topaz-crystal / topaz Goto Github PK

I'd like to check if a model is persisted. It will help with validations later too.

class Thing < Topaz::Model
  columns(
    description: String
  )
end

x = Thing.new("a")
x.persisted?
# => false
x.save
x.persisted?
# => true

y = Thing.create("b")
y.persisted?
# => true

has_many(childlen: {model: SampleChild, key: parent_id})

The Crystal community has somewhat decided to use crystal as the language's topic.

To avoid fragmentation in the community, please drop the crystal-language topic, that only has ~28 repositories at the moment. We're sticking to crystal, instead.

Any contributor can change it from the repo's homepage.

See crystal-lang/crystal#4241 for more context.

Thank you very much :)

Is a DB connection opened on each save? Opening a connection is really, really slow. It should be done once in your application. I'm talking about these lines.

I've made a benchmark:

require "benchmark"
require "pg"

DB.open "postgres:///" do |db|
  db.exec "drop table if exists contacts"
  db.exec "create table contacts (num integer)"
end

shared_db = DB.open "postgres:///"

Benchmark.ips do |x|
  x.report("open each time") do
    DB.open "postgres:///" do |db|
      db.exec "insert into contacts (num) values (1)"
    end
  end
  x.report("open once") do
    shared_db.exec "insert into contacts (num) values (1)"
  end
end

Results:

open each time  73.53  (  13.6ms) (± 4.26%) 58.82× slower
     open once   4.32k (231.23µs) (± 1.68%)       fastest

Almost 60 times slower to open and close a connection each time.

I'm not sure, but it could be the reason why you don't get very good performance here

This is more a suggestion than an issue itself, I was just wondering if you guys are considering integrate Topaz with micrate for handling database migrations.

From all the ORM that I have been evaluating, I really like Topaz because for its simplicity but I feel that the way Topaz handles migrations will be not enough for large-scale projects.

Thanks

We may need to update the dependency builds for TravisCI

Get:1 http://ppa.launchpad.net/travis-ci/sqlite3/ubuntu precise/main amd64 libsqlite3-dev amd64 3.7.15.1-1~travis1 [582 kB]
Err:1 http://ppa.launchpad.net/travis-ci/sqlite3/ubuntu precise/main amd64 libsqlite3-dev amd64 3.7.15.1-1~travis1
  Hash Sum mismatch
Get:2 http://ppa.launchpad.net/travis-ci/sqlite3/ubuntu precise/main amd64 libsqlite3-0 amd64 3.7.15.1-1~travis1 [459 kB]
Err:2 http://ppa.launchpad.net/travis-ci/sqlite3/ubuntu precise/main amd64 libsqlite3-0 amd64 3.7.15.1-1~travis1
  Hash Sum mismatch
Get:3 http://ppa.launchpad.net/travis-ci/sqlite3/ubuntu precise/main amd64 sqlite3 amd64 3.7.15.1-1~travis1 [122 kB]
Err:3 http://ppa.launchpad.net/travis-ci/sqlite3/ubuntu precise/main amd64 sqlite3 amd64 3.7.15.1-1~travis1
  Hash Sum mismatch
Fetched 1,163 kB in 2s (572 kB/s)
E: Failed to fetch http://ppa.launchpad.net/travis-ci/sqlite3/ubuntu/pool/main/s/sqlite3/libsqlite3-dev_3.7.15.1-1~travis1_amd64.deb  Hash Sum mismatch
E: Failed to fetch http://ppa.launchpad.net/travis-ci/sqlite3/ubuntu/pool/main/s/sqlite3/libsqlite3-0_3.7.15.1-1~travis1_amd64.deb  Hash Sum mismatch
E: Failed to fetch http://ppa.launchpad.net/travis-ci/sqlite3/ubuntu/pool/main/s/sqlite3/sqlite3_3.7.15.1-1~travis1_amd64.deb  Hash Sum mismatch
E: Unable to fetch some archives, maybe run apt-get update or try with --fix-missing?
The command "sudo apt-get install sqlite3=3.7.15.1-1~travis1" failed and exited with 100 during .
Your build has been stopped.

This is causing all tests to fail (and not even run).

Just wanted to know if its possible to support existing db which has column names with spaces. Such as
"full name", "pin code", etc

After some testing and experimenting it looks like the default time format in the base Model class doesn't play nice with PG.

https://github.com/topaz-crystal/topaz/blob/master/src/topaz/model.cr#L9

Example using a sample table:

sampledb> insert into sample_table (name, created_at, updated_at) 
  values ('sdf', '2017-03-05 08:43:37.681', '2017-03-05 08:43:37.681');
INSERT 0 1
Time: 0.022s


sampledb> insert into sample_table (name, created_at, updated_at)
  values ('sdf', '2017-03-05 08:43:37:681', '2017-03-05 08:43:37:681');

invalid input syntax for type timestamp: "2017-03-05 08:43:37:681"
LINE 1: ...me, created_at, updated_at) values ('sdf', '2017-03-0...

First example using the period. Second example using the colon.

I overridded it in my own project's base model class and that looks like it works fine

require "topaz"

class SampleTable < Topaz::Model
  TIME_FORMAT = "%F %T.%L %z"

How do you update an existing model from JSON? I can't seem to figure out the use of from_json and update with this.

Say I want to have a default value for a model schema. Let's use a Todo app as an example:

class Todo < Topaz::Model
  columns(
    description: String,
    completion: Time?
  )
end

The completion parameter should be assumed to be nil if no value is given unless nullable: false has be set as in https://github.com/topaz-crystal/topaz/blob/master/spec/model/models.cr#L48.

So if the union has a type of Nil in it I'd like the parameters in the create method to have = nil in it by default. Again, to avoid that I'd use nullable: false.

As far as other default values I'm not sure how you'd want to do it. Maybe something like?

class Todo < Topaz::Model
  columns(
    description: String, {default: ""},
    attempts: Int32, {default: 0}
  )
end

Currently there is a way to create a many to many relationship with 2 models ?

Thanks