initial prototype for auto schema generation (piccolo-orm#157)

* initial prototype for auto schema generation

* make sure the correct column type is being instantiated

* use black to nicely format the string

* add Timestamptz

* detect if column is nullable

* introspect Varchar length

* slight refactor, and output required imports

* added initial tests

* fix tests

* remove some kwargs from output

* fix bug with duplicate id columns

* add extra imports if foreign keys are present

* add docs

* fixing a bug with primary key detection

* add test for unique constraints

* log the db version instead of printing to avoid adding noise to output

Otherwise if someone pipes `piccolo schema generate` straight to a file, it won't be valid Python as it'll have something like 'Running Postgres version 9.6' at the top.

* increase code coverage

* add a link from the schema docs to the `piccolo schema generate` docs

Author: dantownsend

docs/src/piccolo/projects_and_apps/included_apps.rst

@@ -60,6 +60,23 @@ Lets you create a new ``piccolo_conf.py`` file. See :ref:`PiccoloProjects`.
 
     piccolo project new
 
+.. _SchemaApp:
+
+schema
+~~~~~~
+
+Lets you auto generate Piccolo ``Table`` classes from an existing database.
+Make sure the credentials in ``piccolo_conf.py`` are for the database you're
+interested in, then run the following:
+
+.. code-block:: bash
+
+    piccolo schema generate > tables.py
+
+.. warning:: This feature is still a work in progress. However, even in it's
+    current form it will save you a lot of time. Make sure you check the
+    generated code to make sure it's correct.
+
 shell
 ~~~~~
 

docs/src/piccolo/schema/defining.rst

@@ -21,6 +21,10 @@ columns. Here's a very simple schema:
 
 For a full list of columns, see :ref:`ColumnTypes`.
 
+.. hint:: If you're using an existing database, see Piccolo's
+ :ref:`auto schema generation command<SchemaApp>`, which will save you some
+ time.
+
 -------------------------------------------------------------------------------
 
 Primary Key