Spaces:
Sleeping
Sleeping
Commit
·
6cf24d5
1
Parent(s):
7ff92da
Added a button for download- testing event listener is attached correctly
Browse files
app.py
CHANGED
|
@@ -32,26 +32,6 @@ def get_named_entities(ocr_text: str):
|
|
| 32 |
return entities
|
| 33 |
|
| 34 |
|
| 35 |
-
# If you don't have tesseract executable in your PATH, include the following:
|
| 36 |
-
# pytesseract.pytesseract.tesseract_cmd = r'<full_path_to_your_tesseract_executable>'
|
| 37 |
-
# Example tesseract_cmd = r'C:\Program Files (x86)\Tesseract-OCR\tesseract'
|
| 38 |
-
|
| 39 |
-
# Simple image to string
|
| 40 |
-
# print(pytesseract.image_to_string(Image.open('eurotext.png')))
|
| 41 |
-
|
| 42 |
-
# # French text image to string
|
| 43 |
-
# print(pytesseract.image_to_string(Image.open('test-european.jpg'), lang='fra'))
|
| 44 |
-
|
| 45 |
-
# # Get bounding box estimates
|
| 46 |
-
# print(pytesseract.image_to_boxes(Image.open('test.png')))
|
| 47 |
-
|
| 48 |
-
# # Get verbose data including boxes, confidences, line and page numbers
|
| 49 |
-
# print(pytesseract.image_to_data(Image.open('test.png')))
|
| 50 |
-
|
| 51 |
-
# # Get information about orientation and script detection
|
| 52 |
-
# print(pytesseract.image_to_osd(Image.open('test.png'))
|
| 53 |
-
|
| 54 |
-
|
| 55 |
def run(image, lang="eng"):
|
| 56 |
result = pytesseract.image_to_string(image, lang=None if lang == [] else lang)
|
| 57 |
|
|
@@ -59,6 +39,15 @@ def run(image, lang="eng"):
|
|
| 59 |
return result, ner
|
| 60 |
|
| 61 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 62 |
with gr.Blocks() as demo:
|
| 63 |
gr.Markdown("## Theatre Programmer")
|
| 64 |
with gr.Row():
|
|
@@ -67,12 +56,15 @@ with gr.Blocks() as demo:
|
|
| 67 |
lang = gr.Dropdown(choices, value="eng")
|
| 68 |
btn = gr.Button("Run")
|
| 69 |
with gr.Column():
|
| 70 |
-
|
| 71 |
with gr.Column():
|
| 72 |
ner = gr.TextArea(label="Named entities")
|
| 73 |
# with gr.Column():
|
| 74 |
# gr.CheckboxGroup(ner, label="Named entities")
|
|
|
|
|
|
|
| 75 |
|
| 76 |
-
btn.click(fn=run, inputs=[image_in, lang], outputs=[
|
|
|
|
| 77 |
|
| 78 |
demo.launch()
|
|
|
|
| 32 |
return entities
|
| 33 |
|
| 34 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 35 |
def run(image, lang="eng"):
|
| 36 |
result = pytesseract.image_to_string(image, lang=None if lang == [] else lang)
|
| 37 |
|
|
|
|
| 39 |
return result, ner
|
| 40 |
|
| 41 |
|
| 42 |
+
def download_output(ocr_text: str, named_entities: str):
|
| 43 |
+
print("Download output!")
|
| 44 |
+
|
| 45 |
+
print("OCR text: ", len(ocr_text))
|
| 46 |
+
print("Named Entities: ", len(named_entities))
|
| 47 |
+
|
| 48 |
+
return True
|
| 49 |
+
|
| 50 |
+
|
| 51 |
with gr.Blocks() as demo:
|
| 52 |
gr.Markdown("## Theatre Programmer")
|
| 53 |
with gr.Row():
|
|
|
|
| 56 |
lang = gr.Dropdown(choices, value="eng")
|
| 57 |
btn = gr.Button("Run")
|
| 58 |
with gr.Column():
|
| 59 |
+
ocr_text = gr.TextArea(label="OCR output")
|
| 60 |
with gr.Column():
|
| 61 |
ner = gr.TextArea(label="Named entities")
|
| 62 |
# with gr.Column():
|
| 63 |
# gr.CheckboxGroup(ner, label="Named entities")
|
| 64 |
+
with gr.Row():
|
| 65 |
+
download_btn = gr.Button("Download output")
|
| 66 |
|
| 67 |
+
btn.click(fn=run, inputs=[image_in, lang], outputs=[ocr_text, ner])
|
| 68 |
+
download_btn.click(fn=download_output, inputs=[ocr_text, ner], outputs=[])
|
| 69 |
|
| 70 |
demo.launch()
|