Thback commited on
Commit
2759478
β€’
1 Parent(s): 59b5f1e

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +19 -7
app.py CHANGED
@@ -26,13 +26,25 @@ def get_pdf_text(pdf_docs):
26
  # μ•„λž˜ ν…μŠ€νŠΈ μΆ”μΆœ ν•¨μˆ˜λ₯Ό μž‘μ„±
27
 
28
  def get_text_file(text_docs):
29
- temp_dir = tempfile.TemporaryDirectory() # μž„μ‹œ 디렉토리λ₯Ό μƒμ„±ν•©λ‹ˆλ‹€.
30
- temp_filepath = os.path.join(temp_dir.name, text_docs.name) # μž„μ‹œ 파일 경둜λ₯Ό μƒμ„±ν•©λ‹ˆλ‹€.
31
- with open(temp_filepath, "wb") as f: # μž„μ‹œ νŒŒμΌμ„ λ°”μ΄λ„ˆλ¦¬ μ“°κΈ° λͺ¨λ“œλ‘œ μ—½λ‹ˆλ‹€.
32
- f.write(text_docs.getvalue()) # PDF λ¬Έμ„œμ˜ λ‚΄μš©μ„ μž„μ‹œ νŒŒμΌμ— μ”λ‹ˆλ‹€.
33
- text_loader = TextLoader(temp_filepath) # PyPDFLoaderλ₯Ό μ‚¬μš©ν•΄ PDFλ₯Ό λ‘œλ“œν•©λ‹ˆλ‹€.
34
- text_doc = text_loader.load() # ν…μŠ€νŠΈλ₯Ό μΆ”μΆœν•©λ‹ˆλ‹€.
35
- return text_doc # μΆ”μΆœν•œ ν…μŠ€νŠΈλ₯Ό λ°˜ν™˜ν•©λ‹ˆλ‹€.
 
 
 
 
 
 
 
 
 
 
 
 
36
 
37
 
38
 
 
26
  # μ•„λž˜ ν…μŠ€νŠΈ μΆ”μΆœ ν•¨μˆ˜λ₯Ό μž‘μ„±
27
 
28
  def get_text_file(text_docs):
29
+ # μž„μ‹œ 디렉토리λ₯Ό μƒμ„±ν•©λ‹ˆλ‹€.
30
+ temp_dir = tempfile.TemporaryDirectory()
31
+
32
+ # μž„μ‹œ 파일 경둜λ₯Ό μƒμ„±ν•©λ‹ˆλ‹€.
33
+ temp_filepath = os.path.join(temp_dir.name, "temp_file.txt")
34
+
35
+ # ν…μŠ€νŠΈ λ¬Έμ„œμ˜ λ‚΄μš©μ„ μž„μ‹œ νŒŒμΌμ— μ”λ‹ˆλ‹€.
36
+ with open(temp_filepath, "w", encoding="utf-8") as f:
37
+ f.write(text_docs.getvalue())
38
+
39
+ # μž„μ‹œ νŒŒμΌμ—μ„œ ν…μŠ€νŠΈλ₯Ό μ½μ–΄μ˜΅λ‹ˆλ‹€.
40
+ with open(temp_filepath, "r", encoding="utf-8") as f:
41
+ text_content = f.read()
42
+
43
+ # μž„μ‹œ 디렉토리λ₯Ό μžλ™μœΌλ‘œ μ •λ¦¬ν•©λ‹ˆλ‹€.
44
+ temp_dir.cleanup()
45
+
46
+ return [text_content] # μΆ”μΆœν•œ ν…μŠ€νŠΈλ₯Ό λ¦¬μŠ€νŠΈμ— λ‹΄μ•„ λ°˜ν™˜ν•©λ‹ˆλ‹€.
47
+
48
 
49
 
50