fixed linting
Browse files- lightrag/api/lightrag_server.py +49 -45
- lightrag/api/requirements.txt +1 -1
lightrag/api/lightrag_server.py
CHANGED
@@ -175,7 +175,11 @@ def parse_args():
|
|
175 |
class DocumentManager:
|
176 |
"""Handles document operations and tracking"""
|
177 |
|
178 |
-
def __init__(
|
|
|
|
|
|
|
|
|
179 |
self.input_dir = Path(input_dir)
|
180 |
self.supported_extensions = supported_extensions
|
181 |
self.indexed_files = set()
|
@@ -357,26 +361,22 @@ def create_app(args):
|
|
357 |
),
|
358 |
)
|
359 |
|
360 |
-
|
361 |
-
|
362 |
async def index_file(file_path: Union[str, Path]) -> None:
|
363 |
-
"""
|
364 |
-
|
365 |
Args:
|
366 |
file_path: Path to the file to be indexed (str or Path object)
|
367 |
-
|
368 |
Raises:
|
369 |
ValueError: If file format is not supported
|
370 |
FileNotFoundError: If file doesn't exist
|
371 |
"""
|
372 |
if not pm.is_installed("aiofiles"):
|
373 |
pm.install("aiofiles")
|
374 |
-
|
375 |
-
|
376 |
-
|
377 |
# Convert to Path object if string
|
378 |
file_path = Path(file_path)
|
379 |
-
|
380 |
# Check if file exists
|
381 |
if not file_path.exists():
|
382 |
raise FileNotFoundError(f"File not found: {file_path}")
|
@@ -384,23 +384,24 @@ def create_app(args):
|
|
384 |
content = ""
|
385 |
# Get file extension in lowercase
|
386 |
ext = file_path.suffix.lower()
|
387 |
-
|
388 |
match ext:
|
389 |
case ".txt" | ".md":
|
390 |
# Text files handling
|
391 |
async with aiofiles.open(file_path, "r", encoding="utf-8") as f:
|
392 |
content = await f.read()
|
393 |
-
|
394 |
case ".pdf":
|
395 |
if not pm.is_installed("pypdf2"):
|
396 |
pm.install("pypdf2")
|
397 |
from pypdf2 import PdfReader
|
|
|
398 |
# PDF handling
|
399 |
reader = PdfReader(str(file_path))
|
400 |
content = ""
|
401 |
for page in reader.pages:
|
402 |
content += page.extract_text() + "\n"
|
403 |
-
|
404 |
case ".docx":
|
405 |
if not pm.is_installed("docx"):
|
406 |
pm.install("docx")
|
@@ -409,11 +410,12 @@ def create_app(args):
|
|
409 |
# Word document handling
|
410 |
doc = Document(file_path)
|
411 |
content = "\n".join([paragraph.text for paragraph in doc.paragraphs])
|
412 |
-
|
413 |
case ".pptx":
|
414 |
if not pm.is_installed("pptx"):
|
415 |
pm.install("pptx")
|
416 |
from pptx import Presentation
|
|
|
417 |
# PowerPoint handling
|
418 |
prs = Presentation(file_path)
|
419 |
content = ""
|
@@ -421,7 +423,7 @@ def create_app(args):
|
|
421 |
for shape in slide.shapes:
|
422 |
if hasattr(shape, "text"):
|
423 |
content += shape.text + "\n"
|
424 |
-
|
425 |
case _:
|
426 |
raise ValueError(f"Unsupported file format: {ext}")
|
427 |
|
@@ -433,9 +435,6 @@ def create_app(args):
|
|
433 |
else:
|
434 |
logging.warning(f"No content extracted from file: {file_path}")
|
435 |
|
436 |
-
|
437 |
-
|
438 |
-
|
439 |
@app.on_event("startup")
|
440 |
async def startup_event():
|
441 |
"""Index all files in input directory during startup"""
|
@@ -559,6 +558,7 @@ def create_app(args):
|
|
559 |
)
|
560 |
except Exception as e:
|
561 |
raise HTTPException(status_code=500, detail=str(e))
|
|
|
562 |
@app.post(
|
563 |
"/documents/file",
|
564 |
response_model=InsertResponse,
|
@@ -566,14 +566,14 @@ def create_app(args):
|
|
566 |
)
|
567 |
async def insert_file(file: UploadFile = File(...), description: str = Form(None)):
