Spaces:
Running
Running
Update app.py
Browse files
app.py
CHANGED
|
@@ -4,8 +4,21 @@ import os
|
|
| 4 |
import tempfile
|
| 5 |
from pathlib import Path
|
| 6 |
|
| 7 |
-
#
|
|
|
|
| 8 |
from working_yolo_pipeline import run_document_pipeline, DEFAULT_LAYOUTLMV3_MODEL_PATH, WEIGHTS_PATH
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 9 |
|
| 10 |
|
| 11 |
def process_pdf(pdf_file, layoutlmv3_model_path=None):
|
|
@@ -47,16 +60,13 @@ def process_pdf(pdf_file, layoutlmv3_model_path=None):
|
|
| 47 |
output_filename = f"{Path(pdf_path).stem}_analysis.json"
|
| 48 |
temp_output = tempfile.NamedTemporaryFile(mode='w', delete=False, suffix='.json', prefix='analysis_')
|
| 49 |
|
|
|
|
| 50 |
with open(temp_output.name, 'w', encoding='utf-8') as f:
|
| 51 |
json.dump(result, f, indent=2, ensure_ascii=False)
|
| 52 |
|
| 53 |
# Format JSON for display
|
| 54 |
json_display = json.dumps(result, indent=2, ensure_ascii=False)
|
| 55 |
|
| 56 |
-
# Truncate if too long for display
|
| 57 |
-
# if len(json_display) > 50000:
|
| 58 |
-
# json_display = json_display[:50000] + "\n\n... (truncated for display, download full file)"
|
| 59 |
-
|
| 60 |
return json_display, temp_output.name
|
| 61 |
|
| 62 |
except Exception as e:
|
|
@@ -64,7 +74,8 @@ def process_pdf(pdf_file, layoutlmv3_model_path=None):
|
|
| 64 |
|
| 65 |
|
| 66 |
# Create Gradio interface
|
| 67 |
-
|
|
|
|
| 68 |
gr.Markdown("""
|
| 69 |
# π Document Analysis Pipeline
|
| 70 |
|
|
|
|
| 4 |
import tempfile
|
| 5 |
from pathlib import Path
|
| 6 |
|
| 7 |
+
# NOTE: You must ensure that 'working_yolo_pipeline.py' exists
|
| 8 |
+
# and defines the following items correctly:
|
| 9 |
from working_yolo_pipeline import run_document_pipeline, DEFAULT_LAYOUTLMV3_MODEL_PATH, WEIGHTS_PATH
|
| 10 |
+
# Since I don't have this file, I am assuming the imports are correct.
|
| 11 |
+
|
| 12 |
+
# Define placeholders for assumed constants if the pipeline file isn't present
|
| 13 |
+
# You should replace these with your actual definitions if they are missing
|
| 14 |
+
try:
|
| 15 |
+
from working_yolo_pipeline import run_document_pipeline, DEFAULT_LAYOUTLMV3_MODEL_PATH, WEIGHTS_PATH
|
| 16 |
+
except ImportError:
|
| 17 |
+
print("Warning: 'working_yolo_pipeline.py' not found. Using dummy paths.")
|
| 18 |
+
def run_document_pipeline(*args):
|
| 19 |
+
return {"error": "Placeholder pipeline function called."}
|
| 20 |
+
DEFAULT_LAYOUTLMV3_MODEL_PATH = "./models/layoutlmv3_model"
|
| 21 |
+
WEIGHTS_PATH = "./weights/yolo_weights.pt"
|
| 22 |
|
| 23 |
|
| 24 |
def process_pdf(pdf_file, layoutlmv3_model_path=None):
|
|
|
|
| 60 |
output_filename = f"{Path(pdf_path).stem}_analysis.json"
|
| 61 |
temp_output = tempfile.NamedTemporaryFile(mode='w', delete=False, suffix='.json', prefix='analysis_')
|
| 62 |
|
| 63 |
+
# Dump results to the temporary file
|
| 64 |
with open(temp_output.name, 'w', encoding='utf-8') as f:
|
| 65 |
json.dump(result, f, indent=2, ensure_ascii=False)
|
| 66 |
|
| 67 |
# Format JSON for display
|
| 68 |
json_display = json.dumps(result, indent=2, ensure_ascii=False)
|
| 69 |
|
|
|
|
|
|
|
|
|
|
|
|
|
| 70 |
return json_display, temp_output.name
|
| 71 |
|
| 72 |
except Exception as e:
|
|
|
|
| 74 |
|
| 75 |
|
| 76 |
# Create Gradio interface
|
| 77 |
+
# FIX APPLIED: Removed 'theme=gr.themes.Soft()' which caused the TypeError
|
| 78 |
+
with gr.Blocks(title="Document Analysis Pipeline") as demo:
|
| 79 |
gr.Markdown("""
|
| 80 |
# π Document Analysis Pipeline
|
| 81 |
|