I always get error, is it a bug?
And when I use aggregations like:
http --session=admin :6543/api/customers _aggs.my_agg.terms.field==state
I found the _aggs parameter can't be used to implement nested aggregations like:
{
"aggs": {
"colors": {
"terms": {
"field": "color"
},
"aggs": {
"avg_price": {
"avg": {
"field": "price"
}
}
}
}
}
} http --session=admin :6543/api/customers body=="{_source:false,aggs:{types:{terms:{field:\"type\"},aggs:{avg_id:{avg:{field:\"customer_id\"}}}}}}"
But I always get documents from server instead of a aggregation!
Most of important I found I can't change ES mappings in my schema
I think all nefertari or nefertari_sqa must be improved to solve above issues.
nefertari (0.7.0)
nefertari-sqla (0.4.2)
pyramid (1.7.3)
pyramid-jinja2 (2.6.2)
pyramid-sqlalchemy (1.6)
pyramid-tm (1.0)
ramses (0.5.3)
SQLAlchemy (1.1.2)
SQLAlchemy-Utils (0.32.9)
I've just installed ramses and tryed the pyramid scaffold associated
I've installed mongodb and elastcisearch, I manage to run the example
but I can't manage to POST to the api :/
when I run:
http POST 192.168.1.7:6543/api/items name=banane description='bananes jaunes'
I got:
"message": "Bad or missing param '_asdict'",
Any idea??
Hi there - have a couple of questions of the architecture of this product. 1) Can it be plugged in to an existing Elasticsearch index and provide an api for it. 2) Why do the tutorial seem to need both postgressql and elasticsearch? e.g https://www.elastic.co/blog/make-an-elasticsearch-powered-rest-api-for-any-data-with-ramses
@RHSman you might want to set your lift_name as a primary key:https://ramses.readthedocs.io/en/stable/fields.html#primary-key
re: ES geo data, it is not supported but it shouldn't be too hard to add. if you have some spare time and would like to implement it, feel free to submit a PR
@GHapps 1) it could but it would require to write a RAML file that describes the exact same schemas/mappings as your ES index. one way to achieve this would be to create a Ramses app, and compare the ES index generated by Ramses to the existing ES index. And then step-by-step, try to reproduce the same mappings. another way would be to write you RAML file and build your Ramses app without caring too much about the old EX index and write a script that takes the data in your old index and POSTs it to your new API... I am sure there are other ways.
2) currently Postgres (or Mongodb) is used for writes and Elasticsearch for reads, that's the way it is designed. there has been some discussions and work around decoupling the two and allowing reads/writes to Postgres only as well as r/w to Elasticsearch only for some scenarios, it mights makes sense actually. However that work has never been completed. Feel free to take a look though, this work is in a branch of Neferari (the lower-level API framework layer): https://github.com/ramses-tech/nefertari/tree/engine-refactor
I'm quite new to python, so my question may sound dumb, pardon me in advance. I was wondering why you've chosen Pyramid and not Falcon (https://falconframework.org) - is that intended?
@jstoiko The ability to choose a backend (pg or ES) I think would be really useful, I have had to 'manipulate' ES mappings to be for my use case (mostly identifying key records where I don't have an ID as an example). Mappings to ES, as a hack I override the mapping part as its just too limiting and then run a mapping against the indexes. If I have get time I'll get to that Mappings component as it really adds another layer of control via the RAML.
Re key: I set the lift_name as primary but what if I have a lift with same name but different resort? I have fixed my issue in the workflow but again, super useful at the RAML level.
GEO: Yes, I'll get on that as it really takes ES as a backend service to the next level!!!
Re key: I set the lift_name as primary but what if I have a lift with same name but different resort? I have fixed my issue in the workflow but again, super useful at the RAML level.
GEO: Yes, I'll get on that as it really takes ES as a backend service to the next level!!!
On another note, I'll be trying KiBi (https://siren.solutions/kibi/) on this as a logical analytics layer.
Hello, i have troubles making a many to many relationship with ramses. I think that ramses doesn't take it into account.
We have two tables Structures (IdStruct (pk), ...) & Fields (IdField (pk), ...) with a many to many relationship. Thus, this relationship results in a third table Propriety (IdStruct (fk), IdField (fk)).
We would like to implement these relations but failed to find some examples.... Do you know how to do that with Ramses?
We have two tables Structures (IdStruct (pk), ...) & Fields (IdField (pk), ...) with a many to many relationship. Thus, this relationship results in a third table Propriety (IdStruct (fk), IdField (fk)).
We would like to implement these relations but failed to find some examples.... Do you know how to do that with Ramses?
I am completely new to Ramses. When following the guide at https://www.elastic.co/blog/make-an-elasticsearch-powered-rest-api-for-any-data-with-ramses,
pserve local.ini complains Exception: TransportError(400, u'mapper_parsing_exception', u'No handler for type [string] declared on field [_pk]')
What might cause such error?
All setting is is default exception psql connection url is set to my local postgresql which is just a clean install as well (db is created with gii_api).
Change nefertari's elasticsearch.py file around at line 249 forcing type, obtained from model_cls.get_es_mapping(), in dict to text temporarily fix the problem.
But where are this data coming from? Which file or path do engine read that particular mapping? I've modified all items.json type of string to text already under folder venv and gii_api, but that doesn't fix the problem.
But it's weird even changing schema json file's type to text, ramses still complains _pk and country type is string
That is already been changed to "type": "text" within country section.
I have more than 1k fields so I would rather avoid manual annotate all fields.