Skip to content

Commit

Permalink
Cria módulo de segmentação para suportar novos segmentadores
Browse files Browse the repository at this point in the history
- Cria módulo `segmentation` para organizar o código dos segmentadores
- O módulo segue o padrão de classe base e classes derivadas para cada
  segmentador específico
- O primeiro segmentador base é o `AssociationSegmenter` para segmentar
  documentos de associações de municípios
- Uma dataclass para segmentos `GazetteSegment` também foi criada
- Os segmentadores são instanciados por meio de uma fábrica
  `get_segmenter`
  • Loading branch information
Jefersonalves authored and ogecece committed Nov 27, 2023
1 parent 9555960 commit 14537e2
Show file tree
Hide file tree
Showing 12 changed files with 276 additions and 182 deletions.
5 changes: 5 additions & 0 deletions segmentation/__init__.py
Original file line number Diff line number Diff line change
@@ -0,0 +1,5 @@
from .factory import get_segmenter

__all__ = [
"get_segmenter",
]
7 changes: 7 additions & 0 deletions segmentation/base/__init__.py
Original file line number Diff line number Diff line change
@@ -0,0 +1,7 @@
from .gazette_segment import GazetteSegment
from .association_segmenter import AssociationSegmenter

__all__ = [
"GazetteSegment",
"AssociationSegmenter",
]
27 changes: 27 additions & 0 deletions segmentation/base/association_segmenter.py
Original file line number Diff line number Diff line change
@@ -0,0 +1,27 @@
from typing import Union, Dict, List
from segmentation.base import GazetteSegment


class AssociationSegmenter:
def __init__(self, association_gazette: str):
self.association_gazette = association_gazette

def get_gazette_segments(self, *args, **kwargs) -> List[Union[GazetteSegment, Dict]]:
"""
Returns a list of GazetteSegment
"""
raise NotImplementedError

def split_text_by_territory(self, *args, **kwargs) -> Union[Dict[str, str], List[str]]:
"""
Segment a association text by territory
and returns a list of text segments
"""
raise NotImplementedError

def build_segment(self, *args, **kwargs) -> GazetteSegment:
"""
Returns a GazetteSegment
"""
raise NotImplementedError

26 changes: 26 additions & 0 deletions segmentation/base/gazette_segment.py
Original file line number Diff line number Diff line change
@@ -0,0 +1,26 @@
from datetime import date, datetime
from dataclasses import dataclass


@dataclass
class GazetteSegment:
"""
Dataclass to represent a gazette segment of a association
related to a city
"""
territory_name: str
source_text: str
date: date
edition_number: str
is_extra_edition: bool
power: str
file_checksum: str
scraped_at: datetime
created_at: datetime
processed: bool
file_path: str
file_url: str
state_code: str
territory_id: str
file_raw_txt: str
url: str
42 changes: 42 additions & 0 deletions segmentation/factory.py
Original file line number Diff line number Diff line change
@@ -0,0 +1,42 @@
from typing import Any

from segmentation.base import AssociationSegmenter
from segmentation import segmenters


def get_segmenter(territory_id: str, association_gazzete: dict[str, Any]) -> AssociationSegmenter:
"""
Factory method to return a AssociationSegmenter
Example
-------
>>> association_gazette = {
"territory_name": "Associação",
"created_at": datetime.datetime.now(),
"date": datetime.datetime.now(),
"edition_number": 1,
"file_path": 'raw/pdf.pdf',
"file_url": 'localhost:8000/raw/pdf.pdf',
"is_extra_edition": True,
"power": 'executive',
"scraped_at": datetime.datetime.now(),
"state_code": 'AL',
"source_text": texto,
}
>>> from segmentation import get_segmenter
>>> segmenter = get_segmenter(territory_id, association_gazette)
>>> segments = segmenter.get_gazette_segments()
Notes
-----
This method implements a factory method pattern.
See: https://github.com/faif/python-patterns/blob/master/patterns/creational/factory.py
"""

territory_to_segmenter_class = {
"2700000": "ALAssociacaoMunicipiosSegmenter",
}

segmenter_class_name = territory_to_segmenter_class[territory_id]
segmenter_class = getattr(segmenters, segmenter_class_name)
return segmenter_class(association_gazzete)
5 changes: 5 additions & 0 deletions segmentation/segmenters/__init__.py
Original file line number Diff line number Diff line change
@@ -0,0 +1,5 @@
from .al_associacao_municipios import ALAssociacaoMunicipiosSegmenter

__all__ = [
"ALAssociacaoMunicipiosSegmenter",
]
136 changes: 136 additions & 0 deletions segmentation/segmenters/al_associacao_municipios.py
Original file line number Diff line number Diff line change
@@ -0,0 +1,136 @@
import re

from typing import Any
from segmentation.base import AssociationSegmenter, GazetteSegment
from tasks.utils import get_checksum

class ALAssociacaoMunicipiosSegmenter(AssociationSegmenter):
def __init__(self, association_gazzete: dict[str, Any]):
super().__init__(association_gazzete)
# No final do regex, existe uma estrutura condicional que verifica se o próximo match é um \s ou SECRETARIA. Isso foi feito para resolver um problema no diário de 2018-10-02, em que o município de Coité do Nóia não foi percebido pelo código. Para resolver isso, utilizamos a próxima palavra (SECRETARIA) para tratar esse caso.
# Exceções Notáveis
# String: VAMOS, município Poço das Trincheiras, 06/01/2022, ato CCB3A6AB
self.RE_NOMES_MUNICIPIOS = (
r"ESTADO DE ALAGOAS(?:| )\n{1,2}PREFEITURA MUNICIPAL DE (.*\n{0,2}(?!VAMOS).*$)\n\s(?:\s|SECRETARIA)"
)
self.association_source_text = self.association_gazette["source_text"]

