paradizelost/paperless-ngx
1
0
Fork 0
mirror of https://github.com/paperless-ngx/paperless-ngx.git synced 2025-04-27 10:59:27 -05:00
Code Issues Packages Projects Releases Wiki Activity

Detect language only on one page of PDF

Browse Source 
To detect the language currently the entire document gets processed. If
a different language has been detected than the default one, the entire
document will be processed again for the new language.

This PR analyzes the middle page for its language and either processes
the remaining pages with the default language if it didn't differ, or
processes all pages for the new guessed language.

The amount of processed pages comes down from the worst case `2n` to
worst case `n+1`.
This commit is contained in:
Pit Kleyersburg 2016-02-14 16:13:34 +01:00
parent 9162e41507
commit aeab9a0e81
1 changed files with 16 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

17
src/documents/consumer.py
View File

@ -9,6 +9,7 @@ import random
import re
import subprocess
import math
import pyocr
from PIL import Image
@ -152,9 +153,14 @@ class Consumer(Renderable):
simple language detection trial & error.
"""
if not pngs:
raise OCRError
self._render(" OCRing the document", 2)
raw_text = self._ocr(pngs, self.DEFAULT_OCR_LANGUAGE)
# Since the division gets rounded down by int, this calculation works for every edge-case, i.e. 1
middle = int(len(pngs) / 2)
raw_text = self._ocr([pngs[middle]], self.DEFAULT_OCR_LANGUAGE)
guessed_language = self._guess_language(raw_text)
@ -166,10 +172,14 @@ class Consumer(Renderable):
"with what we have.",
1
)
raw_text = self._ocr(pngs[:middle], self.DEFAULT_OCR_LANGUAGE) + raw_text
raw_text += self._ocr(pngs[middle+1:], self.DEFAULT_OCR_LANGUAGE)
return raw_text
raise OCRError
if ISO639[guessed_language] == self.DEFAULT_OCR_LANGUAGE:
raw_text = self._ocr(pngs[:middle], self.DEFAULT_OCR_LANGUAGE) + raw_text
raw_text += self._ocr(pngs[middle+1:], self.DEFAULT_OCR_LANGUAGE)
return raw_text
try:
@ -183,6 +193,8 @@ class Consumer(Renderable):
),
0
)
raw_text = self._ocr(pngs[:middle], self.DEFAULT_OCR_LANGUAGE) + raw_text
raw_text += self._ocr(pngs[middle+1:], self.DEFAULT_OCR_LANGUAGE)
return raw_text
raise OCRError
@ -191,6 +203,9 @@ class Consumer(Renderable):
Performs a single OCR attempt.
"""
if not pngs:
return ""
self._render(" Parsing for {}".format(lang), 2)
with Pool(processes=self.THREADS) as pool: