pytion 1.2.3

Unofficial Python client for official Notion API

pytion

Independent unofficial Python client for the official Notion API (for internal integrations only)

Client is built with its own object model based on API

So if you are using notion.so and want to automate some stuff with the original API, you're welcome!
You can read any available data, create basic models, and even work with databases.

Current Notion API version = "2022-02-22"

*does not use notion-sdk-py client

Contents

1. Quick Start
2. Pytion API
   1. pytion.api.Element
3. Models
   1. pytion.models
   2. Supported block types
   3. Block creating examples
4. Logging

Quick start

pip install pytion

Create new integration and get your Notion API Token at notion.so -> here.
Invite your new integration 'manager' to your pages or databases.

from pytion import Notion; no = Notion(token=SOME_TOKEN)

Or put your token for Notion API into file token at script directory and use simple no = Notion()

from pytion import Notion
no = Notion(token=SOME_TOKEN)
page = no.pages.get("PAGE ID")  # retrieve page data (not content) and create object
blocks = page.get_block_children()  # retrieve page content and create list of objects
database = no.databases.get("Database ID")  # retrieve database data (not content) and create object
# retrieve database content by filtering with sorting
pages = database.db_filter(property_name="Done", property_type="checkbox", value=False, descending="title")

In [1]: from pytion import Notion

In [2]: no = Notion(token=SOME_TOKEN)

In [3]: page = no.pages.get("a458613160da45fa96367c8a594297c7")
In [4]: print(page)
Notion/pages/Page(Example page)

In [5]: blocks = page.get_block_children_recursive()

In [6]: print(blocks)
Notion/blocks/BlockArray(## Migration planning [x] Rese)

In [7]: print(blocks.obj)
## Migration planning
[x] Reset new switch 2022-05-12T00:00:00.000+03:00 → 2022-05-13T01:00:00.000+03:00 
	- reboot
	- hold reset button
[x] Connect to console with baud rate 9600
[ ] Skip default configuration dialog
Use LinkTo(configs) 
[Integration changes](https://developers.notion.com/changelog?page=2)

In [8]: print(blocks.obj.simple)
Migration planning
Reset new switch 2022-05-12T00:00:00.000+03:00 → 2022-05-13T01:00:00.000+03:00 
	reboot
	hold reset button
Connect to console with baud rate 9600
Skip default configuration dialog
Use https://api.notion.com/v1/pages/90ea1231865f4af28055b855c2fba267 
https://developers.notion.com/changelog?page=2

Pytion API

Almost all operations are carried out through Notion or Element object:

page = no.pages.get("a458613160da45fa96367c8a594297c7")

# no -> Notion
# pages -> URI in https://api.notion.com/v1/PAGES/a458613160da45fa96367c8a594297c7
# get -> pytion API method
# page -> Element
# page.obj -> Page (main data structure)

so:

• isinstance(no, Notion) == True
• isinstance(no.pages, Element) == True
• isinstance(no.databases, Element) == True
• isinstance(page, Element) == True
• isinstance(page.obj, Page) == True

and if you want to retrieve a database - you must use "databases" URI

database = no.databases.get("123412341234")

and the same applies for "blocks" and "users". Valid URI-s are:

• pages
• blocks
• databases
• users

When you work with existing Element object like page above, all methods below will be applied to this Page:

new_page = page.page_update(title="new page name 2")

# new_page is not a new page, it is updated page
# new_page.obj is equal page.obj except title and last_edited properties

pytion.api.Element

There is a list of available methods for communicate with api.notion.com. These methods are better structured in next chapter.

.get(id_) - Get Element by ID.

.get_parent(id_) - Get parent object of current object if possible.

.get_block_children(id_, limit) - Get children Block objects of current Block object (tabulated texts) if exist.

.get_block_children_recursive(id_, max_depth, limit, force) - Get children Block objects of current Block object (tabulated texts) if exist recursive.

.get_page_property(property_id, id_, limit) - Retrieve a page property item.

.db_query(id_, limit, filter_, sorts) - Query Database.

.db_filter(...see desc...) - Query Database.

.db_create(database_obj, parent, properties, title) - Create Database.

There is no way to delete a database object yet!

.db_update(id_, title, properties) - Update Database.

.page_create(page_obj, parent, properties, title) - Create Page.

.page_update(id_, properties, title, archived) - Update Page.

.block_update(id_, block_obj, new_text, archived) - Update text in Block.

.block_append(id_, block, blocks) - Append block or blocks children.

.from_linkto(linkto) - Creates new Element object based on LinkTo information.

.from_object(model) - Creates new Element object from Page, Block or Database object. Usable while Element object contains an Array.

More details and usage examples of these methods you can see into func descriptions.

Models

pytion.models

There are classes based on API structures:

