Skip to content

Commit

Permalink
Merge c19c341 into 4d61a0e
Browse files Browse the repository at this point in the history
  • Loading branch information
AminovE99 committed Nov 20, 2019
2 parents 4d61a0e + c19c341 commit ea78687
Showing 1 changed file with 11 additions and 6 deletions.
17 changes: 11 additions & 6 deletions erpnext_ocr/erpnext_ocr/doctype/ocr_read/ocr_read.py
Original file line number Diff line number Diff line change
Expand Up @@ -50,35 +50,40 @@ def read_document(path, lang='eng'):
# external link
fullpath = requests.get(path, stream=True).raw

frappe.publish_realtime("ocr_progress_bar", {"progress": "0"}, user=frappe.session.user)
text = " "

if path.endswith('.pdf'):
from wand.image import Image as wi
pdf = wi(filename=fullpath, resolution=300)
pdf_image = pdf.convert('jpeg')
i = 0
size = len(pdf_image.sequence)
size = len(pdf_image.sequence) * 3
for img in pdf_image.sequence:
img_page = wi(image=img)
image_blob = img_page.make_blob('jpeg')
frappe.publish_realtime("ocr_progress_bar", {"progress": [i, size]}, user=frappe.session.user)
i += 1

recognized_text = " "

image = Image.open(io.BytesIO(image_blob))
frappe.publish_realtime("ocr_progress_bar", {"progress": [i, size]}, user=frappe.session.user)
i += 1

recognized_text = pytesseract.image_to_string(image, lang)
text = text + recognized_text

frappe.publish_realtime("ocr_progress_bar", {"progress": [i, size]})
frappe.publish_realtime("ocr_progress_bar", {"progress": [i, size]}, user=frappe.session.user)
i += 1

else:
frappe.publish_realtime("ocr_progress_bar", {"progress": "0"}, user=frappe.session.user)

image = Image.open(fullpath)
frappe.publish_realtime("ocr_progress_bar", {"progress": [33, 100]}, user=frappe.session.user)

text = pytesseract.image_to_string(image, lang=lang)
frappe.publish_realtime("ocr_progress_bar", {"progress": [66, 100]}, user=frappe.session.user)

text.split(" ")
frappe.publish_realtime("ocr_progress_bar", {"progress": [100, 100]}, user=frappe.session.user)

return text

Expand Down

0 comments on commit ea78687

Please sign in to comment.