|
568 |
"""Insert a file directly into the RAG system
|
569 |
-
|
570 |
Args:
|
571 |
file: Uploaded file
|
572 |
description: Optional description of the file
|
573 |
-
|
574 |
Returns:
|
575 |
InsertResponse: Status of the insertion operation
|
576 |
-
|
577 |
Raises:
|
578 |
HTTPException: For unsupported file types or processing errors
|
579 |
"""
|
@@ -581,19 +581,19 @@ def create_app(args):
|
|
581 |
content = ""
|
582 |
# Get file extension in lowercase
|
583 |
ext = Path(file.filename).suffix.lower()
|
584 |
-
|
585 |
match ext:
|
586 |
case ".txt" | ".md":
|
587 |
# Text files handling
|
588 |
text_content = await file.read()
|
589 |
content = text_content.decode("utf-8")
|
590 |
-
|
591 |
case ".pdf":
|
592 |
if not pm.is_installed("pypdf2"):
|
593 |
pm.install("pypdf2")
|
594 |
from pypdf2 import PdfReader
|
595 |
from io import BytesIO
|
596 |
-
|
597 |
# Read PDF from memory
|
598 |
pdf_content = await file.read()
|
599 |
pdf_file = BytesIO(pdf_content)
|
@@ -601,25 +601,27 @@ def create_app(args):
|
|
601 |
content = ""
|
602 |
for page in reader.pages:
|
603 |
content += page.extract_text() + "\n"
|
604 |
-
|
605 |
case ".docx":
|
606 |
if not pm.is_installed("docx"):
|
607 |
pm.install("docx")
|
608 |
from docx import Document
|
609 |
from io import BytesIO
|
610 |
-
|
611 |
# Read DOCX from memory
|
612 |
docx_content = await file.read()
|
613 |
docx_file = BytesIO(docx_content)
|
614 |
doc = Document(docx_file)
|
615 |
-
content = "\n".join(
|
616 |
-
|
|
|
|
|
617 |
case ".pptx":
|
618 |
if not pm.is_installed("pptx"):
|
619 |
pm.install("pptx")
|
620 |
from pptx import Presentation
|
621 |
from io import BytesIO
|
622 |
-
|
623 |
# Read PPTX from memory
|
624 |
pptx_content = await file.read()
|
625 |
pptx_file = BytesIO(pptx_content)
|
@@ -629,7 +631,7 @@ def create_app(args):
|
|
629 |
for shape in slide.shapes:
|
630 |
if hasattr(shape, "text"):
|
631 |
content += shape.text + "\n"
|
632 |
-
|
633 |
case _:
|
634 |
raise HTTPException(
|
635 |
status_code=400,
|
@@ -641,10 +643,10 @@ def create_app(args):
|
|
641 |
# Add description if provided
|
642 |
if description:
|
643 |
content = f"{description}\n\n{content}"
|
644 |
-
|
645 |
await rag.ainsert(content)
|
646 |
logging.info(f"Successfully indexed file: {file.filename}")
|
647 |
-
|
648 |
return InsertResponse(
|
649 |
status="success",
|
650 |
message=f"File '{file.filename}' successfully inserted",
|
@@ -661,6 +663,7 @@ def create_app(args):
|
|
661 |
except Exception as e:
|
662 |
logging.error(f"Error processing file {file.filename}: {str(e)}")
|
663 |
raise HTTPException(status_code=500, detail=str(e))
|
|
|
664 |
@app.post(
|
665 |
"/documents/batch",
|
666 |
response_model=InsertResponse,
|
@@ -668,13 +671,13 @@ def create_app(args):
|
|
668 |
)
|
669 |
async def insert_batch(files: List[UploadFile] = File(...)):
|
670 |
"""Process multiple files in batch mode
|
671 |
-
|
672 |
Args:
|
673 |
files: List of files to process
|
674 |
-
|
675 |
Returns:
|
676 |
InsertResponse: Status of the batch insertion operation
|
677 |
-
|
678 |
Raises:
|
679 |
HTTPException: For processing errors
|
680 |
"""
|
@@ -686,41 +689,43 @@ def create_app(args):
|
|
686 |
try:
|
687 |
content = ""
|
688 |
