mirror of https://github.com/MISP/misp-modules
fix: [ocr_enrich] Fixed tesseract input format
- It looks like the `image_to_string` method now assumes RGB format and the `imdecode` method seems to give BGR format, so we convert the image array beforepull/488/head
parent
859e490173
commit
611bb6fa9e
|
@ -6,14 +6,21 @@ import pytesseract
|
|||
|
||||
misperrors = {'error': 'Error'}
|
||||
mispattributes = {'input': ['attachment'],
|
||||
'output': ['freetext', 'text']}
|
||||
moduleinfo = {'version': '0.1', 'author': 'Sascha Rommelfangen',
|
||||
'output': ['freetext']}
|
||||
moduleinfo = {'version': '0.2', 'author': 'Sascha Rommelfangen',
|
||||
'description': 'OCR decoder',
|
||||
'module-type': ['expansion']}
|
||||
|
||||
moduleconfig = []
|
||||
|
||||
|
||||
def filter_decoded(decoded):
|
||||
for line in decoded.split('\n'):
|
||||
decoded_line = line.strip('\t\x0b\x0c\r ')
|
||||
if decoded_line:
|
||||
yield decoded_line
|
||||
|
||||
|
||||
def handler(q=False):
|
||||
if q is False:
|
||||
return False
|
||||
|
@ -31,9 +38,16 @@ def handler(q=False):
|
|||
image = img_array
|
||||
image = cv2.imdecode(img_array, cv2.IMREAD_COLOR)
|
||||
try:
|
||||
decoded = pytesseract.image_to_string(image)
|
||||
return {'results': [{'types': ['freetext'], 'values': decoded, 'comment': "OCR from file " + filename},
|
||||
{'types': ['text'], 'values': decoded, 'comment': "ORC from file " + filename}]}
|
||||
decoded = pytesseract.image_to_string(cv2.cvtColor(image, cv2.COLOR_BGR2RGB))
|
||||
return {
|
||||
'results':[
|
||||
{
|
||||
'types': ['freetext'],
|
||||
'values': list(filter_decoded(decoded)),
|
||||
'comment': f"OCR from file {filename}"
|
||||
}
|
||||
]
|
||||
}
|
||||
except Exception as e:
|
||||
print(e)
|
||||
err = "Couldn't analyze file type. Only images are supported right now."
|
||||
|
|
Loading…
Reference in New Issue