def assemble_corpus(
corpus_reader: CorpusReader,
types_requested: List[str],
type_dirs: Dict[str, List[str]] = None,
type_files: Dict[str, List[str]] = None
) -> Tuple[CorpusReader, List[str], Set[str]]:
"""
Create a filtered corpus.
:param corpus_reader: This get mutated
:param types_requested: a list of string types, which are to be found in the type_dirs and
type_files mappings
:param type_dirs: a dict of corpus types to directories
:param type_files: a dict of corpus types to files
:return: a Tuple(CorpusReader object containing only the mappings desired,
fileid_names: A list of file ids of the matching corpus files, and
categories_found: a set of word categories used to build the reader
"""
fileid_names = [] # type: List[str]
categories_found = set() # type: Set[str]
try:
ALL_FILE_IDS = list(corpus_reader.fileids())
CLEAN_IDS_TYPES = [] # type: List[Tuple[str, str]]
if type_files:
for key, valuelist in type_files.items():
if key in types_requested:
for value in valuelist:
if value in ALL_FILE_IDS:
if key:
CLEAN_IDS_TYPES.append((value, key))
if type_dirs:
for key, valuelist in type_dirs.items():
if key in types_requested:
for value in valuelist:
corrected_dir = value.replace('./', '')
corrected_dir = '{}/'.format(corrected_dir)
for name in ALL_FILE_IDS:
if name and name.startswith(corrected_dir):
CLEAN_IDS_TYPES.append((name, key))
CLEAN_IDS_TYPES.sort(key=lambda x: x[0])
fileid_names, categories = zip(*CLEAN_IDS_TYPES) # type: ignore
categories_found = set(categories) # type: Set[str]
corpus_reader._fileids = fileid_names
except Exception:
LOG.exception('failure in corpus building')
return (corpus_reader, fileid_names, categories_found)
def assemble_corpus(
corpus_reader: CorpusReader,
types_requested: List[str],
type_dirs: Dict[str, List[str]] = None,
type_files: Dict[str, List[str]] = None,
) -> CorpusReader:
"""
Create a filtered corpus.
:param corpus_reader: This get mutated
:param types_requested: a list of string types, which are to be found in the type_dirs and
type_files mappings
:param type_dirs: a dict of corpus types to directories
:param type_files: a dict of corpus types to files
:return: a CorpusReader object containing only the mappings desired
"""
fileid_names = [] # type: List[str]
try:
all_file_ids = list(corpus_reader.fileids())
clean_ids_types = [] # type: List[Tuple[str, str]]
if type_files:
for key, valuelist in type_files.items():
if key in types_requested:
for value in valuelist:
if value in all_file_ids:
if key:
clean_ids_types.append((value, key))
if type_dirs:
for key, valuelist in type_dirs.items():
if key in types_requested:
for value in valuelist:
corrected_dir = value.replace("./", "")
corrected_dir = "{}/".format(corrected_dir)
for name in all_file_ids:
if name and name.startswith(corrected_dir):
clean_ids_types.append((name, key))
clean_ids_types.sort(key=lambda x: x[0])
fileid_names, categories = zip(*clean_ids_types) # type: ignore
corpus_reader._fileids = fileid_names
return corpus_reader
except Exception:
LOG.exception("failure in corpus building")
def assemble_corpus(corpus_reader: CorpusReader,
types_requested: List[str],
type_dirs: Dict[str, List[str]] = None,
type_files: Dict[str, List[str]] = None) -> CorpusReader:
"""
Create a filtered corpus.
:param corpus_reader: This get mutated
:param types_requested: a list of string types, which are to be found in the type_dirs and
type_files mappings
:param type_dirs: a dict of corpus types to directories
:param type_files: a dict of corpus types to files
:return: a CorpusReader object containing only the mappings desired
"""
fileid_names = [] # type: List[str]
try:
all_file_ids = list(corpus_reader.fileids())
clean_ids_types = [] # type: List[Tuple[str, str]]
if type_files:
for key, valuelist in type_files.items():
if key in types_requested:
for value in valuelist:
if value in all_file_ids:
if key:
clean_ids_types.append((value, key))
if type_dirs:
for key, valuelist in type_dirs.items():
if key in types_requested:
for value in valuelist:
corrected_dir = value.replace('./', '')
corrected_dir = '{}/'.format(corrected_dir)
for name in all_file_ids:
if name and name.startswith(corrected_dir):
clean_ids_types.append((name, key))
clean_ids_types.sort(key=lambda x: x[0])
fileid_names, categories = zip(*clean_ids_types) # type: ignore
corpus_reader._fileids = fileid_names
return corpus_reader
except Exception:
LOG.exception('failure in corpus building')
