Spaces:
Sleeping
Sleeping
Commit
·
78bc6bc
1
Parent(s):
9618468
Fix: Replace magic_pdf.api import with direct magic_pdf usage
Browse files
app.py
CHANGED
|
@@ -1,6 +1,9 @@
|
|
| 1 |
from fastapi import FastAPI, UploadFile, File
|
| 2 |
from fastapi.responses import JSONResponse
|
| 3 |
-
|
|
|
|
|
|
|
|
|
|
| 4 |
|
| 5 |
app = FastAPI()
|
| 6 |
|
|
@@ -8,7 +11,34 @@ app = FastAPI()
|
|
| 8 |
async def extract(file: UploadFile = File(...)):
|
| 9 |
content = await file.read()
|
| 10 |
try:
|
| 11 |
-
|
| 12 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 13 |
except Exception as e:
|
| 14 |
return JSONResponse(status_code=500, content={"error": str(e)})
|
|
|
|
| 1 |
from fastapi import FastAPI, UploadFile, File
|
| 2 |
from fastapi.responses import JSONResponse
|
| 3 |
+
import magic_pdf
|
| 4 |
+
import tempfile
|
| 5 |
+
import os
|
| 6 |
+
import json
|
| 7 |
|
| 8 |
app = FastAPI()
|
| 9 |
|
|
|
|
| 11 |
async def extract(file: UploadFile = File(...)):
|
| 12 |
content = await file.read()
|
| 13 |
try:
|
| 14 |
+
# Save the uploaded PDF to a temporary file
|
| 15 |
+
with tempfile.NamedTemporaryFile(suffix=".pdf", delete=False) as temp_pdf:
|
| 16 |
+
temp_pdf.write(content)
|
| 17 |
+
temp_pdf_path = temp_pdf.name
|
| 18 |
+
|
| 19 |
+
# Process the PDF using magic_pdf.PDF class
|
| 20 |
+
result = magic_pdf.PDF(temp_pdf_path).parse()
|
| 21 |
+
|
| 22 |
+
# Convert result to dictionary
|
| 23 |
+
output = {
|
| 24 |
+
"pages": []
|
| 25 |
+
}
|
| 26 |
+
|
| 27 |
+
for page in result.pages:
|
| 28 |
+
page_data = {
|
| 29 |
+
"page_num": page.page_num,
|
| 30 |
+
"text": "\n".join([block.text for block in page.text_blocks]),
|
| 31 |
+
"tables": []
|
| 32 |
+
}
|
| 33 |
+
|
| 34 |
+
for table in page.tables:
|
| 35 |
+
page_data["tables"].append(table.to_markdown())
|
| 36 |
+
|
| 37 |
+
output["pages"].append(page_data)
|
| 38 |
+
|
| 39 |
+
# Clean up the temporary file
|
| 40 |
+
os.unlink(temp_pdf_path)
|
| 41 |
+
|
| 42 |
+
return {"result": output}
|
| 43 |
except Exception as e:
|
| 44 |
return JSONResponse(status_code=500, content={"error": str(e)})
|