MIME-Version: 1.0
References: <20200713154907.GH81733@artax.karlin.mff.cuni.cz>
In-Reply-To: <20200713154907.GH81733@artax.karlin.mff.cuni.cz>
From: Daniele Varrazzo <daniele.varrazzo@gmail.com>
Date: Mon, 13 Jul 2020 17:15:19 +0100
Message-ID: 
 <CA+mi_8bPLhcgA3KgyZ-g+=oSkUBjaB-JoZD5oAKVjPqbm8iiog@mail.gmail.com>
Subject: Re: register_adapter Json with custom JSONEncoder
To: Hans Ginzel <hans@matfyz.cz>
Cc: psycopg@lists.postgresql.org
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

Hello,

On Mon, 13 Jul 2020 at 16:49, Hans Ginzel <hans@matfyz.cz> wrote:
[...]

> But I need to use custom JSONEncoder because of bson.ObectId type
> # https://stackoverflow.com/a/16586277/2556118

This doc page has more examples:
https://www.psycopg.org/docs/extras.html#json-adaptation


> There is a parameter dumps in Json.__init__(self, adapted, dumps=None),
> but how to set it when used with register_adapter(),

You can use a partial function for instance, of the like of: `lambda
obj: Json(obj, dumps=mydumps)`.


> https://www.psycopg.org/docs/extensions.html#psycopg2.extensions.register_adapter?
>
> Should I write myself the whole Json class?

No, if you really want you can subclass the one we provide and only
override `_dumps()`: the one psycopg2 provides consists in pretty
much:

    class Json(object):
        def __init__(self, adapted, dumps=None):
            self.adapted = adapted
           self._dumps = dumps or json.dumps

        def dumps(self, obj):
            return self._dumps(obj)

so you can either create your instances using a wrapper in
register_adapter, or subclass the _dumps() method of the class: as you
can see it is equivalent. I guess the first method is better as you
don't need to care about the implementation of the `Json` class.


-- Daniele