Update app.py
Browse files
app.py
CHANGED
@@ -1,51 +1,59 @@
|
|
|
|
1 |
from transformers import TrOCRProcessor, VisionEncoderDecoderModel
|
2 |
from PIL import Image
|
3 |
import os
|
4 |
import re
|
5 |
|
6 |
-
# Load
|
7 |
processor = TrOCRProcessor.from_pretrained("microsoft/trocr-base-stage1")
|
8 |
model = VisionEncoderDecoderModel.from_pretrained("microsoft/trocr-base-stage1")
|
9 |
|
10 |
-
#
|
11 |
-
PATIENT_RECORDS_DIR = "records
|
|
|
12 |
|
13 |
-
#
|
14 |
def extract_patient_name(file_name):
|
15 |
-
|
16 |
-
|
17 |
-
|
18 |
-
|
19 |
-
|
20 |
-
|
21 |
-
|
22 |
-
image = Image.open(image_path).convert("RGB")
|
23 |
pixel_values = processor(images=image, return_tensors="pt").pixel_values
|
24 |
generated_ids = model.generate(pixel_values)
|
25 |
-
|
26 |
-
return
|
27 |
-
|
28 |
-
# Save
|
29 |
-
def
|
30 |
-
os.
|
31 |
-
|
32 |
-
|
33 |
-
|
34 |
-
|
35 |
-
|
36 |
-
|
37 |
-
|
38 |
-
|
39 |
-
|
40 |
if not patient_name:
|
41 |
-
return "β
|
|
|
|
|
|
|
|
|
|
|
42 |
|
43 |
-
|
44 |
-
|
45 |
-
|
|
|
|
|
|
|
|
|
|
|
46 |
|
47 |
-
# Example usage
|
48 |
if __name__ == "__main__":
|
49 |
-
|
50 |
-
result = process_uploaded_lab_result(file_to_upload)
|
51 |
-
print(result)
|
|
|
1 |
+
import gradio as gr
|
2 |
from transformers import TrOCRProcessor, VisionEncoderDecoderModel
|
3 |
from PIL import Image
|
4 |
import os
|
5 |
import re
|
6 |
|
7 |
+
# Load OCR model
|
8 |
processor = TrOCRProcessor.from_pretrained("microsoft/trocr-base-stage1")
|
9 |
model = VisionEncoderDecoderModel.from_pretrained("microsoft/trocr-base-stage1")
|
10 |
|
11 |
+
# Folder to store extracted records
|
12 |
+
PATIENT_RECORDS_DIR = "records"
|
13 |
+
os.makedirs(PATIENT_RECORDS_DIR, exist_ok=True)
|
14 |
|
15 |
+
# Extract patient name from filename
|
16 |
def extract_patient_name(file_name):
|
17 |
+
# Example: JuanDelaCruz_2025-06-13.png β JuanDelaCruz
|
18 |
+
match = re.match(r"([A-Za-z]+[A-Za-z]*)_.*\.(jpg|jpeg|png)$", file_name)
|
19 |
+
return match.group(1) if match else None
|
20 |
+
|
21 |
+
# OCR logic
|
22 |
+
def perform_ocr(image_file):
|
23 |
+
image = Image.open(image_file).convert("RGB")
|
|
|
24 |
pixel_values = processor(images=image, return_tensors="pt").pixel_values
|
25 |
generated_ids = model.generate(pixel_values)
|
26 |
+
text = processor.batch_decode(generated_ids, skip_special_tokens=True)[0].strip()
|
27 |
+
return text
|
28 |
+
|
29 |
+
# Save to patient record file
|
30 |
+
def save_record(patient_name, ocr_text):
|
31 |
+
file_path = os.path.join(PATIENT_RECORDS_DIR, f"{patient_name}_records.txt")
|
32 |
+
with open(file_path, "a") as f:
|
33 |
+
f.write("\n\n===== New Lab Result =====\n")
|
34 |
+
f.write(ocr_text)
|
35 |
+
|
36 |
+
# Gradio function
|
37 |
+
def process_lab_result(image):
|
38 |
+
file_name = os.path.basename(image.name)
|
39 |
+
patient_name = extract_patient_name(file_name)
|
40 |
+
|
41 |
if not patient_name:
|
42 |
+
return "β Cannot extract patient name from filename. Please name the file like JuanDelaCruz_2025-06-13.png"
|
43 |
+
|
44 |
+
ocr_text = perform_ocr(image)
|
45 |
+
save_record(patient_name, ocr_text)
|
46 |
+
|
47 |
+
return f"β
OCR completed. Lab result saved to `{patient_name}_records.txt`.\n\n---\nπ Extracted Text:\n{ocr_text}"
|
48 |
|
49 |
+
# Gradio interface
|
50 |
+
iface = gr.Interface(
|
51 |
+
fn=process_lab_result,
|
52 |
+
inputs=gr.File(label="Upload Lab Result Image (.png, .jpg)", type="file"),
|
53 |
+
outputs="text",
|
54 |
+
title="π©Ί Lab Result OCR with Patient Linking",
|
55 |
+
description="Upload a lab result image named like `JuanDelaCruz_2025-06-13.png`. The system will extract the text and save it to the patient's record."
|
56 |
+
)
|
57 |
|
|
|
58 |
if __name__ == "__main__":
|
59 |
+
iface.launch()
|
|
|
|