• RichText based on Rich text object
• RichTextArray is a list of RichText objects with useful methods
  • You can create object by simple RichTextArray.create("My title text") and then use it in any methods
  • Any Rich Text getting from API will be RichTextArray
  • str() returns plain text of all "Rich Texts". ez
  • + operator is available with str and RichTextArray
  • .simple property returns stripped plain text without any markdown syntax. useful for URL
• Database based on Database object
  • You can create object Database.create(...) and/or use .db_create(...) API method
  • attrs represent API model. Complex structures like created_by are wrapped in internal objects
  • use .db_update() API method for modify a real database (for ex. properties or title)
  • use .db_query() to get all database content (it will be PageArray)
  • use .db_filter() to get database content with filtering and/or sorting
• Page based on Page object
  • You can create object Page.create(...) and/or use .page_create(...) API method
  • use .page_update() method to modify attributes or delete the page
  • use .get_block_children() to get page content (without nested blocks) (it will be BlockArray)
  • use .get_block_children_recursive() to get page content with nested blocks
  • use .get_page_property() to retrieve the specific PropertyValue of the page
• Block based on Block object
  • You can create object Block.create(...) of specific type from support matrix below and then use it while creating pages or appending
  • use .block_update() to replace content or change extension attributes or delete the block
  • use .block_append() to add a new block to a page or add a nested block to another block
  • use .get_block_children() to get first level nested blocks
  • use .get_block_children_recursive() to get all levels nested blocks
• User based on User object
  • You can create object User.create(...) and use it in some properties like people type property
  • You can retrieve more data about a User by his ID using .get()
• Property based on Property object
  • You can create object Property.create(...) while creating or editing database: .db_create() or .db_update()
  • formula, relation, rollup type properties configuration is not supported
• PropertyValue based on Property values
  • You can create object PropertyValue.create(...) to set or edit page properties by .page_create() or .page_update()
  • files, relation, formula, rollup type properties are not editable

There are also useful internal classes:

• BlockArray is found when API returns page content in "list of blocks" format
  • it is useful to represent all content by str()
  • also it has simple property like RichTextArray object
  • it automatically indents str output of nested blocks
• PageArray is found when API returns the result of database query (list of pages)
• LinkTo is basic internal model to link to any Notion object
  • You can create object LinkTo.create() and use it in many places and methods
  • use LinkTo(from_object=my_page1) to quickly create a link to any existing object of pytion.models
  • link property of LinkTo returns expanded URL

And every model has a .get() method that returns API friendly JSON.

Supported block types

At present the API only supports the block types which are listed in the reference below. Any unsupported block types will continue to appear in the structure, but only contain a type set to "unsupported". Colors are not yet supported.

Every Block has mandatory attributes and extension attributes. There are mandatory:

• id: str - UUID-64 without hyphens
• object: str - always "block" (from API)
• created_time: datetime - from API
• created_by: User - from API
• last_edited_time: datetime - from API
• last_edited_by: User - from API
• type: str - the type of block (from API)
• has_children: bool - does the block have children blocks (from API)
• archived: bool - does the block marked as deleted (from API)
• text: Union[str, RichTextArray] - main content
• simple: str - only simple text string (url expanded)

Extension attributes are listed below in support matrix:

Block Type	Description	Read support	Create support	Can have children	Extension attributes
paragraph	Simple Block with text	+	+	+
heading_1	Heading Block with text highest level	+	-	-
heading_2	Heading Block with text medium level	+	-	-
heading_3	Heading Block with text lowest level	+	-	-
bulleted_list_item	Text Block with bullet	+	-	+
numbered_list_item	Text Block with number	+	-	+
to_do	Text Block with checkbox	+	+	+	checked: bool
toggle	Text Block with toggle to children blocks	+	-	+
code	Text Block with code style	+	+	+	language: str, caption: RichTextArray
child_page	Page inside	+	-	+
child_database	Database inside	+	-	+
embed	Embed online content	+	-	-	caption: RichTextArray
image	Embed image content	+	-	-	caption: RichTextArray, expiry_time: datetime
video	Embed video content	+	-	-	caption: RichTextArray, expiry_time: datetime
file	Embed file content	+	-	-	caption: RichTextArray, expiry_time: datetime
pdf	Embed pdf content	+	-	-	caption: RichTextArray, expiry_time: datetime
bookmark	Block for URL Link	+	-	-	caption: RichTextArray
callout	Highlighted footnote text Block	+	-	+	icon: dict
quote	Text Block with quote style	+	-	+
equation	KaTeX compatible text Block	+	-	-
divider	Simple line to divide the page	+	-	-
table_of_contents	Block with content structure in the page	+	-	-
column		-	-	+
column_list		-	-	-
link_preview	Same as bookmark	+	-	-
synced_block	Block for synced content aka parent	+	-	+	synced_from: LinkTo
template	Template Block title	+	-	+
link_to_page	Block with link to particular page @...	+	-	-	link: LinkTo
table	Table Block with some attrs	+	-	+	table_width: int
table_row	Children Blocks with table row content	+	-	-
breadcrumb	Empty Block actually	+	-	-
unsupported	Blocks unsupported by API	+	-	-

API converts toggle heading Block to simple heading Block.

Block creating examples

Create paragraph block object and add it to Notion:

from pytion.models import Block
my_text_block = Block.create("Hello World!")
my_text_block = Block.create(text="Hello World!", type_="paragraph")  # the same

# indented append my block to other known block:
no.blocks.block_append("5f60073a9dda4a9c93a212a74a107359", block=my_text_block)

# append my block to a known page (in the end)
no.blocks.block_append("9796f2525016128d9af4bf12b236b555", block=my_text_block)  # the same operation actually

# another way to append:
my_page = no.pages.get("9796f2525016128d9af4bf12b236b555")
my_page.block_append(block=my_text_block)

Create to_do block object:

from pytion.models import Block
my_todo_block = Block.create("create readme documentation", type_="to_do")
my_todo_block2 = Block.create("add 'create' method", type_="to_do", checked=True)

Create code block object:

from pytion.models import Block
my_code_block = Block.create("code example here", type_="code", language="javascript")
my_code_block2 = Block.create("another code example", type_="code", caption="it will be plain text code block with caption")

Logging

Logging is muted by default. To enable to stdout and/or to file:

from pytion import setup_logging

setup_logging(level="debug", to_console=True, filename="pytion.log")