ext = Path(file.filename).suffix.lower()
|
689 |
-
|
690 |
match ext:
|
691 |
case ".txt" | ".md":
|
692 |
text_content = await file.read()
|
693 |
content = text_content.decode("utf-8")
|
694 |
-
|
695 |
case ".pdf":
|
696 |
if not pm.is_installed("pypdf2"):
|
697 |
pm.install("pypdf2")
|
698 |
from pypdf2 import PdfReader
|
699 |
from io import BytesIO
|
700 |
-
|
701 |
pdf_content = await file.read()
|
702 |
pdf_file = BytesIO(pdf_content)
|
703 |
reader = PdfReader(pdf_file)
|
704 |
for page in reader.pages:
|
705 |
content += page.extract_text() + "\n"
|
706 |
-
|
707 |
case ".docx":
|
708 |
if not pm.is_installed("docx"):
|
709 |
pm.install("docx")
|
710 |
from docx import Document
|
711 |
from io import BytesIO
|
712 |
-
|
713 |
docx_content = await file.read()
|
714 |
docx_file = BytesIO(docx_content)
|
715 |
doc = Document(docx_file)
|
716 |
-
content = "\n".join(
|
717 |
-
|
|
|
|
|
718 |
case ".pptx":
|
719 |
if not pm.is_installed("pptx"):
|
720 |
pm.install("pptx")
|
721 |
from pptx import Presentation
|
722 |
from io import BytesIO
|
723 |
-
|
724 |
pptx_content = await file.read()
|
725 |
pptx_file = BytesIO(pptx_content)
|
726 |
prs = Presentation(pptx_file)
|
@@ -728,7 +733,7 @@ def create_app(args):
|
|
728 |
for shape in slide.shapes:
|
729 |
if hasattr(shape, "text"):
|
730 |
content += shape.text + "\n"
|
731 |
-
|
732 |
case _:
|
733 |
failed_files.append(f"{file.filename} (unsupported type)")
|
734 |
continue
|
@@ -771,7 +776,6 @@ def create_app(args):
|
|
771 |
logging.error(f"Batch processing error: {str(e)}")
|
772 |
raise HTTPException(status_code=500, detail=str(e))
|
773 |
|
774 |
-
|
775 |
@app.delete(
|
776 |
"/documents",
|
777 |
response_model=InsertResponse,
|
|
|
175 |
class DocumentManager:
|
176 |
"""Handles document operations and tracking"""
|
177 |
|
178 |
+
def __init__(
|
179 |
+
self,
|
180 |
+
input_dir: str,
|
181 |
+
supported_extensions: tuple = (".txt", ".md", ".pdf", ".docx", ".pptx"),
|
182 |
+
):
|
183 |
self.input_dir = Path(input_dir)
|
184 |
self.supported_extensions = supported_extensions
|
185 |
self.indexed_files = set()
|
|
|
361 |
),
|
362 |
)
|
363 |
|
|
|
|
|
364 |
async def index_file(file_path: Union[str, Path]) -> None:
|
365 |
+
"""Index all files inside the folder with support for multiple file formats
|
366 |
+
|
367 |
Args:
|
368 |
file_path: Path to the file to be indexed (str or Path object)
|
369 |
+
|
370 |
Raises:
|
371 |
ValueError: If file format is not supported
|
372 |
FileNotFoundError: If file doesn't exist
|
373 |
"""
|
374 |
if not pm.is_installed("aiofiles"):
|
375 |
pm.install("aiofiles")
|
376 |
+
|
|
|
|
|
377 |
# Convert to Path object if string
|
378 |
file_path = Path(file_path)
|
379 |
+
|
380 |
# Check if file exists
|
381 |
if not file_path.exists():
|
382 |
raise FileNotFoundError(f"File not found: {file_path}")
|
|
|
384 |
content = ""
|
385 |
# Get file extension in lowercase
|
386 |
ext = file_path.suffix.lower()
|
387 |
+
|
388 |
match ext:
|
389 |
case ".txt" | ".md":
|
390 |
# Text files handling
|
391 |
async with aiofiles.open(file_path, "r", encoding="utf-8") as f:
|
392 |
content = await f.read()
|
393 |
+
|
394 |
case ".pdf":
|
395 |
if not pm.is_installed("pypdf2"):
|
396 |
pm.install("pypdf2")
|
397 |
from pypdf2 import PdfReader
