def generate_packaging_keyword_processor(packaging: Optional[List[str]] = None):
p = (
text_file_iter(settings.OCR_PACKAGING_DATA_PATH)
if packaging is None
else packaging
)
return generate_keyword_processor(p)
def generate_brand_keyword_processor(
brands: Iterable[str], blacklist: bool = True,
):
blacklisted_brands: Optional[Set[str]] = None
if blacklist:
blacklisted_brands = BRAND_BLACKLIST_STORE.get()
keep_func = functools.partial(
keep_brand_from_taxonomy, blacklisted_brands=blacklisted_brands,
)
return generate_keyword_processor(brands, keep_func=keep_func)
def generate_label_keyword_processor(labels: Optional[List[str]] = None):
if labels is None:
labels = text_file_iter(settings.OCR_LABEL_FLASHTEXT_DATA_PATH)
return generate_keyword_processor(labels)
def generate_fishing_code_keyword_processor() -> KeywordProcessor:
codes = text_file_iter(settings.OCR_FISHING_FLASHTEXT_DATA_PATH)
return generate_keyword_processor(
("{}||{}".format(c.upper(), c) for c in codes))
def generate_trace_keyword_processor(labels: Optional[List[str]] = None):
if labels is None:
labels = list(text_file_iter(settings.OCR_TRACE_ALLERGEN_DATA_PATH))
return generate_keyword_processor(labels)