parse address splits correct post code into house_number

Aknilam commented 4 years ago

Hi!

I was checking out libpostal, and saw something that could be improved.

My country is

DE

Here's how I'm using libpostal

rest api

Here's what I did

query=Wendenstr. 414-424 20537 Hamburg

Here's what I got

  {
    "label": "road",
    "value": "wendenstr. 414-424 2053"
  },
  {
    "label": "house_number",
    "value": "7"
  },
  {
    "label": "city",
    "value": "hamburg"
  }
]

Here's what I was expecting

  {
    "label": "road",
    "value": "wendenstr."
  },
  {
    "label": "house_number",
    "value": "414-424"
  },
  {
    "label": "postcode",
    "value": "20537"
  },
  {
    "label": "city",
    "value": "hamburg"
  }
]

For parsing issues, please answer "yes" or "no" to all that apply.

Does the input address exist in OpenStreetMap?

yes
Do all the toponyms exist in OSM (city, state, region names, etc.)?

yes
If the address uses a rare/uncommon format, does changing the order of the fields yield the correct result?

yes and no, but the provided format is common query=20537 Hamburg Wendenstr. 414-424 is also not working as then post code is detected as house_number as well
```
{
"label": "house_number",
"value": "20537"
},
{
"label": "road",
"value": "hamburg wendenstr."
},
{
"label": "house_number",
"value": "414-424"
}
]
```
If the address does not contain city, region, etc., does adding those fields to the input improve the result?

it contains city
If the address contains apartment/floor/sub-building information or uncommon formatting, does removing that help? Is there any minimum form of the address that gets the right parse?

only street, house_number (with hyphen!), post code and city are provided

Here's what I think could be improved

Fix for incorrect parsing/splitting of the address - somehow postcode is split and its last number is extracted as a house_number.

tobwen commented 3 years ago

libpostal seems to be trained against a schema like street name house number, postcode city-district for the German data, where "street name" isn't allowed to contain a number at the beginning, etc. Parsing works with Wendenstr. 414-424, 20537 Hamburg - I've just added the missing comma.

Just my 2 cent: I've stopped using libpostal, since the input data needs to be too clean.

ChargedMonk commented 2 years ago

Just my 2 cent: I've stopped using _libpostal, since the input data needs to be too clean.

@tobwen, Any good working alternative?

openvenues / libpostal