|
398 |
+
|
399 |
# PDF handling
|
400 |
reader = PdfReader(str(file_path))
|
401 |
content = ""
|
402 |
for page in reader.pages:
|
403 |
content += page.extract_text() + "\n"
|
404 |
+
|
405 |
case ".docx":
|
406 |
if not pm.is_installed("docx"):
|
407 |
pm.install("docx")
|
|
|
410 |
# Word document handling
|
411 |
doc = Document(file_path)
|
412 |
content = "\n".join([paragraph.text for paragraph in doc.paragraphs])
|
413 |
+
|
414 |
case ".pptx":
|
415 |
if not pm.is_installed("pptx"):
|
416 |
pm.install("pptx")
|
417 |
from pptx import Presentation
|
418 |
+
|
419 |
# PowerPoint handling
|
420 |
prs = Presentation(file_path)
|
421 |
content = ""
|
|
|
423 |
for shape in slide.shapes:
|
424 |
if hasattr(shape, "text"):
|
425 |
content += shape.text + "\n"
|
426 |
+
|
427 |
case _:
|
428 |
raise ValueError(f"Unsupported file format: {ext}")
|
429 |
|
|
|
435 |
else:
|
436 |
logging.warning(f"No content extracted from file: {file_path}")
|
437 |
|
|
|
|
|
|
|
438 |
@app.on_event("startup")
|
439 |
async def startup_event():
|
440 |
"""Index all files in input directory during startup"""
|
|
|
558 |
)
|
559 |
except Exception as e:
|
560 |
raise HTTPException(status_code=500, detail=str(e))
|
561 |
+
|
562 |
@app.post(
|
563 |
"/documents/file",
|
564 |
response_model=InsertResponse,
|
|
|
566 |
)
|
567 |
async def insert_file(file: UploadFile = File(...), description: str = Form(None)):
|
568 |
"""Insert a file directly into the RAG system
|
569 |
+
|
570 |
Args:
|
571 |
file: Uploaded file
|
572 |
description: Optional description of the file
|
573 |
+
|
574 |
Returns:
|
575 |
InsertResponse: Status of the insertion operation
|
576 |
+
|
577 |
Raises:
|
578 |
HTTPException: For unsupported file types or processing errors
|
579 |
"""
|
|
|
581 |
content = ""
|
582 |
# Get file extension in lowercase
|
583 |
ext = Path(file.filename).suffix.lower()
|
584 |
+
|
585 |
match ext:
|
586 |
case ".txt" | ".md":
|
587 |
# Text files handling
|
588 |
text_content = await file.read()
|
589 |
content = text_content.decode("utf-8")
|
590 |
+
|
591 |
case ".pdf":
|
592 |
if not pm.is_installed("pypdf2"):
|
593 |
pm.install("pypdf2")
|
594 |
from pypdf2 import PdfReader
|
595 |
from io import BytesIO
|
596 |
+
|
597 |
# Read PDF from memory
|
598 |
pdf_content = await file.read()
|
599 |
pdf_file = BytesIO(pdf_content)
|
|
|
601 |
content = ""
|
602 |
for page in reader.pages:
|
603 |
content += page.extract_text() + "\n"
|
604 |
+
|
605 |
case ".docx":
|
606 |
if not pm.is_installed("docx"):
|
607 |
pm.install("docx")
|
608 |
from docx import Document
|
609 |
from io import BytesIO
|
610 |
+
|
611 |
# Read DOCX from memory
|
612 |
docx_content = await file.read()
|
613 |
docx_file = BytesIO(docx_content)
|
614 |
doc = Document(docx_file)
|
615 |
+
content = "\n".join(
|
616 |
+
[paragraph.text for paragraph in doc.paragraphs]
|
617 |
+
)
|
618 |
+
|
619 |
case ".pptx":
|
620 |
if not pm.is_installed("pptx"):
|
621 |
pm.install("pptx")
|
622 |
from pptx import Presentation
|
623 |
from io import BytesIO
|
624 |
+
|
625 |
# Read PPTX from memory
|
626 |
pptx_content = await file.read()
|
627 |
pptx_file = BytesIO(pptx_content)
|
|
|
631 |
for shape in slide.shapes:
|
632 |
if hasattr(shape, "text"):
|
633 |
