Marmelator/image-classification-service
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases 29 Wiki Activity

refactor scanned page filtering WIP

Browse Source
This commit is contained in:
Julius Unverfehrt 2023-02-01 15:14:27 +01:00
parent 436a32ad2b
commit 0f440bdb09
1 changed files with 8 additions and 15 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

23
image_prediction/image_extractor/extractors/parsable.py
View File

@ -9,11 +9,14 @@ from typing import Iterable, Iterator, List
import fitz
from PIL import Image
from funcy import merge, pluck, curry, compose, rcompose
from funcy import merge, pluck, curry, compose, rcompose, remove
from image_prediction.formatter.formatters.enum import EnumFormatter
from image_prediction.image_extractor.extractor import ImageExtractor, ImageMetadataPair
from image_prediction.image_extractor.filters import filter_metadata_for_scanned_pages
from image_prediction.image_extractor.filters import (
filter_metadata_for_scanned_pages,
__breaches_image_to_page_quotient,
)
from image_prediction.info import Info
from image_prediction.stitching.stitching import stitch_pairs
from image_prediction.stitching.utils import validate_box_coords, validate_box_size
@ -47,7 +50,7 @@ class ParsablePDFImageExtractor(ImageExtractor):
def __process_images_on_page(self, page: fitz.fitz.Page):
metadata = list(get_metadata_for_images_on_page(page))
metadata = filter_metadata_for_scanned_pages(metadata)
metadata = list(filter_out_page_sized_images(metadata))
metadata = list(filter_out_tiny_images(metadata))
metadata = list(filter_invalid_metadata(metadata))
@ -108,18 +111,8 @@ def filter_invalid_metadata(metadata):
return compose(validate_size_and_passthrough, validate_coords_and_passthrough)(metadata)
# def get_metadata_for_images_on_page_2(page: fitz.fitz.Page):
# """Effectively the same as image_prediction.image_extractor.extractors.parsable.get_metadata_for_images_on_page,
# however without the validation steps since not required here and take a significant amount of time.
# """
# # temporary solution to avoid circular imports without changing the original code
# from image_prediction.image_extractor.extractors.parsable import get_image_metadata, add_page_metadata
#
# image_infos = page.get_image_info(xrefs=True)
# metadata = lmap(get_image_metadata, image_infos)
# metadata = add_page_metadata(page, metadata)
#
# return metadata
def filter_out_page_sized_images(metadata):
yield from remove(__breaches_image_to_page_quotient, metadata)
@lru_cache(maxsize=None)