kabutobashi

Core Concept

@block-decorator and Flow-class is important. @block automatically generates input and output functions, allowing you to focus solely on the processing. Flow allows you to focus solely on the process flow and input parameters.

About `@block`-decorator

simple decorator is like below.

def simple_decorator(func):
    def wrap_func() -> str:
        res = func()
        return f"Hello, {res}"
    return wrap_func


@simple_decorator
def world() -> str:
    return "world"


world()  # => "Hello, world"

A decorator is something that dynamically generates and adds processes to functions or classes, similar to its name.

First, prepare a function as follows and decorate it with @block.

from kabutobashi import block

@block()
class UdfBlock:
    term: int = 10

    def _process(self):
        return {"doubled_term": self.term * 2}

The classes above is equivalent to the following class definition.

import pandas as pd
from kabutobashi.domain.entity.blocks import BlockGlue

class UdfBlock:
    series: pd.DataFrame = None
    params: dict = None
    term: int = 10
    block_name: str = "udf_block"

    def _process(self) -> dict:
        return {"doubled_term": self.term * 2}
    
    def process(self) -> BlockGlue:
        # _process() method can be Tuple[Optional[dict], Optional[pd.DataFrame]]
        res = self._process()
        return BlockGlue(params=res, series=None, block_outputs={})

    def factory(self, glue: BlockGlue) -> "UdfBlock":
        # Omitted. In reality, processes are described.
        ...

    def _factory(self, glue: BlockGlue) -> dict:
        # Omitted. In reality, processes are described.
        ...

    def glue(self, glue: BlockGlue) -> BlockGlue:
        # Omitted. In reality, processes are described.
        ...

In classes decorated with @block, it is not recommended to execute the __init__() method. Instead, it is recommended to use the factory() class-method.

factory() method description. process() method description. glue() method description.

sequenceDiagram
  autonumber
  participant G as glue()
  participant UC as UdfBlock::class
  create participant S1 as factory()
  UC->>S1: create
  create participant S2 as _factory()
  UC->>S2: create or defined by user
  create participant P1 as process()
  UC->>P1: create
  create participant P2 as _process()
  UC->>P2: create or defined by user
  Note over S1: Generate udf_block_instance
  G->>+S1: Request
  S1->>+S2: Request
  Note over S2: User can modify _factory()
  S2-->>S2: get params from glue
  S2-->>S2: get series from glue
  S2-->>-S1: params and series
  create participant UI as UdfBlock::instance
  S1->>UI: UdfBlock(params, series)
  S1->>UI: setattr params to udf_block_instance
  S1-->>-G: udf_block_instance
  G->>+UI: udf_block_instance.process()
  UI->>+P1: process()
  Note over P1: execute process()
  P1->>P2: Request
  Note over P2: execute user defined function
  P2-->>P1: params or series
  P1-->>-UI: BlockGlue(params, series)
  UI-->>-G: block_glue_instance

Up to this point, the use of the @block decorator with classes such as UdfClass has described, but using the Block class on its own is not intended. Please read the following explanation of the Flow class for more details.

About `Flow`-class

Blocks are meant to be combined.

Processes always consist of combinations of multiple simple operations. And the only tedious part is aligning their inputs and outputs.

Therefore, in Flow-class, it automatically resolves the sequence of those processes for users, as long as you provide the initial values.

usage

import kabutobashi as kb

# n日前までの営業日の日付リストを取得する関数
target_date = "2020-01-01"
date_list = kb.get_past_n_days(target_date, n=40)

initialize Database

import kabutobashi as kb
kb.KabutobashiDatabase().initialize()

# add data
kb.crawl_info_multiple(code="1375", page="1", database_dir="...")
kb.crawl_info_multiple(code="1375", page="2", database_dir="...")
kb.crawl_info_multiple(code="1375", page="etc...", database_dir="...")

# add data daily
kb.crawl_info(code="1375", database_dir="...")

# analysis and add data
kb.analysis(code="1375", database_dir="...")

Metadata

Release files for kabutobashi 0.8.7

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for kabutobashi 0.8.7
File	Size	Uploaded
kabutobashi-0.8.7.tar.gz	32.7 kB	Aug 16, 2024	Details

Built distribution (wheel)

Table of built distributions (wheels) for kabutobashi 0.8.7
File	Interpreter	ABI	Platform	Reset
kabutobashi-0.8.7-py3-none-any.whl 56.4 kB Aug 16, 2024	Python 3	none	any	Details

Total release size: 89.2 kB

Release files / kabutobashi-0.8.7.tar.gz

Download URL	kabutobashi-0.8.7.tar.gz
Size	32.7 kB
Tags	Source
SHA-256 checksum How to use checksums	`26b1b575addf0e621acb78f6b8f893c7b7745a364ccbdb245bc993206d821f89`
BLAKE2b-256 checksum How to use checksums	`769167713d64c5aa88332d107d3ef48012e1a87eff54e945c37278af4edd55ad`
Upload date	Aug 16, 2024
Uploaded using Trusted Publishing? What is trusted publishing?	No
Uploaded via	`poetry/1.8.3 CPython/3.10.14 Linux/6.5.0-1025-azure`

Release files / kabutobashi-0.8.7-py3-none-any.whl

Download URL	kabutobashi-0.8.7-py3-none-any.whl
Size	56.4 kB
Tags	Python 3
SHA-256 checksum How to use checksums	`976d719b2d57aec1c185119e10249a6358c0cc199a25a39dd6e6eb6fbac738be`
BLAKE2b-256 checksum How to use checksums	`78440e7006b9436aba56ba6c4ed4246496e7066db871d255b05f43b09a9dbf35`
Upload date	Aug 16, 2024
Uploaded using Trusted Publishing? What is trusted publishing?	No
Uploaded via	`poetry/1.8.3 CPython/3.10.14 Linux/6.5.0-1025-azure`