Any idea about
elias@1-openaddresses-rjk4r:/code/pelias/openaddresses$ ./bin/download
info: [openaddresses-download] Attempting to download all data
error: [openaddresses-download] Failed to download data message=Command failed: unzip -o -qq -d /data/openaddresses /tmp/202123-37-12tjcyy.caci.zip
[/tmp/202123-37-12tjcyy.caci.zip]
End-of-central-directory signature not found. Either this file is not ??
@hamzarhaiem You need to run some kind of Ingress controller like Istio or Traffik. This has nothing to do with Pelias.
Hi everyone,
I am currently running a pelias instance using custom csv-imported data for Switzerland.
It seems to be working quite well, but some addresses are not found.
For example, the following query:
qry = "Chemin de la Pré-de-la-Raisse 11"
qry = urllib.parse.quote(req)
requests.get(f"{pelias}search?text={qry}&layers=address").json()Yields:
{
"geocoding": {
"version": "0.2",
"attribution": "http://api.yourpelias.com/attribution",
"query": {
"text": "Chemin de la Pré-de-la-Raisse 11",
"size": 10,
"layers": ["address"],
"private": False,
"lang": {
"name": "English",
"iso6391": "en",
"iso6393": "eng",
"via": "default",
"defaulted": True,
},
"querySize": 20,
"parser": "pelias",
"parsed_text": {
"subject": "11 Chemin de la Pré-de-la-Raisse",
"street": "Chemin de la Pré-de-la-Raisse",
"housenumber": "11",
},
},
"engine": {"name": "Pelias", "author": "Mapzen", "version": "1.0"},
"timestamp": 1615458787863,
},
"type": "FeatureCollection",
"features": [],
}With no features.
While if I explore directly elasticsearch using this query:
GET pelias/_search
{
"query":{
"match": {
"address_parts.street": "Chemin du Pré-de-la-Raisse"
}
}
}I do have an exact match with the following document:
{
"_index" : "pelias",
"_type" : "_doc",
"_id" : "admin_ch:address:295517056-0",
"_score" : 17.968594,
"_source" : {
"center_point" : {
"lon" : 6.130037,
"lat" : 46.167932
},
"parent" : {
"continent" : [
"Europe"
],
"continent_id" : [
"102191581"
],
"continent_a" : [
null
],
"country" : [
"Switzerland"
],
"country_id" : [
"85633051"
],
"country_a" : [
"CHE"
],
"region" : [
"Geneva"
],
"region_id" : [
"85682291"
],
"region_a" : [
"GE"
],
"county" : [
"Genève"
],
"county_id" : [
"102062917"
],
"county_a" : [
null
],
"locality" : [
"Plan-les-Ouates"
],
"locality_id" : [
"1125887615"
],
"locality_a" : [
null
],
"localadmin" : [
"Plan-les-Ouates"
],
"localadmin_id" : [
"404328619"
],
"localadmin_a" : [
null
]
},
"name" : { },
"address_parts" : {
"street" : "Chemin du Pré-de-la-Raisse",
"number" : "5",
"zip" : "1228"
},
"source" : "admin_ch",
"source_id" : "295517056-0",
"layer" : "address"
}
}I don't understand why this particular exemple doesn't work, as many of my other test addresses do indeed work.
Does anyone have an idea where to look for? How to debug that?
Thanks in advance :)
As an example, if I wanted to find all points of interest in a specific area, I could do that with overpassAPI. Does this have the same functionality? And if not, curious as to whether it is due to a clear design choice or just a resource limitation. I see it does have the concept of places.
It describes the geocoder and reverse geocoder functionality.
Nice, didnt see that page although I did scour through the different repositories.
I guess I needed more clarification on the 'venues' portion.
A good example would be if I wanted to make a request targetting particularly OSM (but others as well if they have it) data on point of interest.
Something like 'all entertainment points of interest in a 50 mile radius of this point'
I can do that with the overpassAPI, I can understand the structure of the data and OSM has a whole wiki on POI.
From what ive seen you cannot do that with pelias.
However the reason I wanted to make sure is because the system in place could easily support this functionality given a bit of legwork, and this seems like a very well organized project!
Like I cant do 'waterfountains near london bridge' with the geocoder I dont think. Just wanted to confirm
Hello! With pelias-docker I would like to use my own openstreetmap pbf file. At the moment I just configured as import.openstreetmap.download.sourceURL a not existing URL. Then I copy my own pbf file into the directory before doing the pelias prepare all. While this works, it produces an error. I'm wondering if there is a better way of handling this situation?
I have noticed that when trying to forward geocode an address like this 10955 Old River RD , Komoka, ON, then pelias is not able to find/match the zipcode to the address.
Could someone advise me how I could include the candaian postal code geocode data from here: https://www.serviceobjects.com/blog/unique-us-canadian-zip-code-files-available-download/ into pelias ?
Could someone advise me how I could include the candaian postal code geocode data from here: https://www.serviceobjects.com/blog/unique-us-canadian-zip-code-files-available-download/ into pelias ?
and my total disk usage is staying more or less constant at around 163GB
not seeing any build output either
the odd thing is that my cpu is showing very low utilization.. but the script is still in a Interruptible Sleep state
I've got a North America build for Pelias and we've noticed some issues related to the source data (fields duplicated to incorrect fields, missing fields, etc.) and I was wondering if there was a way to identify and clean up such occurrences. Also, is there a preferred way to update the data? I can't seem to find any documentation addressing either of these questions.
@jcushner: Some are there and some are not. Missing zip codes are definitely one of the most common issues I'm seeing though.
I am also interested in a method to update the data
Yeah, I know I can certainly do a full re-download, then wipe the data from elasticsearch and re-import, but doing so would result in down-time, which is obviously problematic for larger datasets. I could also just try re-importing, but I am not certain if that will result in duplicated records.
Yeah, scrolling through the history here, it seems to have come up a few times. I'm not seeing any answers though. Definitely should be something mentioned in the docs, I think
Hello, I currently see pip-service loading only a small portion of features:
Is there any idea how to force loading more files?
"openstreetmap": {
"leveldbpath": "/tmp",
"datapath": "/share/data/osm",
"importVenues": true,
"import": [
{"filename": "*"}
]
"leveldbpath": "/tmp",
"datapath": "/share/data/osm",
"importVenues": true,
"import": [
{"filename": "*"}
]
this doesn't seem to work. i get an error that it "failed to stat pbf file: /share/data/osm/*
how do i use wildcards..
like if I only wanted a*, etc..
I see this param but doesn't make sense to me on how to use it imports.openstreetmap.import[0].filename
@jcushner: We have found that a large part of our data quality issues (particularly with autocomplete) are from the openaddresses us/fl/_loveland dataset. Have you noticed a similar pattern?
This dataset seems to be missing ZIP codes and has a very large number of occurrences where the house number and street name are the same number. Some spot checking finds that these are bogus addresses.