def get_gazette_segments(self) -> list[dict[str, Any]]:
"""
Returns a list of dicts with the gazettes metadata
"""
territory_to_text_split = self.split_text_by_territory()
gazette_segments = []
for municipio, texto_diario in territory_to_text_split.items():
segmento = self.build_segment(municipio, texto_diario)
gazette_segments.append(segmento.__dict__)
return gazette_segments

def split_text_by_territory(self) -> dict[str, str]:
"""
Segment a association text by territory
and returns a dict with the territory name and the text segment
"""
texto_diario_slice = self.association_source_text.lstrip().splitlines()

# Processamento
linhas_apagar = [] # slice de linhas a ser apagadas ao final.
ama_header = texto_diario_slice[0]
ama_header_count = 0
codigo_count = 0
codigo_total = self.association_source_text.count("Código Identificador")

for num_linha, linha in enumerate(texto_diario_slice):
# Remoção do cabeçalho AMA, porém temos que manter a primeira aparição.
if linha.startswith(ama_header):
ama_header_count += 1
if ama_header_count > 1:
linhas_apagar.append(num_linha)

# Remoção das linhas finais
if codigo_count == codigo_total:
linhas_apagar.append(num_linha)
elif linha.startswith("Código Identificador"):
codigo_count += 1

# Apagando linhas do slice
texto_diario_slice = [l for n, l in enumerate(
texto_diario_slice) if n not in linhas_apagar]

# Inserindo o cabeçalho no diário de cada município.
territory_to_text_split = {}
nomes_municipios = re.findall(
self.RE_NOMES_MUNICIPIOS, self.association_source_text, re.MULTILINE)
for municipio in nomes_municipios:
nome_municipio_normalizado = self._normalize_territory_name(municipio)
territory_to_text_split[nome_municipio_normalizado] = ama_header + '\n\n'

num_linha = 0
municipio_atual = None
while num_linha < len(texto_diario_slice):
linha = texto_diario_slice[num_linha].rstrip()

if linha.startswith("ESTADO DE ALAGOAS"):
nome = self._extract_territory_name(texto_diario_slice, num_linha)
if nome is not None:
nome_normalizado = self._normalize_territory_name(nome)
municipio_atual = nome_normalizado

# Só começa, quando algum muncípio for encontrado.
if municipio_atual is None:
num_linha += 1
continue

# Conteúdo faz parte de um muncípio
territory_to_text_split[municipio_atual] += linha + '\n'
num_linha += 1

return territory_to_text_split

def build_segment(self, territory, segment_text) -> GazetteSegment:
file_checksum = get_checksum(segment_text)
processed = True
territory_name = territory
source_text = segment_text.rstrip()

# TODO: get territory data and replace the None values
territory_id = None
# file_raw_txt = f"/{territory_id}/{date}/{file_checksum}.txt"
file_raw_txt = None
# url = file_raw_txt
url = None

return GazetteSegment(
# same association values
created_at=self.association_gazette.get("created_at"),
date=self.association_gazette.get("date"),
edition_number=self.association_gazette.get("edition_number"),
file_path=self.association_gazette.get("file_path"),
file_url=self.association_gazette.get("file_url"),
is_extra_edition=self.association_gazette.get("is_extra_edition"),
power=self.association_gazette.get("power"),
scraped_at=self.association_gazette.get("scraped_at"),
state_code=self.association_gazette.get("state_code"),
url=self.association_gazette.get("url"),

# segment specific values
file_checksum=file_checksum,
processed=processed,
territory_name=territory_name,
source_text=source_text,
territory_id=territory_id,
file_raw_txt=file_raw_txt,
)

def _normalize_territory_name(self, municipio: str) -> str:
municipio = municipio.rstrip().replace('\n', '') # limpeza inicial
# Alguns nomes de municípios possuem um /AL no final, exemplo: Viçosa no diário 2022-01-17, ato 8496EC0A. Para evitar erros como "vicosa-/al-secretaria-municipal...", a linha seguir remove isso.
municipio = re.sub("(\/AL.*|GABINETE DO PREFEITO.*|PODER.*|http.*|PORTARIA.*|Extrato.*|ATA DE.*|SECRETARIA.*|Fundo.*|SETOR.*|ERRATA.*|- AL.*|GABINETE.*)", "", municipio)
return municipio

def _extract_territory_name(self, texto_diario_slice: list[str], num_linha: int):
texto = '\n'.join(texto_diario_slice[num_linha:num_linha+10])
match = re.findall(self.RE_NOMES_MUNICIPIOS, texto, re.MULTILINE)
if len(match) > 0:
return match[0].strip().replace('\n', '')
return None
2 changes: 1 addition & 1 deletion tasks/__init__.py
Original file line number Diff line number Diff line change
Expand Up @@ -11,4 +11,4 @@
)
from .list_gazettes_to_be_processed import get_gazettes_to_be_processed
from .list_territories import get_territories_gazettes

from .get_territorie_info import get_territorie_info
78 changes: 0 additions & 78 deletions tasks/diario_ama.py

This file was deleted.

Loading

0 comments on commit 14537e2

Please sign in to comment.