forked from Archives/langchain
add docstring and move encoding detection to own func (#4479)
This commit is contained in:
parent
191da06d9b
commit
760a4f72f4
@ -8,7 +8,18 @@ logger = logging.getLogger(__name__)
|
|||||||
|
|
||||||
|
|
||||||
class TextLoader(BaseLoader):
|
class TextLoader(BaseLoader):
|
||||||
"""Load text files."""
|
"""Load text files.
|
||||||
|
|
||||||
|
|
||||||
|
Args:
|
||||||
|
file_path: Path to the file to load.
|
||||||
|
|
||||||
|
encoding: File encoding to use. If `None`, the file will be loaded
|
||||||
|
with the default system encoding.
|
||||||
|
|
||||||
|
autodetect_encoding: Whether to try to autodetect the file encoding
|
||||||
|
if the specified encoding fails.
|
||||||
|
"""
|
||||||
|
|
||||||
def __init__(
|
def __init__(
|
||||||
self,
|
self,
|
||||||
@ -29,7 +40,7 @@ class TextLoader(BaseLoader):
|
|||||||
text = f.read()
|
text = f.read()
|
||||||
except UnicodeDecodeError as e:
|
except UnicodeDecodeError as e:
|
||||||
if self.autodetect_encoding:
|
if self.autodetect_encoding:
|
||||||
detected_encodings = self.detect_file_encodings()
|
detected_encodings = detect_file_encodings(self.file_path)
|
||||||
for encoding in detected_encodings:
|
for encoding in detected_encodings:
|
||||||
logger.debug("Trying encoding: ", encoding["encoding"])
|
logger.debug("Trying encoding: ", encoding["encoding"])
|
||||||
try:
|
try:
|
||||||
@ -48,15 +59,17 @@ class TextLoader(BaseLoader):
|
|||||||
metadata = {"source": self.file_path}
|
metadata = {"source": self.file_path}
|
||||||
return [Document(page_content=text, metadata=metadata)]
|
return [Document(page_content=text, metadata=metadata)]
|
||||||
|
|
||||||
def detect_file_encodings(self, timeout: int = 5) -> List[dict]:
|
def detect_file_encodings(file_path: str) -> List[dict]:
|
||||||
"""Try to detect the file encoding."""
|
"""Try to detect the file encoding."""
|
||||||
import chardet
|
import chardet
|
||||||
|
|
||||||
with open(self.file_path, "rb") as f:
|
|
||||||
rawdata = f.read()
|
|
||||||
encodings = chardet.detect_all(rawdata)
|
|
||||||
|
|
||||||
if all(encoding["encoding"] is None for encoding in encodings):
|
with open(file_path, "rb") as f:
|
||||||
raise RuntimeError(f"Could not detect encoding for {self.file_path}")
|
rawdata = f.read()
|
||||||
res = [encoding for encoding in encodings if encoding["encoding"] is not None]
|
encodings = chardet.detect_all(rawdata)
|
||||||
return cast(List[dict], res)
|
|
||||||
|
|
||||||
|
if all(encoding["encoding"] is None for encoding in encodings):
|
||||||
|
raise RuntimeError(f"Could not detect encoding for {file_path}")
|
||||||
|
res = [encoding for encoding in encodings if encoding["encoding"] is not None]
|
||||||
|
return cast(List[dict], res)
|
||||||
|
Loading…
Reference in New Issue
Block a user