\\\"type\\\":\\\"security_exception\\\",\\\"reason\\\":\\\"missing authentication credentials for REST request [/pelias/_search?search_type=dfs_query_then_fetch]\\\",\\\"header\\\":{\\\"WWW-Authenticate\\\":[\\\"Bearer realm=\\\\\\\"security\\\\\\\"\\\",\\\"ApiKey\\\",\\\"Basic realm=\\\\\\\"security\\\\\\\" charset=\\\\\\\"UTF-8\\\\\\\"\\\"]}}],\\\"type\\\":\\\"security_exception\\\",\\\"reason\\\":\\\"missing authentication credentials for REST request [/pelias/_search?search_type=dfs_query_then_fetch]\\\",\\\"header\\\":{\\\"WWW-Authenticate\\\":[\\\"Bearer realm=\\\\\\\"security\\\\\\\"\\\",\\\"ApiKey\\\",\\\"Basic realm=
    \\\\\\\"security\\\\\\\" charset=\\\\\\\"UTF-8\\\\\\\"\\\"]}},\\\"status\\\":401}\",\"wwwAuthenticateDirective\":\"Bearer realm=\\\"security\\\", ApiKey, Basic realm=\\\"security\\\" charset=\\\"UTF-8\\\"\"}","label":"api","timestamp":"2021-01-28T16:26:48.806Z"}

    pelias.json

    {
        "esclient": {
          "hosts": [{
            "host": "x.43.155.194",
            "port": 9200,
            "protocol": "https"
          }],
            "auth": { "apiKey": "VU9ETlNYY0J3cEtTcWdJT2pjLWM6ZXAwTldIRkRTTkMyYTNIdjNYZGRqZw==" }
        }
    curl request works
    curl -H "Authorization: ApiKey VU9ETlNYY0J3cEtTcWdJT2pjLWM6ZXAwTldIRkRTTkMyYTNIdjNYZGRqZw==" -k https://x.43.155.194:9200/_cluster/health
    
    {"cluster_name":"elasticsearch","status":"green","timed_out":false,"number_of_nodes":2,"number_of_data_nodes":2,"active_primary_shards":1,"active_shards":2,"relocating_shards":0,"initializing_shards":0,"unassigned_shards":0,"delayed_unassigned_shards":0,"number_of_pending_tasks":0,"number_of_in_flight_fetch":0,"task_max_waiting_in_queue_millis":0,"active_shards_percent_as_number":100.0}
    Phlegx Systems OG
    @phlegx
    Hi there! I was wondering that if with https://github.com/pelias/openstreetmap we can also import already existing files, or if we do have to always pass an URL to the file?
    3 replies
    Phlegx Systems OG
    @phlegx
    And another question: This whoisonfirst importer is not available in the helm chart, right? I'm running into this issue, when I try to import some OSM DB: pelias/openstreetmap#544 So that means I would need to run the whoisfirstimport first before the openstreetmap iimporter, right?
    1 reply
    Oleksandr Zholob
    @ShoSashko
    Hello everyone,
    I see from the documentation that I can't install Pelias from the portland-metro example on Windows locally.
    What If I am using linux-containers for docker on Windows? Is that still be a problem? :)
    spshealy
    @spshealy
    Does Pelias use Tiger files for forward geocoding of street addresses in the US? Reason I am asking isI modified the north-america docker project to download all the tiger files. But when I forward geocoded some street addresses(that have been around forever) in the county next to mine it can't find a high confidence hit and returns the centorid of the locality from whosonfirst. I notice that there is no openaddresses file(Richland County,SC) but I would still think the address would be estimate-able form TIGER address data. Perhaps i am misunderstanding how Pelias uses TIGER data. TIA
    benbiju
    @benbiju
    hi, i have setup pelias docker and imported OSM data. I am testing the /search endpoint and some fields are missing in the response compared to https://pelias.github.io/compare/#/v1/search. For example, when i searched for Nicollet Mall in the server, it returned
    "layer": "venue", "source": "openstreetmap", "source_id": "node/3452611421", "name": "The Nicollet Diner", "housenumber": "1428", "street": "Nicollet Avenue", "postalcode": "55403", "confidence": 0.8, "match_type": "fallback", "accuracy": "point", "label": "The Nicollet Diner",
    But if i search for Nicollet Mall in https://pelias.github.io/compare/#/v1/search?sources=osm&text=nicollet+mall, i am getting more fields like county, region, country names etc. Is there a way i can expose these information in the /search response from my pelias setup
    Phlegx Systems OG
    @phlegx
    For the whoseonefirst importer, do we have a PELIAS_CONFIG as well? Checked the code and doesn't seem to be there. How else can I set the filename for the pelias confi?
    2 replies
    Nirmal Kumar
    @nirmalkq

    Hi, I am getting this error with whoseonefirst importer . Any idea how to resolve this .Creating pelias_whosonfirst_run ... done
    /code/pelias/whosonfirst/import.js:18
    throw new Error(err.message);
    ^

    Error
    at /code/pelias/whosonfirst/import.js:18:11
    at getDBList (/code/pelias/whosonfirst/src/bundleList.js:54:12)
    at Object.getList [as generateBundleList] (/code/pelias/whosonfirst/src/bundleList.js:61:12)
    at Object.<anonymous> (/code/pelias/whosonfirst/import.js:15:9)
    at Module._compile (internal/modules/cjs/loader.js:955:30)
    at Object.Module._extensions..js (internal/modules/cjs/loader.js:991:10)
    at Module.load (internal/modules/cjs/loader.js:811:32)
    at Function.Module._load (internal/modules/cjs/loader.js:723:14)
    at Function.Module.runMain (internal/modules/cjs/loader.js:1043:10)
    at internal/main/run_main_module.js:17:11

    cyfugr
    @cyfugr
    Hi, i want to rank results that are first in a specific country and then everything else, any ideas ?
    kkflorian
    @kkflorian
    Does pelias provide pre-built search indices (like for example photon does)?