Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -1,6 +1,8 @@
|
|
| 1 |
from fastapi import FastAPI, File, UploadFile
|
| 2 |
import pdfplumber
|
| 3 |
import io
|
|
|
|
|
|
|
| 4 |
|
| 5 |
app = FastAPI()
|
| 6 |
|
|
@@ -33,6 +35,33 @@ async def extract(file: UploadFile = File(...)):
|
|
| 33 |
"stream": stream
|
| 34 |
}
|
| 35 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 36 |
@app.get("/health")
|
| 37 |
async def health():
|
| 38 |
return {
|
|
|
|
| 1 |
from fastapi import FastAPI, File, UploadFile
|
| 2 |
import pdfplumber
|
| 3 |
import io
|
| 4 |
+
from markitdown import MarkItDown
|
| 5 |
+
import shutil
|
| 6 |
|
| 7 |
app = FastAPI()
|
| 8 |
|
|
|
|
| 35 |
"stream": stream
|
| 36 |
}
|
| 37 |
|
| 38 |
+
md = MarkItDown()
|
| 39 |
+
|
| 40 |
+
@app.post("/convert")
|
| 41 |
+
async def convert_to_markdown(file: UploadFile = File(...)):
|
| 42 |
+
temp_path = f"temp_{file.filename}"
|
| 43 |
+
|
| 44 |
+
try:
|
| 45 |
+
with open(temp_path, "wb") as buffer:
|
| 46 |
+
content = await file.read()
|
| 47 |
+
buffer.write(content)
|
| 48 |
+
|
| 49 |
+
# MarkItDown conversion
|
| 50 |
+
result = md.convert(temp_path)
|
| 51 |
+
|
| 52 |
+
return {
|
| 53 |
+
"filename": file.filename,
|
| 54 |
+
"markdown": result.text_content
|
| 55 |
+
}
|
| 56 |
+
|
| 57 |
+
except Exception as e:
|
| 58 |
+
# Added missing HTTPException import
|
| 59 |
+
raise HTTPException(status_code=500, detail=str(e))
|
| 60 |
+
|
| 61 |
+
finally:
|
| 62 |
+
if os.path.exists(temp_path):
|
| 63 |
+
os.remove(temp_path)
|
| 64 |
+
|
| 65 |
@app.get("/health")
|
| 66 |
async def health():
|
| 67 |
return {
|