content += shape.text + "\n"
|
634 |
+
|
635 |
case _:
|
636 |
raise HTTPException(
|
637 |
status_code=400,
|
|
|
643 |
# Add description if provided
|
644 |
if description:
|
645 |
content = f"{description}\n\n{content}"
|
646 |
+
|
647 |
await rag.ainsert(content)
|
648 |
logging.info(f"Successfully indexed file: {file.filename}")
|
649 |
+
|
650 |
return InsertResponse(
|
651 |
status="success",
|
652 |
message=f"File '{file.filename}' successfully inserted",
|
|
|
663 |
except Exception as e:
|
664 |
logging.error(f"Error processing file {file.filename}: {str(e)}")
|
665 |
raise HTTPException(status_code=500, detail=str(e))
|
666 |
+
|
667 |
@app.post(
|
668 |
"/documents/batch",
|
669 |
response_model=InsertResponse,
|
|
|
671 |
)
|
672 |
async def insert_batch(files: List[UploadFile] = File(...)):
|
673 |
"""Process multiple files in batch mode
|
674 |
+
|
675 |
Args:
|
676 |
files: List of files to process
|
677 |
+
|
678 |
Returns:
|
679 |
InsertResponse: Status of the batch insertion operation
|
680 |
+
|
681 |
Raises:
|
682 |
HTTPException: For processing errors
|
683 |
"""
|
|
|
689 |
try:
|
690 |
content = ""
|
691 |
ext = Path(file.filename).suffix.lower()
|
692 |
+
|
693 |
match ext:
|
694 |
case ".txt" | ".md":
|
695 |
text_content = await file.read()
|
696 |
content = text_content.decode("utf-8")
|
697 |
+
|
698 |
case ".pdf":
|
699 |
if not pm.is_installed("pypdf2"):
|
700 |
pm.install("pypdf2")
|
701 |
from pypdf2 import PdfReader
|
702 |
from io import BytesIO
|
703 |
+
|
704 |
pdf_content = await file.read()
|
705 |
pdf_file = BytesIO(pdf_content)
|
706 |
reader = PdfReader(pdf_file)
|
707 |
for page in reader.pages:
|
708 |
content += page.extract_text() + "\n"
|
709 |
+
|
710 |
case ".docx":
|
711 |
if not pm.is_installed("docx"):
|
712 |
pm.install("docx")
|
713 |
from docx import Document
|
714 |
from io import BytesIO
|
715 |
+
|
716 |
docx_content = await file.read()
|
717 |
docx_file = BytesIO(docx_content)
|
718 |
doc = Document(docx_file)
|
719 |
+
content = "\n".join(
|
720 |
+
[paragraph.text for paragraph in doc.paragraphs]
|
721 |
+
)
|
722 |
+
|
723 |
case ".pptx":
|
724 |
if not pm.is_installed("pptx"):
|
725 |
pm.install("pptx")
|
726 |
from pptx import Presentation
|
727 |
from io import BytesIO
|
728 |
+
|
729 |
pptx_content = await file.read()
|
730 |
pptx_file = BytesIO(pptx_content)
|
731 |
prs = Presentation(pptx_file)
|
|
|
733 |
for shape in slide.shapes:
|
734 |
if hasattr(shape, "text"):
|
735 |
content += shape.text + "\n"
|
736 |
+
|
737 |
case _:
|
738 |
failed_files.append(f"{file.filename} (unsupported type)")
|
739 |
continue
|
|
|
776 |
logging.error(f"Batch processing error: {str(e)}")
|
777 |
raise HTTPException(status_code=500, detail=str(e))
|
778 |
|
|
|
779 |
@app.delete(
|
780 |
"/documents",
|
781 |
response_model=InsertResponse,
|
lightrag/api/requirements.txt
CHANGED
@@ -7,6 +7,7 @@ nest_asyncio
|
|
7 |
numpy
|
8 |
ollama
|
9 |
openai
|
|
|
10 |
python-dotenv
|
11 |
python-multipart
|
12 |
tenacity
|
@@ -15,4 +16,3 @@ torch
|
|
15 |
tqdm
|
16 |
transformers
|
17 |
uvicorn
|
18 |
-
pipmaster
|
|
|
7 |
numpy
|
8 |
ollama
|
9 |
openai
|
10 |
+
pipmaster
|
11 |
python-dotenv
|
12 |
python-multipart
|
13 |
tenacity
|
|
|
16 |
tqdm
|
17 |
transformers
|
18 |
uvicorn
|
|